Genesis by James Miller

Shooting spectacular raw video on the 5D Mark III requires UDMA 7 compact flash cards. Ideally you need a 1000x 64GB or 128GB card and certainly more than one for anything but very short shoots.

Here’s my guide to which ones to go for.

There are a lot of 1000x compact flash cards out there, with big variance in performance and price. Personally I have had the best results with the KomputerBay 64GB 1000x card, which only cost me 89 Euros on Amazon. Some KomputerBay and Lexar cards have been able to achieve 120MB/s sustained write speeds in Magic Lantern’s benchmark test, in theory enough for a Blackmagic Cinema Camera rivalling 2.5K 2560 x 1152 which requires 118.1MB/s at 24fps. That’s also fast enough for slow-mo 1720 x 672 at 60fps.

However at the moment sustained raw video write speeds on the 5D Mark III whilst recording are lower than the benchmark scores. The best I’ve been able to do relatively reliably is 1920 x 1080, 1720 x 1280 and 2880 x 720. Pretty great but I’d love to be able to get my 1920 x 1280 reliable for anamorphic, currently it stops after around 10-20 seconds, go for even higher vertical resolutions in 2.8K 1:1 crop mode and get reliable 25p plus 48p at lower resolutions for slow-motion work.

Is the 4GB file size a limitation? Thankfully no, with future build of Magic Lantern you can go until the card is full as the 5D Mark III supports the more modern ExFat file system for larger files than 4GB. Already people have recorded long takes of raw video in a single 12GB file!

Click here to view the embedded video.

Work by the Magic Lantern developers is being done to optimise the way the code handles the buffer and card writing to improve speeds. I’ve also found that the first couple of clips you record tend to start off slowly before gradually climbing, so the risk of dropped frames is higher when you first turn on the camera and start recording. It is sensible to leave some margin and not push resolution or frame rate right to the limit. 1920 x 720 at 24fps (as in James’s video above) is a sweet spot for most of the 1000x cards.

Here’s the full benchmark chart from A1ex at Magic Lantern -

(Click to enlarge)

Below is the benchmark for SD cards. As you can see the card controller on the 600D and 550D seems to top out at 20MB/s regardless of card speed but the Canon 6D is twice as fast. Both cameras are still a long way from the 5D Mark III but the SD card slot on that also suffers from much slower performance than the CF card interface does.

So which are the current best cards for the 5D Mark III?

Here’s a guide…

KomputerBay 1000x

128GB – $189 – Buy It Now

64GB – $106 – Buy It Now

What IS the story with KomputerBay? They appeared from nowhere and offer very good performance for very small prices. Well the truth has been outed. These are quality control failed memory chips. Is this a bad thing? Actually no. In the same way that Red offer Q/C failed Epic parts in the form of the much cheaper Scarlet, rather than dumping the hardware altogether, KomputerBay card just has more variance in performance. Reliability seems unaffected. Thanks to a happy coincidence reported here at DVXUser we now know the memory inside was destined for the much more expensive Lexar 1000x cards.

“Interesting fact I learned today that affects anyone currently buying 1000x cards at the moment for the ML raw hack. I set out this morning to buy Komputerbay cards via Amazon. I wanted them overnight. After ordering I noticed they werent coming until Wednesday. I canceled the order, then tried contacting Komputerbay directly. I noticed it was a 770- area code. 770 is my area code as well. Turns out they are 15 mins from me. I contacted them and picked up locally. I had a long talk with the senior manager and he told me the FULL truth about these cards. These cards have the IDENTICAL components as the LEXAR 1000x cards. He even went as far as slicing a few open to show me the identical Micron chips inside same as LEXAR. He was very matter of fact about this info. He went on to tell me exactly how and why this was possible. He said the only difference was slight QC variance from Lexar. Keep on mind these cards come direct from Komputerbay with a 100 percent guarentee. Meaning that if you get a card that was missed in the QC variance then return and get a different one. Rest assured I am only a filmaker and not a CF card seller. Please take this information as a gift as I see it as VERY reliable having seen and heard with my own two eyes and ears.”

Indeed this backs up exactly what I am seeing with my own two KomputerBay cards.

The 128GB 1000X tops out at around 73MB/s write and 122MB/s read. The chipset is SiliconMotion SM2236AC.

However my 64GB 1000x does 95MB/s write and 120MB/s read. Same chips.

When buying a KomputerBay card, some capacities will likely have greater variance than others. The 64GB cards certainly seem to be achieving the better results. Make sure you purchase from eBay or Amazon where you have the best return policies in case of a slower card.

When you get a good one, you get a total bargain.

Lexar 1000x

128GB – $654 – Buy It Now

64GB – $309 – Buy It Now

If for an urgent shoot you need to guarantee a good card without having to take a chance on KomputerBay parts and return them you can spend more money to solve the problem. The thing which discourages me from spending so much money on memory though, is that prices are notoriously volatile and constantly decreasing over time.

Toshiba 1066x

64GB – $299 – Buy It Now

Now here’s a genuinely exciting option – Toshiba quote actual write speeds on the card label. They claim 150MB/s!

Will it live up to expectations? Nobody has tested the card yet. As soon as the scores come in I will update the blog.

I am very tempted to buy one of these myself as $300 isn’t bad should the card live up to Toshiba’s spec.

My feeling is that the 150MB/s is a peak burst rate not sustained. In the little video clip logo on the card they quote 65MB/s sustained and reliable. However the Transcend 1000x cards quotes 20MB/s for this video standard, yet achieve sustained raw video at 80MB/s with Magic Lantern. It’s a VERY conservative figure.

Sandisk Extreme Pro 100MB/s

128GB – $832 – Buy It Now

64GB – $339 – Buy It Now

Andrew Wonder has tested the 128GB card and isn’t quite getting the raw resolutions I am getting with my KomputerBay 64GB.

I can’t really recommend SanDisk right now. They do a premium 100MB/s 128GB card and a much cheaper 90MB/s 64GB card and both frankly are absurdly overpriced. Why so much more than the 95MB/s Extreme Pro SD cards? The 128GB 100MB/s model does basically what it says on the box – 100MB/s write speeds – that is a full 20MB/s off the KomputerBay 64GB on the Magic Lantern benchmarks for 5x the price. Hardly value for money OR top performance. I think Sandisk need to up their game here. The cheapest card I found in the US was over $1000 and the cheapest genuine one from Hong Kong is linked to above.

Hoodman Steel 1000x

64GB – $330 – Buy It Now

The Hoodman Steel is advertised as having SSD NAND flash chips inside to get as close to the 167MB/s limit of UDMA 7 as possible. However the Hoodman Steel seems to perform identically to Lexar’s 1000x cards. I think they are pretty much a Lexar reseller or using the same chipset. Pricing is also very similar.

Transcend 1000x

32GB – $131 – Buy It Now

The 32GB card is right up there on the benchmarks for a low price but the KomputerBay card offers double the capacity for less money. Considering they are both El Cheapo brands that is a big difference. Another big discrepancy with the best is that Transcend only guarantee sustained video rates at 20MB/s whilst Toshiba guarantee it at 65MB/s on their 1066x – though it doesn’t seem to have much of a bearing on the pretty nifty benchmark results.

What about attaching an SSD to the 5D Mark III?

There is actually a way to connect an SSD drive to the 5D Mark III’s card slot, but it is currently untested. The adapter in question is the Sycard CF Extend 165. This goes from the CF card slot on your camera to a PCIExpress interface with ExpressCard slot on the end. However as it is a very niche hardware developer’s tool, I’m having a hard time finding one in Europe.

Various options exist for ExpressCard slots. There are ExpressCard SSDs like this one which top out around 256GB, though whether it will be any faster than the best CF cards remains to be seen. There’s also the question of power draw and whether the CF card slot can provide enough power to the SSD.

You can even attach the super fast 2.5″ SSDs for PCs and Macs to the ExpressCard slot via an E-Sata adapter and caddy, though I very much doubt the CF card slot on the 5D Mark III can power it all. You’d very likely need an external power source for that and a E-Sata adapter that would allow you to connect one.

Watch this space

The post Which Compact Flash cards for 5D Mark III raw video? appeared first on EOSHD.com.

Download the sample DNG frames from continuous raw video on the 600D

Though most of the development at Magic Lantern has been focussed on the powerful 5D Mark III, the latest firmware builds have also been ported to the 600D. How does it perform for raw video?

Click the image above to see the full 1440x720p version

The 600D is probably best compared to the Blackmagic Pocket Cinema Camera. Like that camera it records to SD cards and is extremely affordable. Currently it has a Super 16mm crop in the most reliable current raw recording resolution.

The card speeds achieved are much lower than the 5D Mark III. Therefore it does not yet record 1080p raw on the camera. You have to drop to the Super 16mm crop of the Super 35mm sized sensor, which gives you very nice albiet low resolution raw video at 960×480. On my Sandisk 95MB/s card I only get 19MB/s write speeds but that is enough for 960×480 at 24fps (17MB/s) which upscales very nicely to 1440x720p.

If write speeds can be improved in future updates, then the maximum resolution is 1740×1160 and Super 35mm and that uses the full sensor. Essentially 1080p but in 3:2 aspect ratio. That will be amazing if they get it working. Big if though!

Here are some original 960×480 frames converted from DNG – just a straight convert with Raw2DNG and no upscaling. (Click them to see at 960×480)

The low resolution does however have the advantage of being MUCH easier on card space, much more like ProRes in terms of file sizes whilst still giving you the sharpness, dynamic range, beautiful colour and overall lovely image quality of 14bit uncompressed raw – as well as the creative freedom to change the in post and forget about most of the camera settings whilst shooting. Just set shutter angle and aperture, and shoot raw.

Moire and aliasing are improved from the rather poor H.264 Canon video mode but there’s some false detail on very fine details similar to the Blackmagic Cinema Camera with red blue & green error pixels. On the shot below you can see some moire too on the metal flight case.

The sensor in the 600D does not sample the RGB image data quite as cleanly as the 5D Mark III.

Here’s another upscaling example. Click the images of course to expand them…

1920 x 960 (bit soft at this level)

1280×640

The original 960×480

The camera gives you crop marks on the screen as you shoot so you can frame the shot, but it would be nice in future updates to use the whole LCD. Remember this is still very early development code.

Right now the 5D Mark III is definitely worth the money in terms of the upgrade from the 600D if you want to shoot raw video. However it will be interesting to see how the 600D improves as Magic Lantern develop the camera further.

If you want to more accurately get an idea of image quality do feel free to download all my DNGs from this page in a zip file from Dropbox.

You can open them in Photoshop and have a play. It isn’t a patch on the 5D Mark III’s quality but still great to have and a leap from the normal 600D video mode.

Viva the revolution!

The post A look at raw video on the Canon 600D appeared first on EOSHD.com.

This was a great week. It is the first time we’ve seen this kind of moving image from a full frame sized sensor in raw. A pristine 2K 14bit image that’s like a 24fps digital film negative or a digital scan of VistaVision. It’s also a breakthrough for accessibility bringing raw video to many talents for the first time, when the only other accessible raw shooting cameras out there have been in extremely limited supply.

The news of 5D Mark III raw video (and the huge quantum leap in image quality that has given us on DSLRs) has generated the largest and most enthusiastic response I’ve ever seen on EOSHD. Even though raw video and development code are relatively arcane subjects, the subject garnered mass audience attention with tech blog Gizmodo picking up the story and digital photography authority DPReview running it as headline news. By the 5th day the quickly shot video posted late on Sunday evening had totalled 487,000 impressions and 98,000 plays, and mine was by no means the only one out there! I’m amazed the Magic Lantern forum didn’t melt and they handled the eager questions of 5D Mark III users very calmly.

The only cameras that have a similar aesthetic at the moment are the Red Epic with Dragon sensor and the Arri Alexa Studio in 4:3 with anamorphic glass. The Cinema EOS range is all 8bit aside from the raw tap on the C500, but then that isn’t $3000 and isn’t full frame and the Blackmagic Cinema Camera has a much smaller sensor, and therefore a very different rendering of the lens.

The 5D Mark III with Magic Lantern really is unique and a new image quality benchmark under $20,000. Perhaps the Canon 1D C comes close in feel but it only shoots 8bit MJPEG so cannot be bossed around in post with the same degree of creative control.

This is the biggest single leap in image quality for DSLRs since the advent of DSLR video and as significant as the original introduction of the 5D Mark II back in 2008/09.

Creative control

Like everything creative, post production is time consuming but I love and need the control raw gives you. Raw on the 5D Mark III is a creative dream.

It’s also viewed as an unnecessary potential nightmare for pros working to tight deadlines.

Many of the professionals who used to be happy with workarounds for better image quality, be it Super 35mm adapters or DSLRs, are now shooting comfortably with Cinema EOS cameras. They were far more muted in their excitement, preferring to highlight some of the practical issues around shooting raw rather than mentioning the image quality and creative possibilities of the hack.

In some filmmaking environments a camera is like a hammer, a tool which gets out of the way and gets a job done efficiently and effectively. But aside from the practicalities of getting done a job to spec and to a time frame, think about all that artistry inherent in using a camera and delivering a film. We all want convenience but for me, the camera is more than tool, it is emotional and intuitive, not just functional and technical. In the same way that film stocks give a certain feel, the same way even a colour temperature does – it’s all part of our emotional palette as humans.

We have a unique image here and we should cherish it. It’s even free.

Price and accessibility being obvious reasons why to get excited about this but creatively it matters the most.

Camera aesthetics are linked directly to emotional reaction. Style is a subliminal trigger for our thoughts as we watch a movie. It matters.

Vincent Laforet’s thoughts on 5D Mark III raw

Vincent Laforet is a photographer and filmmaker with a high profile track record in using DSLRs at the high end of video and movie production.

I asked his thoughts about the recent developments -

Vincent: Too little too late in my opinion. C Series and other cameras are what people should be focusing on… but I am looking into it to be honest… the RAW (sic) is cool – just the HDSLR limitations are still pretty severe relative to other cameras and production needs.

EOSHD: What do you think the biggest hindrance is – Hack reliability? File sizes?

Vincent: Hack reliability – and can the body TRULY survive doing this over time (heat/damage?)

I think Vincent will change his mind further down the line when the raw recording module is more developed, the workflow better and the camera totally reliable. I actually think high end productions will find a lot to like about raw on DSLRs too. The shortcomings of DSLRs didn’t stop Hollywood shooting reams of feature film scenes on DSLRs. There’s still stealth factor (Natalie Portman and Aronofsky on the New York metro), POV and helmet cameras, crash and vehicle cameras and the full frame look for framing faces so ably put to use on House in the hands of Gale Tattersall. To have all that in Alexa rivalling raw is ground breaking. New compact gyro stabilised rigs like MoVI work better with a small DSLR than an Alexa or fully kitted out Epic, so there’s another reason why a high end production team could and should embrace the 5D Mark III with raw not to mention the fact that the image just looks nicer than the Cinema EOS C300. Raw from a full frame sensor as good as the one in the 5D Mark III is just flat out better looking than compressed 8bit MPEG!

When Vincent shot Reverie that image was considered cutting edge, now it is superseded by the C300, 1D C, etc. The raw image from the 5D Mark III supersedes even those in some ways. That is exciting.

Blackmagic

When Blackmagic first announced their Cinema Camera, and again this year at NAB with the Pocket Cinema Camera and 4K Production model, they generated a similarly excited response at NAB many times larger than the usual camera launch.

Raw on the 5D Mark III is in many ways even bigger news because the image is better and the design of the camera is lighter, smaller and it happens to do superb stills as well. Pricing is just as attractive as the Blackmagic at around $3000 body only.

In this situation buyers look to DSLR community leaders to advise them. One’s response to this news is more muted than his reaction to Blackmagic and both raw-related reactions were more muted to the fan fair given to the $6000 Canon 1D X. The blogger admits to loving a full frame image and that on the 1D X for your extra $3000 you get a slight sharpness improvement over the 5D Mark III. Now the hack has taken the full frame image of the 5D Mark III to a whole new realm and way beyond the 1D X, there’s not the same enthusiasm from him for it. I’m baffled by this, and similar views from others, who have pigeonholed this groundbreaking event as a hobbyist thing. C’mon! Raw is the future. Storage space and media capacities are improving so quickly that they have given us a glimpse of this future already.

Similar to me with my love of raw on the BMCC and 5D Mark III, he’s now a Canon 1D C owner for the image,and the image alone, not the headaches of actually shooting with it. As I said in my review of the 1D C it has many practical and ergonomic issues due to being a DSLR with huge file sizes on compact flash cards, but the image is beautiful. The blogger in question happily endorses such an expensive camera based on the superior image, but not the arguably superior raw shooting 5D Mark III for $3000!  This is inconsistent.

Magic Lantern

Magic Lantern have given Canon customers a lot of superb functionality over the years for free and I believe they should be richly rewarded for doing so. They are an incredibly skilful team of developers, with unique knowledge. To call Magic Lantern a ‘hack’ is not accurate. It is as sophisticated as a complex software package. What we have here is not a hack as we know it with the GH2, where encryption was broken and existing parameters simply changed from within. We have a bespoke firmware for advanced filmmakers and photographers.

Without any inside knowledge they have taken an existing product aimed towards a wide audience and improved on it to make it more tailored to the needs of advanced creatives and pros. Sales of the 5D Mark III will go up as a result, but the Magic Lantern team probably won’t see a cent of it unfortunately. I hope this can change.

“Story matters”

In the 1970′s Pink Floyd were locked in a room for years experimenting, shaping and discovering as they went along a completely new sound, the sound that would later shape Dark Side of The Moon. I think young cinematographers are doing the same with raw on the 5D Mark III and Blackmagic Cinema Camera, shaping their ‘feel’ to images and with it filmmaking itself.

But whilst the soundscapes were taking shape someone keeps shouting at the band as they play – “but what about the lyrics!? That’s all that matters!!”

I’m seeing this more and more and very rarely does a leap in image quality go by – be it Red, full frame, raw, 4K or even laser projection without the cry of “none of it matters”. This is ridiculous of course it matters.

Without technological progress we’d all be shooting only in black and white. Technology broadens your creative scope. Black and white is beautiful but sometimes it isn’t what you want to shoot with.

There’s too much emphasis on storytelling as the be all and end all of filmmaking. Again the music analogy holds up. Many songs have lyrics but instrumentals and orchestral scores are still emotionally charged without any. Furthermore some of the best lyrics aren’t narrative, they’re allegorical, and the same goes for films.

There’s not one aspect of a film that matters most, everything does. A good story is unconvincing without great actors. Great acting is let down if the cinematography doesn’t also convey the same strength of feeling. Good writing cannot hide poor audio and bad music cannot ascend to the realms of greatness simply by adding better lyrics. Everything matters. Everything. It is not possible to appreciate one aspect in isolation if it is not in isolation!

Arri and 4K

4K is only one of 15 things that makes up a cinematic looking image and Arri recently said they’re holding off on a 4K update to the Alexa as 2K upscales so brilliantly. So it does – I’ve seen this done on prototype Sony hardware and it works amazingly.

However, to hold back advancing one of those ingredients because of the current status quo is a poor future product strategy.

Steve Jobs (and his quotes seem to be very much en vogue at the moment, like a mythical deity of some kind rather than the creative and acute businessman he was) used a sporting analogy in that he always tried to move ahead to where the ball was going rather than follow where it has just been, the claim being he was able to catch the zeitgeist ahead of his rivals. Apple certainly did that under Steve Jobs, and comforting though it is to think the tools you have now will be cutting edge for years, one day the 2K Alexa will be tired and old and obsolete.

Hopefully raw on the 5D Mark III will be embraced by pros and will give Canon pause for thought on how conservatively they have pushed their technology recently.

Always go to where the ball is going.

Right now the ball is in raw’s court.

The post The impact of 5D Mark III raw video and what does Vincent Laforet think? appeared first on EOSHD.com.

The Speed Booster for Micro Four Thirds gives cameras like the Panasonic GH2, GH3 and Blackmagic Cinema Camera a Super 35mm / APS-C sized sensor using clever optics, in the same way that the E-mount version turned cameras like the Sony FS100 and NEX 7 into valid alternatives to full frame.

As well as shrinking the image circle of a full frame lens to fit a smaller sensor, the Speed Booster lives up to its name by giving you a 1 stop increase in brightness so that F2.8 effectively becomes F2.0.

I’ve been testing the Leica R mount version of the new Micro Four Thirds Speed Booster on the Panasonic GH3 and here are some of my early observations.

First I noticed the glass is totally different. This is not just the E-mount version in a different shell.

The optics are new.

The specs however are similar – it aims to do the same thing. 0.71x compression of the image circle into a smaller sensor than the lens was designed for. Unlike the E-Mount / NEX version, on the Micro Four Thirds version of the Speed Booster you can use APS-C glass as well as full frame, though of course all Leica R glass is designed for full frame.

The Leica R version is of course passive mount, no electronics – as is the Contax Yashica version which I also hope to test with some of my fast F1.4 Contax Yashica glass. There’s no Smart EF mount electronics yet from Conorus though I believe one is still planned for release later in the year. This will finally give you full control of your modern Canon glass on Micro Four Thirds cameras with an active mount.

There’s now a much needed grub screw in place to secure the optics once you’ve rotated it in the internal thread to find infinity.

My adapter is a beta version for testing, it was far off infinity when it reached me and finding infinity yourself is quite time consuming – a process of trial and error and lots of tiny adjustments, but it is relatively straight forward – just a case of rotating the rear element in small steps.

I have a collection of Leica R glass which I am testing on the adapter -

• 28mm F2.8 Elmarit-R
• 35mm F2.8 Elmarit-R
• 50mm F2.0 Summicron-R
• 90mm F2.8 Elmarit-R
• 28-70mm F3.5-F4.5 Vario-Elmar-R

The zoom in particular becomes a very nice F2.5 – F3.5! The 50mm F2.0 becomes F1.4 and the F2.8 glass goes to F2.0 with the Speed Booster.

All this glass was procured pretty cheaply, it pays to shop around second hand stores and eBay for bargains. The 50mm, 35mm and 90mm cost 300 euros altogether. Some people are really undervaluing the glass. Leica R was discontinued ages ago but the optics are superbly sharp.

Of course Speed Booster increases sharpness in the centre of the frame but not at the edges.

Centre frame on my early tests with the Leica 28mm F2.8, wide open looks like F5.6… Stunningly sharp! A lot of that is down to Leica but some gain also is thanks to the Speed Booster glass.

The corners however are pretty soft and stopping down to F5.6 doesn’t improve them a great deal, so for landscape shots at infinity you’re looking at stopping down to F8 to get acceptable sharpness across the frame.

Left is centre, right is corner…

In terms of the exposure speed boost, the adapter does give the smaller Micro Four Thirds a real helping hand and will be great for GH2 and GH3 users.

When my Blackmagic Cinema Camera MFT finally does arrive (where on earth!?) it will boost low light performance on that too and give it effectively a Super 35mm sensor, albeit with rather soft corners.

Here’s the Leica 28mm F2.8 with Speed Booster (F2.0 effective) -

Here’s without (same aperture) -

Here’s the Samyang 35mm F1.4 at F2.0 (seems slightly brighter, better T-stop than the Leica with Speed Booster?)

Next up let’s see how the GH3 compares to Super 35mm.

The NEX 7 has a APS-C sensor, the same approximate 1.5x crop over full frame that Super 35mm is, although really as a cinema standard Super 35mm is not thought of as a ‘crop sensor’. Micro Four Thirds is a 2x crop but with the Speed Booster that becomes more like 1.5x. Not full frame, but much more in line with the cinema standard and the field of view / glass selection cinematographers are most used to.

As you can see on the GH3′s 2x crop sensor the Leica 28mm with Speed Booster very closely matches the field of view of 28mm on the NEX 7 with no speed booster (1.5x crop APS-C sensor). Maybe even a tad wider, and don’t forget the gain in low light!

Here’s some more sample images (shot in stills mode) on the Panasonic GH3 with Speed Booster and Leica 28mm F2.8 -

I really love the feel of the lens, and you have more control over depth of field.

For a 28mm F2.8 on a 2x crop sensor, you now get a pretty nice shallow depth of field.

Here’s one shot with the Leica R 80mm F2.8. It’s beautifully sharp though focus is pretty tricky wide open -

Here are the exposed Speed Booster optics from the front.

Looking down the barrel of the adapter with your naked eye, the sensor appears Super 35mm in size.

Build quality, like the E-mount version of Speed Booster is excellent, if not actually even better this time out.

I hope this gives you an early preview and an idea of what the Speed Booster for Micro Four Thirds is capable of. I’m still discovering it myself as I go.

I am working on a full review which sees how the adapter improves the GH3 in low light.

It could be said that Micro Four Thirds needs the Speed Booster more than any other camera system but actually in terms of field of view, I have got used to matching my lenses pretty well with the 2x crop sensor in the GH3 (and in the GH2 it is a 1.86x crop). What the Speed Booster does do is give you some more super wide angle options other than the Lumix 7-14mm F4. APS-C glass like the Tokina 11-16mm F2.8 for example will be nice on the Canon mount version.

Full frame glass is kind of in the middle – a 28mm is still not wide angle like it is on the E-mount version of the Speed Booster, but it still does give you a very valuable extra stop of low light performance and quite often a more atmospheric look to the lens, as well as giving you a more standard Super 35mm format via a neat optical trick.

The post Metabones Speed Booster for Micro Four Thirds – First Look and GH3 images appeared first on EOSHD.com.

First the news. Magic Lantern have playback of raw files working with early stage code. Black and white and slow frame rate but works – already. Experimental audio recording has been added. Testers and early adopters will need to see if it in sync.

Future nightly builds of the Magic Lantern 5D Mark III Alpha firmware will see raw recording start / stop added to the physical movie record button as well as many other usability improvements.

KomputerBay cards are giving different write speeds from the same 1000x models. My 64GB card is fastest with a sustained write speed of 95MB/s, but my 128GB card is a dog, topping out at 73MB/s – not enough for reliable 1920×1080 raw let alone 1920×1280 for anamorphic. May have to return it to Amazon.

Much more testing needs to be done to find the best cards. The KomputerBay 64GB 1000x has been superb for me, so it is a surprise the 128GB didn’t live up to expectations.

Buffer overflow and dropped frames are handled differently now. You can choose to have recording stop, or to have recording drop frames. Once the buffer is full, usually the camera continues to drop frames until you stop it, so I recommend the former option (stop altogether upon first dropped frame).

4GB file size limit

The 4GB file size limit of FAT32 prevents continuous takes over 40 seconds currently. I asked Alex of Magic Lantern about overcoming this and he says it is perfectly possible but not a current priority. There’s a lot that comes before it – even basic stuff like assigning raw recording to a physical button. They’ve done a lot of the hard stuff and I believe seamless file spanning will come, please be patient as it is still early days.

The need for speed

There’s plenty of room for improvement in compact flash card write speeds. Whilst read speeds are generally higher (sustained 130MB/s + on 1000x speed cards, with a theoretical limit of around 150MB/s), write speeds are always much lower than the stated 1000x 150MB/s spec. What resolutions could we see in raw with future improved CF cards or other storage solutions?

Here’s a guide…

• Anamorphic 3:2 @ 1920×1280 /24p = 98.3MB/s
• 30p @ 1080 = 103.6MB/s
• 48p HFR @ 1928 x 672 = 103.4MB/s
• 60p @ 1928 x 672 = 129.8MB/s
• Blackmagic style 2.5K raw @ 2560 x 1320 /24p = 135.3MB/s
• Full 3.5k Super 35 crop mode @ 3592 x 1320 / 24p = 189.9MB/s

At the moment all the above will record but after a few seconds the buffer becomes full if the card can’t keep up. The buffer is 700MB/s and a much faster sort of RAM memory. Hopefully they will put SSD NAND chips in a CF card and bust the CF speed barrier altogether. I also hold out hope for a breakthrough in SD card technology as the camera can of course also uses those. The other option would be to write to an SSD via the USB port, if the master drive can be switched to that from the card slot. Unfortunately unlike the USB 3.0 equipped Nikon D800, Canon chose to cut costs and only included a USB 2.0 port on the 5D Mark III. Madness! USB 2.0 is only capable of 35MB/s. HDMI meanwhile offers data transfer rates of a whopping 1GB/s, but the HDMI controller processes the raw data first, adding stuff to it resulting in a garbled output. Maybe the Odyssey 7Q external recorder could be a future option if they figure out the HDMI port.

Blackmagic Cinema Camera vs 5D Mark III in low light

One of the biggies for me is low light, as this is an area where the only other current affordable raw camera, the BMCC falls down. Whilst it isn’t exactly bad in low light, the 5D Mark III with its full frame sensor should be able to up the game significantly.

Just for the record I love the Blackmagic Cinema Camera and have no plans to replace it or cancel any pre-orders.

However the 5D Mark III blows it away in low light. This is ISO 1600 -

The 5D Mark III raw video at this level of ISO (1600) is pretty much free of noise. It looks like ISO 200 on some cameras!

The test is shot wide open at F1.4 so isn’t a sharpness test. However something quite noticeable struck me whilst producing the above image – I upscaled the 1080p raw from the 5D Mark III to match the size of the Blackmagic’s 2.5K raw and there’s hardly any noticeable difference in detail levels.

What is more, the 5D Mark III produces a cleaner moire free image than the Blackmagic Cinema Camera.

Download the original DNGs here

Photoshop notes – here’s how I graded for ISO 1600 on the 5D Mark III in raw:

Do the same for the Blackmagic DNG but set exposure to 0 (centre on the slider). Note the Temperature and Tint sliders especially – the 5D Mark III has these wildly out in the raw metadata when you first open the image. I expect this to be a formality fix in Magic Lantern, given due course.

Here’s the unscaled 1:1 crop from the DNGs…

Push the exposure past ISO 1600 and onwards towards ISO 6400 and the difference is even more astounding…

Image quality better than wildest dreams?

I truly believe the 5D Mark III is unique on the market with the Magic Lantern raw recording module.

It is giving an Alexa / Epic comparable image at 2K, whilst probably out performing both those cameras in terms of low light sensitivity and noise.

I cannot, frankly, believe my eyes. This is unbelievable stuff I’m seeing.

Opening the CR2 raw file from the 22MP stills mode, and downscaling it to 1080p reveals the same characteristics. So think of 5D Mark III raw video as a moving DSLR raw still photo, and you can begin to get an idea of how mind blowing the quality is.

The post New 5D Raw developments – plus my low light comparison with Blackmagic Cinema Camera appeared first on EOSHD.com.

Click here to download this guide as a PDF document

Here’s a quick and easy way to get raw recording setup on your 5D Mark III thanks to the recent Magic Lantern developments.

1. Prepare your 5D Mark III

Charge the battery fully.

Set the camera to movie mode on the live view dial and set the mode dial to M. I also recommend assigning 5x focus magnification to the SET button.

Attach a lens. DSLRs don’t work without a lens

Tip: if you’re running the new uncompressed HDMI firmware from Canon (version 1.2.1), download and install the older version 1.1.3 instead. Both Mac and Windows firmware archives are in the zip. Unzip the firmware files and copy to your card, then go into the camera menus and select Firmware Update.

2. Download

Download the zip on this page (nightly 5D Mark III Magic Lantern builds)

Unzip it (double click) and copy all the files to the root directory of your card.

• Windows users: Download EOScard.exe and raw2dng.exe to a folder on your computer.
• Mac users: Download macbook.class / macboot.command and Raw2DNG (Mac) to a folder on your computer.

Above: KomputerBay – the Comic Sans of Compact Flash cards

3. Prepare your Compact Flash card

For 1080p will need a 1000x speed card, preferably 64GB as some of the higher capacity 128GB cards currently have issues with lower than expected write speeds.

Mac users should use Mac Boot to prepare the card, whilst Windows users can use EOS Card. You just downloaded these in step 2.

Windows users: Open the EOS Card app and choose your compact flash card from the drop down box (it must be in your PC / USB card reader) click the Magic Lantern logo and click Save. Done.

Mac users: If you have OSX Mountain Lion first double click the macboot.command shell to run it. This must be in same folder as the Mac Boot class app. Then you can run the app. Select card drive in drop down box at the top, click Make DSLR Bootable and then Prepare Card.

Tip: EOS Card can also install one of the existing 600D / 5D Mark II Magic Lantern releases on the card. If it does, delete those files and repeat the part of step 2 where you copy your downloaded Magic Lantern files to the card.

4. Record raw video

Put the card in your camera and go to the Canon main menus. Select Firmware Update and follow the commands on screen. From now on your camera will automatically use the Magic Lantern firmware on the card. (To go back to the normal camera firmware you simply turn off the camera and use a normal SD or CF card).

Enter the Magic Lantern menu with the delete key on the camera (it has a trash can on it). The top dial navigates the tabs and the rear scroll wheel and joystick navigate the menu options. The Q button goes into a sub-menu and back out again.

• You can choose your resolution and start the recording in the Raw Video menu of Magic Lantern
• You can stop the recording by pressing the Playback button (or by going back into the raw menu)
• Copy the raw clips from the DCIM folder on the card to your PC or Mac
• Drag and drop your 5D Mark III raw files into the Raw2DNG app to convert to DNG format
• Be amazed

Click here to view the embedded video.

After Effects Workflow

First step is the Raw2DNG app that you downloaded in the zip. This app too, like the raw recording module on the camera, is hot off the lab and very very early code! This is required to make your raw files compatible with mainstream Adobe, Avid, Apple and Blackmagic post production software. You drag and drop the raw clips into it, and out come DNG raw image sequences like the Blackmagic Cinema Camera. One important difference is that these are standard DNG image sequences, no Cinema DNG metadata. You can’t yet put them straight into Resolve unfortunately.

Once in DNG format you can load individual frames in Photoshop for inspection or load the folder of frames in After Effects to render it out as a sequence. After Effects is how I am doing it at the moment. Drag each folder of DNG clips that Raw2DNG created on your drive into After Effects.

In After Effects you can render the DNG sequence in any codec you like – be it CineForm, ProRes or space saving H.264.

After Effects allows you to grade the first frame using standard Photoshop camera raw controls. That grade is applied to the whole sequence. In the future, using Resolve would be preferable (like Blackmagic) but a stills DNG to Cinema DNG utility needs to be made first for that to happen.

Summary

Of course in the future the raw video feature of Magic Lantern will have more options. Some of the functionality is still in the Canon side of the firmware. For example to activate crop mode and higher resolutions than 1920×1280 you need to enter 5x magnification focus assist mode. I currently have this assigned to the SET button. To access higher frame rates than 30p you need to set the video mode to 720p in the Canon menus rather than 1080p, and use the FPS Override menu in Magic Lantern to choose your frame rate. 24p and 25p work most reliably.

The post Easier 5D Mark III raw guide in 4 steps appeared first on EOSHD.com.

The 5D Mark III is already a very capable low light camera. In its factory guise whilst not quite as clean as the Canon C300 or Sony FS100, it is the best DSLR for low light shooting (though the Nikon D5200 puts up a good fight).

But that was before the latest developments with ML Raw.

Has low light improved even further?

The answer is a resounding yes.

The image is cleaner in low light with raw, with even exposures I rated ISO 12,800 coming out perfectly usable. I’m hesitant to say it as a matter of fact but based on my early observations the 5D Mark III seems to have surpassed the C300 as the ‘low light king’.

The power of raw is that you can fully exploit the dynamic range of the image, and in low light that means going into the shadows and giving them a lift, or crushing them back for more contrast and less noise.

First an important note – don’t judge noise from the Vimeo streaming embedded clip. The heavy compression on the website makes it seem MUCH noiser and blockier than the raw actually is. Download the more lightly compressed file or even better take a look at the ISO 1600 shot below. So clean.

RAW (click it to enlarge)

Download the DNG and grade it yourself in Photoshop

Note on warm cast – that is on purpose, creative choice to match look of scene how my eyes saw it. You can apply cooler white balance to the raw image if you download the DNG.

H.264

Here for comparison is the H.264 video mode 1080p (white balance on automatic). Look at the lovely bulb in the old German table light – vanished!

Also look at the poster – half of it is gone.

This is why I am never going to shoot a serious project in H.264 again, until they give us a better image which more closely resembles raw or ProRes.

The standard compressed video mode is totally washed out in terms of colour as well, at ISO 1600. Dynamic range is shot to bits. I’d say only 6 stops usable in it! Why is it so bad? What are Canon doing to the video image to get from the raw sensor feed to this? Compression alone is not responsible. The highlights are burnt – gone – unrecoverable. Any attempt at recovery leads to horrible banding as well. The blacks are crushed too compared to the raw image.

Latitude

The latitude in the raw shots around the light sources impressed me hugely – in particular watch for the candle flame in the ISO 12,800 H.264 shot.

The whole lantern area is burnt out in the H.264 file but the raw retains everything.

Raw

H.264

And look at how much more detailed the raw shot is!!

Now actually H.264 is a great codec. The screen grabs you see above were taken from the 24Mbit H.264 Vimeo download file and compressed even further to JPEG for the web. It can handle plenty of image quality despite being compressed, as can a JPEG.

I think serious questions need to be asked of why Canon had to hack the image down so much.

The dynamic range is severely clipped from what it could be, and so is resolution. I don’t understand why compression alone has to hurt the image so much, and it is not due to downscaling or the sensor – as the raw feed is already 1920×1080 (when cropped from the original 3:2). There’s no need for further scaling before it goes to the encoder. No need to chop 4 stops dynamic range off it either.

Exposure

In that candle shot you’d be forgiven for thinking the exposure isn’t ISO 12,800 if you only notice the retained highlights, which are no longer over exposed – but if you look at the mid-tones (poster on the wall and the blinds) and shadows (sofa area) you will see I have done my best to match the exposure to the H.264 ISO 12,800 shot. To get those lows and mids in the raw shot, you’d have to shoot ISO 8000 to 12,800 on a normal H.264 camera – and your latitude would be tiny, overall dynamic range probably around 6 stops at ISO 12,800.

Colour cast

I notice that noise in the shadows and blacks in general have a magenta cast if you have a warmer white balance and the magenta tint slider to the right but if you put the magenta tint slider back towards green, the warmer yellows go way too green for my liking. This could be a light issue rather than a camera one. Energy saving practical lights do give a green cast. The solution in post was to avoid anything higher than 3200k white balance – this was more keeping with the LED practical light I had for fill lighting in the scene any way.

Lighting

In the ISO 800 shot I have the table lamp on, I purposefully turned this off for the higher ISO shots to lower the ambient light. A genuine low light test should be done under dim lighting with lots of dark areas, in effect it should be under exposed. Uniformly bright studio lights and stopping the lens down for a normal exposure is not really a good low light test in my view, you need those under exposed tones in the shot!

Better than Blackmagic Cinema Camera?

The 5D Mark III with raw is basically a totally different camera to the factory 5D Mark III and there’s a lot to get our heads around and explore. This is an early test and there’s much more to come. The pace of the Magic Lantern development lately has been astonishing. I have my Blackmagic Cinema Camera sat next to the 5D Mark III and it is going to be very interesting to compare. Already the 5D Mark III has a few advantages on the specs sheet

• Less noisy (in low light)
• Less noisy (no fan!)
• Small interchangeable battery and good run times on one charge considering tiny battery
• Full frame sensor
• 60fps raw (at 720p… In 1080p the data rate is too high for the CF card)
• 22MP stills
• Smaller and lighter form factor
• More physical controls

With that said, it is important I think not to rush to judgement on this as it is such early days. I you are thinking of cancelling your Blackmagic pre-order, I don’t recommend doing so – please wait until this shakes out. Of course Blackmagic will likely still have the resolution advantage on the 4K Production camera and global shutter is still groundbreaking for $4k. Lots to look forward to from them.

I think the 5D Mark Raw is more of an dilemma for Nikon D800 owners looking for video quality. Some of them would have moved to that camera for the gain, and now they will have to come back again big time.

The post Low light test of 5D Mark III raw vs H.264 appeared first on EOSHD.com.

UPDATE: here’s 2.5K raw in 1:1 crop mode -

Click here to view the embedded video.

Huge thanks go to A1ex, g3gg0, 1% and the whole Magic Lantern team.

I have no other words to describe it – this is huge news.

Magic Lantern have done the seemingly impossible and given us a continuous raw recording mode on the Canon 5D Mark III. Once activated in the menus the 5D Mark III becomes essentially a full frame Blackmagic Cinema Camera and amazingly mine has not yet exploded. No more short bursts of raw, this is the real thing.

The image is leaps and bounds ahead of any other DSLR. We’re talking Alexa / Red league here, yet full frame. The first ever raw video shooting full frame camera at that. I’m getting continuous recording at a data rate to the card of 90MB/s and the camera hardly breaks into a sweat.

Here’s the option in Magic Lantern:

I shot the clips in the video above in 1920×1280 mode. I also have some 3.6K clips (3592×1320) too unfortunately they are glitchy with some drop frames, but 3K or 2.5K might be doable, especially with a narrower vertical crop… 3000×1000 maybe. That is approximately Micro Four Thirds area of the sensor at 1:1. The quality is the same as a raw photo. 14 bit, 12 stops dynamic range.

The 1080p raw file sizes are similar to 2.5K on the Blackmagic Cinema Camera. Around 4MB per frame for 1920×1280 (great aspect ratio for anamorphic lenses, utilising a higher than full HD vertical res) or a much more space efficient 2MB for 1920×720 (2.66:1).

You can also record plain old boring 1920×1080!

Needless to say you will need to invest in plenty of 1000x 64GB or 128GB cards especially for higher resolutions such as 3.5K which is over 7MB per frame!

Currently most of the action is being done on the 5D Mark III, once a reference version is finalised it will likely be ported to the 6D, 5D Mark II and even 600D. This is development code not finalised, not even Alpha yet. This is experimental code, just written.

Is heat an issue? Well I don’t know yet – I haven’t shot with it for long enough or in a hot enough place. The card and camera are warmer than normal, but that is to be expected considering the huge data rates generated by raw. I don’t know what – if any – implications for camera life this hack will have, but the sensor is always doing a raw video output in live view mode so sensor heat shouldn’t be any more of an issue than it was with ALL-I H.264 video.

Here’s a full 1920×1280 frame grab from raw. No more soft mushy Canon DSLR quality!

Click the image to enlarge

Breakthrough

As far as my belief goes the breakthrough involved a deeper understanding of the camera’s EDMAC (External DMA controller) and taking clean crops of the image buffer. Unlike in the burst mode, raw data is not converted to DNG format in-camera, you have to transcode it to DNG in post. This is good because you can do a very high quality debayer of the image on a powerful PC. To my eye the image looks better than with the previous nightly builds via DNG burst mode.

Developers can now write utilities to convert the 5D’s raw video files to Cinema DNG compatible with Resolve (not just standard DNG) and the camera can store raw clips in a single ‘.raw’ file. Magic Lantern themselves currently provide a converter for Windows (raw2dng.exe), hence the camera is freed up to just dump the data from the image sensor direct to the card.

The converter can also output to MJPEG if you want to avoid the workflow and get a project out of the door quicker. Even when converted to MJPEG it is a huge step up in image quality from the factory H.264. The camera now has a raw video menu where you can choose your resolution, and even a panning mode (like a digital dolly) in the 1:1 crop formats. The panning can move the sensor window around the frame whilst recording! Higher resolutions than 1920×1280 are a 1:1 crop of the sensor, up to 3592×1320. Here’s a partial list of the currently supported resolutions:

Full frame mode -

• 1920 x 1280 (3:2) – great for anamorphic
• 1920 x 1080 (16:9)
• 1280 x 720 (16:9)
• 1280 x 1280 (1:1)

1:1 crop modes -

• 3592 x 1320 – I get some drop frames at this res
• 2880 x 1320
• 2560 x 1080
• 1920 x 1080 (full HD crop mode like GH2)
• 1280 x 720 (the lower the resolution the more telephoto the crop)

Custom resolutions can be added in the Magic Lantern code, so effectively the raw codec on the 5D Mark III is totally resolution independent and can record in any aspect ratio. 3:2 and 4:3 aspect ratios will be very useful for anamorphic shooters.

The raw format is 14bit linear uncompressed RGGB raw.

Workflow

Raw2dng.exe is currently very simple and runs at the command prompt in Windows, you specify the raw filename and away it goes. After that is done you simply import the DNGs as a sequence into After Effects, though there is a variety of other ways to do it not involving Adobe software. The good thing is that the camera records the raw data to ONE single file in-camera, making it easier to manage your clips on the card. There’s no playback of raw in-camera yet, but live view remains on whilst recording. I am sure a thumbnail reference can be created for clips later.

This continuous raw recording mode came about from a separate development branch of Magic Lantern to the very cool DNG burst stills silent pics. It started off as a module to record raw YUV 422 image data to the card. The burst mode is still in there if you need DNG straight out of the camera and that will remain as a stills feature.

Alex of Magic Lantern rewrote an entire module (lv_rec) to make this happen and you can take a look at some of the source code for the raw recording module here.

The world just exploded

Yes the world just turned on its head and I have no idea what the implications for the market are going to be yet, other than that now Canon’s DSLRs, through no action at all from Canon themselves just became a whole lot more serious as cinema cameras.

There’s no doubt the raw images from the 5D Mark III exceed the quality you get from Canon’s much more expensive cameras, like the C300. The 1D C now also has some serious competition from much cheaper DSLRs. I feel Blackmagic could lose most from this though, and will be a shame if that happens. I’ve not yet tested them head to head but the sensor in the 5D Mark III is far more capable than the one in the Blackmagic Cinema Camera in many ways – generating clean moire free images especially in 1:1 crop mode and cleaner high ISO images from a much larger sensor. The BMCC is no longer as attractive to me.

The 5D Mark III is now top dog for video quality, far surpassing the GH2 and GH3. All this is rather tragic, because quite simply, Canon don’t deserve this success. It is entirely of Magic Lantern’s making, and they did it for us free.

If you want to show your appreciation please donate now to the team!!

The post BIG NEWS – Hands on with CONTINUOUS raw recording on Canon 5D Mark III appeared first on EOSHD.com.

With the rise of J.J. Abrams, TV’s aristocracy has morphed into cinema’s with JJ Abrams to direct Star Wars Episode VII in the UK next year.

J.J. Abrams comes from a family of TV industry figures, his father a TV producer and his mother an executive producer. His sister is a screenwriter. Now with a string of successful TV shows behind him, he’s had a string of hits at the box office too.

For me J.J. Abrams skill is in producing popular escapism, which is exactly what Star Wars is.

Although his mainstream populist TV isn’t what turns me on, J.J. Abrams is another George Lucas. I wouldn’t go as far as saying he’s a filmmaker in the Spielberg mould. Capably written and crafted as Alias, Lost and Super 8 are as, there’s nothing yet in Abrams’s cannon to match Duel, Jaws or Schindler’s List.

Abrams is certainly multi-talented. A composer, writer, producer as well as director. Although J.J. Abrams won an Emmy for Lost and wrote the excellent Alias pilot episode, he also wrote Armageddon.

Lost started well but only held my attention for 3 or 4 episodes. Like Lucas he’s good at concepts, but sometimes short on substance and as a result, a bit forgettable.  I think Disney see him as a safe hand, someone who isn’t going to do anything unexpected or radical with the Star Wars franchise. That said, it is very hard to dislike the guy. Millions get their escapism in the yarns spun by Abrams.

Here’s a glimpse of his directing style -

Click here to view the embedded video.

His latest Star Trek film “Into Darkness” is out now and has met with mixed reviews, but what I’m more interested in is the deal Abrams recently made with Valve, to make films based on Half Life and Portal. Half Life in particularly had a superb narrative and atmosphere. The first game started with a train ride down into a research lab facility and it was incredibly tense. They later ruined it by adding too many monsters. I think Half Life could be incredible sci-fi if he gets someone like Duncan Jones (Moon, Source Code) to direct it.

When J.J. Abrams takes the reigns of the new Star Wars movie, he will be doing so in London, where the franchise is going back to its roots according to Lucasfilm (now Disney owned) – “All of the six previous live-action Star Wars movies have included UK production in such famed studios as Elstree, Shepperton, Leavesden, Ealing and Pinewood Studios. We’ve devoted serious time and attention to revisiting the origins of Star Wars as inspiration for our process on the new movie, and I’m thrilled that returning to the UK for production and utilizing the incredible talent there can be a part of that.”

Torches and anamorphic lenses at the ready sir!

The post Filmmaking tips from J.J. Abrams – plus is he actually any good? appeared first on EOSHD.com.

Beautifully shot indie film Upstream Color debuted at Sundance in January, winning the special jury prize. Since then it has a wide release at cinemas.

The film was shot on the Panasonic GH2 with hack, Voigtlander lenses and the Samyang / Rokinon 85mm F1.4. (The technical page at IMdb confirms)

Here’s the theatrical trailer:

Click here to view the embedded video.

The film is also now available to download / stream from various places, not least of all straight from the director. If you want to wait to see the GH2 perform in a proper theatre (like me) I recommend seeing it in a cinema. The theatrical run started in April 2013 in the US and is still ongoing through May. You can see the latest screenings here. In August / September the film was also begin a theatrical run in the UK.

People say it looks glorious on the big screen.

The film has also been met with a very enthusiastic response by critics, with the Boston Globe calling it a “hypnotic thriller” and the Hollywood Reporter saying the film is worth seeing for “brilliant technique, expressive editing, oblique storytelling and for discovering a significant new actress”. The film scores 80/100 at Metacritic and 87% at Rotten Tomatoes (91% for audience).

The Telegraph’s critic said “my immediate desire when it ended was to stay in my seat and watch it all the way through again.”

The film truly makes use of the unique point of view made possible with such a small camera, placing it in all kinds of places you wouldn’t normally put a heavy Alexa (like on a car dash or in a swimming pool).

And aside from the GH2, the film has another neat trick.

Wrote, directed, starred and scored by Shane Carruth, the plot is allegorical and elliptical. Those words above are from Stanley Kubrick and like Kubrick, Carruth has understood that allowing the audience to unravel a mystery and to be confused by it makes for a more compelling experience.

Here’s the GH2 spotted on set with the 85mm F1.4…

I highly recommend seeing this film, even if you are not interested in seeing how the GH2 performs on the cinema screen.

The post GH2 shot sci-fi Upstream Color breaks $300,000 mark at the US box office appeared first on EOSHD.com.

These past days I’ve been trying to make myself useful to Alex at Magic Lantern, testing the DNG raw video recording on the 5D Mark III with a nightly build of Magic Lantern.

Here’s what the video quality looks like in the DNG video modes at 30fps.

Click the image below to expand…

(Note: the 3.5K crop mode is downscaled to 1928 in this shot)

Image quality wise no matter what the resolution in DNG, it is a HUGE step up from the current H.264 video mode or HDMI in every way – sharpness, detail, dynamic range, grain, colour, roll off, smooth gradation, you name it.

I actually use the term 2K raw for brevity’s sake – the horizontal resolution is actually 1928 not 2048 and the aspect ratio is 3:2 not the cinema standard for 2K. Neither is it 1080p. To get the data rate low enough for the Compact Flash or SD card to handle vertical resolution needs to be cropped from 1285 pixels to 720, giving you an anamorphic style 2.66:1 aspect ratio – 1920 x 720.

It is still an actual resolution increase on the current video mode because the detail is so finely resolved in raw mode and so soft in the current H.264 1080p.

3.5K mode (a 1:1 crop of the sensor) is lovely, but the frame sizes are just too huge (9MB) for an SD card to handle at frame rates high enough for motion pictures so don’t hold your breath for it! Also as you can see from the video in this mode jello is a big issue with handheld footage!

As you can see from the framing in the picture above, 3.5K crop mode samples just over half the sensor horizontally and 1320 pixels vertically. There’s currently a black band on the far right in this mode but that can be chopped off in post.

With the Magic Lantern build I have and the 5D Mark III set to 720p in the movie mode menus, the camera outputs very clean 1928 x 670 raw frames too but my 670p is squashed vertically and need to be shot with an anamorphic lens attached to correct the distortion. The distortion is actually helpful in that respect, saves you having to apply a squeeze in post when shooting with the Iscorama. I believe they have 1920 x 720 working with no distortion but have yet to get this working on my version.

By turning the in-camera CR2 raw recording off and putting JPEG recording on Medium or Small, I can get up to 2 seconds of continuous raw recording on the 5D Mark III at 1928 x 670 and just over 1 second of 1928 x 1285. That is 50 frames of raw video for 1928 x 670 in a burst of up to 30fps and up to 30 frames at the full 1285p. As a timelapse or photographic burst mode it is extremely useful even in it’s early bleeding edge form, very much like the Nikon V2 mirrorless camera but lower resolution.

Above: click to enlarge the 3.5K sample frame (downsampled to 1920 and converted to JPEG)

The short recording time is limited by the size of the camera’s buffer. When the camera dumps the frames from the buffer to the card, it seems to debayer the raw frames for the image review causing the whole process to dramatically slow down. We need real-time recording of DNG to the card for all this to work as a video feature.

Magic Lantern have got the image review working which gives you a live preview of footage as they are slowly written to the card. This debayering is in my opinion very CPU intensive and I think they’d need to find a way to turn it off.

The debayering method used to produce the live view image and H.264 is less CPU intensive but of course results in a very poor image, the source of the soft video we have on the 5D Mark III. Then on top of that dodgy facsimile plenty of compression is added to create the H.264 MOV files. With the latest firmware (1.21) we now know what the compression does not really change image quality, as the uncompressed H.264 output via HDMI looks just as rubbish as the internally recorded compressed H.264 (at 24Mbit in IPB mode).

In theory if Magic Lantern can turn off debayering and simply continuously dump raw frames from the buffer at 24fps to the card, then we have continuously recording raw video and are not limited by the size of the buffer (512MB). The buffer also has to store other stuff, not just the DNG files hence the low number of frames it can store at once before it becomes full.

Writing raw video to the card is not processor intensive, but has a very high date rate.

So the next big hurdle is whether the card – and more importantly the card controller of the 5D Mark III – are fast enough to get that uncompressed DNG raw video out of the camera’s buffer in realtime.

My tests say that in theory yes they are fast enough!

Magic Lantern has an internal benchmark that writes a block of data to the card with the card controller. In the 5D Mark III I got sustained 75MB/s write speeds to my 1000x compact flash card which I used in the 1D C to record 4K MJPEG. Not only is the card capable of that rate, but the internal card controller seems to be too. This part was upgraded in the 5D Mark III to support newer UDMA 7 cards.

The 5D Mark III’s 1080p DNG raw would likely need to be cropped to an anamorphic aspect ratio. Either 1440×1080 4:3 for anamorphic lenses (my preference!), or 1928 x 720 for a faux anamorphic 2.66:1 which many people do in post anyway with 1080p.

Uncompressed 1920 x 1080 raw is around 120MB/s to the card, and on the Blackmagic Pocket Cinema Camera with 2:1 compression it is 60MB/s. Compressed raw doesn’t seem to be a realistic option at this point on the 5D though. The Blackmagic Pocket Cinema Camera does 2:1 compression on its DNG files but it does so with dedicated hardware for the task. Magic Lantern’s DNG module runs in software. It is C code on a general purpose processor. I very much doubt that it has the performance to do complex compression on 2K raw files at 24 frames per second, and if it can do that, your camera will probably explode.

Rather than compress 1080p an anamorphic style 2x crop to 1920 x 540 would produce raw video at 60MB/s, perfectly within the benchmark score for my 1000x compact flash card.

However rather than the camera’s card write LED flickering continuously as the DNG files are written to the card, it does a very brief flicker for each frame and then the camera thinks about something else for around 1 second per frame. With the Magic Lantern card benchmark, the LED flickers manically and continuously.

Could this delay be due to the camera debayering the raw file for image review on the LCD? Hopefully Magic Lantern can find out what the camera is thinking about, and either turn it off or dramatically speed it up. Whatever it is, it isn’t yet a continuous 60MB/s stream of raw video frames to the card, and needs to be if we are to get it to work for video.

As well as a faster card controller than the 5D Mark II the other major advantage of the 5D Mark III is that the raw image appears to be downsampled much more cleanly from the sensor. The DNG raw files have a lot of false detail evident on the 5D Mark II, whilst they are silky smooth and pin sharp on the 5D Mark III. Here’s an example of the difference (at 1:1):

As you can see, there’s a lot of red, blue and green errors on the 5D2 DNG raw with fine detail, it has a very scratchy look and now smooth at all. I put this down to the 5D Mark III doing 3×3 pixel binning on the sensor and the 5D Mark II line skipping over the sensor, producing more false detail – though some very good sources have told me it IS possible to get good quality with nearest neighbour downsampling – essentially the same as skipping lines 3:1 on the sensor to get 1080 lines from 3840 less a crop to 16:9.

The 3:2 aspect ratio 1928 x 1285 frames are 4.8MB each and appear to have around 11 stops of dynamic range. Detail is cleanly rendered and tack sharp, none of the same moire or aliasing issues we usually see. It has less moire than the Blackmagic Cinema Camera too, yet is much cleaner in low light and with a full frame sensor.

The 3.5K frames are activated when the sensor is put into focus assist mode. With the 5x magnification active on the live view display, the 5D’s sensor does a 1:1 crop of the full frame sensor at up to 30fps continuously to the buffer, measuring a huge 3592 x 1320. The aspect ratio is similar to shooting with a 2x anamorphic lens on a 16:9 camera sensor, and I have actually grown to like that look a lot.

Since it is a 1:1 crop the image quality in this mode has no scaling, so the image quality is the same as a raw CR2 photo viewed at 1:1.

The crop however comes from a very oddly positioned window on the full sensor, it isn’t centred. When the focus assist box is moved with the joystick from the far left to the far right, Alex has noticed that it pauses very briefly in the middle and speculates that the sensor changes the windowing mode from the left to the right. As it is, framing a shot in this mode is currently very difficult as the live view display only shows a very narrow (roughly VGA) crop of the much larger 3.5K sensor crop resulting in a very extreme magnification of the full 3.5K image frame, which was after all how it is designed to work (as a focus assist)!

Workflow

Thanks to an EOSHD forum member’s advice in the Nikon V1 60fps raw burst mode topic I was able to get a very quick and simple raw workflow with the 5D Mark III going.

Rather than use DaVinci Resolve (which only works with TIFF sequences and Cinema DNG, not plain photographic DNG sequences) I used After Effects to put the sequences of DNG frames together into video clips.

Ctrl-I brings up an import dialog box and you simply put each sequence of frames into it’s own folder on your drive, now they act like short clips. Choose the first DNG frame in the folder and After Effects will allow you to grade it in Photoshop, before applying the look to the rest of the frames in the sequence automatically as if it was a video clip.

Then I exported each composition to a lossless Quicktime file, and dropped that straight into Premiere. Of course there’s no audio recording to a separate WAV file yet with Magic Lantern so have the dual system audio recorders at the ready.

Summary

All in all the recent Magic Lantern developments have left my jaw on the floor, both with amazement and with bafflement at how Canon can leave out such valuable features. If this team of programmers who have had zero assistance from Canon and zero insider knowledge can put a DNG burst mode of 30fps raw into the 5D Mark III, you have to wonder why Canon didn’t put one in themselves and it is only natural for me to question this, as I am sure everyone else is doing!

Actually we shouldn’t be too harsh on Canon. Although they could have done it at the drop of a hat but didn’t, I don’t think there’s a conspiracy. I believe the issue lies with the economics of the camera industry and the consumer. There are drawbacks for the regular customer – support requests from people with slow SD cards would go through the roof and add to Canon’s overheads if they put the 30fps raw burst mode in there. Sensors might fail due to heat build up and the reliability of the camera might be effected long term resulting in more returns under warranty and the huge implications for Canon’s reputation that could bring. Jello is an issue  in the crop mode, though I’m surprised there’s not a 1080p H.264 crop mode since the 600D has it and the sensor on the 5D Mark III is clearly capable of it.

However as we have seen time and again, whether it is with the GH2 hack or Magic Lantern, we are being given a very small subset of features compared to what the hardware can offer and it really is time this situation was improved. Power to the people.

The post 3.5k Canon 5D Mark III raw video with Magic Lantern and latest updates appeared first on EOSHD.com.

Download my full resolution 6000×4000 photos from the Sony RX1

The Sony RX1 is a compact camera that beats Leica and Hasselblad for image quality.

$2800 may seem a lot for a “compact” but that is before you see the images.

The Sony RX1 truly offers image quality in the league of a Leica M Type 240 outfitted with Leica 35mm / f2.0 Summicron-M.

That kit costs $10,000.

It is hardly any bigger than the RX100 jeans pocketable compact and in some dimensions even smaller than the Canon G1 X and Fuji X100. The lens is barely any larger than the Olympus 45mm F1.8 for Micro Four Thirds yet feeds a sensor 4 times larger. It isn’t jeans pocketable like the RX100 but it is jacket pocketable. It is a miracle.

Image quality

This one is simple.

A) The Sony RX1 is a compact

B) It outscores both the Nikon D4 and Phase One IQ180 Digital Back on DXOMark.

Next consider -

The Phase One IQ180 is $44,000. It is medium format. Not only is the RX1 pocketable – it is $44,000 medium format pocketable.

Anyone still complaining about the $2800 price tag? No? OK I’ll continue…

The lens is better performing than the $1850 Zeiss Distagon T* 35mm F1.4, a flagship 35mm lens Zeiss only recently introduced.

If the base optics are superb then they become nigh on perfect when you turn on in-camera digital compensation. As the lens and body are matched, the image processor can be tailored around the optics. The RX1 has options to turn on or off correction for vignetting, distortion and chromatic aberration depending on your preferences. I turn on correction for distortion and C/A but turn off vignetting correction as I find a slight fall off of brightness at the corners very attractive wide open at F2.0. Despite the high end optical design there’s still a bit of distortion and purple fringing in the raw files but it is correctable in post.

Sony are leading the way with sensor technology at the moment. The only cameras with higher DXOMark scores all have Sony sensors. The Nikon D600 has the same sensor. The D800 has a joint Sony / Nikon 36Mp monster and the D800E the same but without an anti-aliasing filter. The RX1 also dispenses with the OLPF but unlike the D800E, it uses bespoke optics matched to the sensor.

The RX1 comfortably outscores the new Leica M Type 240 for dynamic range at a film rivaling 14.3 stops, a whole stop more than the Leica. The latitude is incredible. Some raw files I thought unsalvageable due to over exposure were brought down 3 stops in Photoshop to a normal level without any burnt highlights or banding in the roll offs. It outscores the Leica in low light, offering cleaner images at ISO 3200 to 25,600 and for colour depth. Is there anything left!?

Oh yes – the 5D Mark III is caned by the RX1 for dynamic range. At 11.7 stops the highly rated Canon sensor lags far behind the RX1 at over 14 stops. Yes you can believe your eyes  - Canon are playing catch up to Sony on sensor quality. DXOMark rated the RX1 raw files better for high ISO performance and colour depth too. Again a whitewash on all key image quality metrics. A 5D Mark III with the new Canon 35mm F2.0 IS gives you less performance for significantly more money than the RX1, a compact camera.

Micro Four Thirds may strike a good balance between size and image quality, but the RX1 takes that philosophy to a completely new level.

Now you can understand my excitement!

Video

Sadly those hoping for high quality video to go with the high end stills will be disappointed by the RX1.

Just as with the Sony VG900 and A99 it’s nowhere near a match for the Canon 5D Mark III, Nikon D5200, GH2 or GH3 for image quality in video mode. It has so much aliasing and moire you can use it like focus peaking. It has a chronic lack of fine detail and comes nowhere near close to a used $500 GH2 body for video quality. The codec isn’t as clean as the one on the Nikon D5200 and uses an AVCHD wrapper in 1080p rather than the more usable Quicktime MOV. The RX1 has a MP4 mode but that is very heavily compressed at 12Mbit and 1440×1080.

Regarding the HDMI output there’s been a lot of misinformation about this, leaving many confused. The RX1 lacks the uncompressed HDMI output of the A99. It doesn’t even have a full screen image via HDMI nor is it clean (free of graphics and text). It is windowed with huge black borders in all display modes other than playback mode.

This is all very strange considering Sony’s progress with video on the NEX cameras a full 2 years ago. Those were the first of the mirrorless and DSLR cameras to give us 1080/60p – but the video quality of the RX1 is a step backwards from the NEX 5N and NEX 7.

The $600 sister camera RX100 also has a nicer image in video mode with more fine detail and less aliasing than the $2800 RX1.

Since this full frame sensor also went into the Sony VG900, a dedicated high end video camera, you’d have thought they’d have optimised the sensor for video a bit and got rid of the issues.

I also have noticed a bug there the last 50 or so lines in AVCHD 1080p mode were blurred (see below).

Although the RX1 has 1080/60p, cinema 24p and has full manual control – everything the big A99 has – I’d rather Sony had just taken it off altogether. It’s a total embarrassment.

Form factor

This camera motivates me to shoot and experiment. “Bag me”, it says. It goes everywhere with you.

Have a look at the usual DSLRs lined up in glass cabinets at the shops – it seems no matter who makes them, they’re broadly homogenous. Not the RX1. This finally breaks the “serious camera” mould, not just in terms of the design and specification of the camera, but in terms of how you use it and the look of the stills.

You cannot get the RX1′s lens on a DSLR. It is seated just a fraction from the sensor rather than adhering to the long 44.5mm flange of a Sony Alpha body. That they were able to get such good corner sharpness with such a short flange is amazing. The lens has a total of THREE Aspherical elements, one of them a new type of Advanced Aspherical. The character of this lens only comes with the matched sensor and body of the RX1 and that character is unique.

All other large sensor compacts on the market such as the Nikon Coolpix A, Fuji X100S and Sony RX100 have comparatively small sensors compared to the full frame RX1. Even APS-C is significantly smaller. Only a large full frame beast such as the Canon 5D Mark III, Sony A99 or Nikon D800, with a very high end 35mm prime, would give you a similar feel to images as the RX1 – but there are even more differences than just spec. The larger cameras almost always wind up being used differently due to their form factor. In terms of stealth factor for street photography for example people can react differently to a larger camera, as it is more easily noticed. With the RX1 they very rarely break the spell by looking directly at the lens. As a result the photos are more natural, more often.

Full frame

Full frame for me is not about extreme shallow depth of field, but about having a more pervasive separation of depth. I don’t always want to have an object off the focus plain, or a background, completely creamed out and unrecognisable but I do want there to be enough separation between points of interest in the image – that subtle blur of objects not directly on the focus plane even if the focus point is quite far from the camera gives a more 3D look to the photo.

With crop sensors at wide angle (even at fast apertures) parts of the scene further away blend sharply into the background, rather than being marked out.

Of course on Micro Four Thirds and APS-C to mitigate this lack of 3D depth you can shoot at longer focal lengths or focus closer but that changes your composition completely. Inspired by Tarantino and Kubrick I’ve always preferred wider focal lengths and full frame gives an unparalleled look to a 35mm F2.0. The 3D look on APS-C is just never quite as pervasive as it is with full frame and the RX1 with the Zeiss 35mm F2 gives you the characterful look of a Voigtlander Nokton 25mm F0.95 on Micro Four Thirds twice as often than the Nokton itself when used on a camera like the OM-D.

Fixed lens

I find I get the best creative results with a prime. It makes me concentrate and focus on the image not the choice of lens. Really you get two focal lengths with the RX1 – one at the time of shooting and one in post. It is 35mm full frame and 50mm (equiv.) APS-C when cropped 1.5x in Photoshop. Resolution is so high and detail so crisply rendered at 1:1 that cropping to a narrower focal length is perfectly doable and to have a zoom in post is very useful. 35mm and 50mm are very versatile focal lengths and I rarely use lenses longer than 135mm for my stills work. I still have my DSLR when I want to use the beautiful Canon 135mm F2.0, or a modded lens like the Flare Factory 58 or a tilt shift.

Focus

I wouldn’t rate the AF system on the RX1 as terrible, but it is certainly a weak area along with video.

So many cameras lack a good manual focus assist or fast AF or both that I assume some companies must think focus is secondary to stitch panorama. Exposure and focus should be priorities on every camera bar none. Simple as that.

AF on the RX1 is much better than the X Pro 1 and X100, but it is nowhere near as fast to lock on as the GH3, OM-D E-M5 or Nikon V2, nor as silent. AF on the RX1 has a more mechanical feel, you can feel and hear the whirring motor drive. Considering Sony have faster and more discrete AF on their NEX cameras and lenses you’d have thought the RX1 would be at least as good, but it actually slower! I do however find it more accurate, which is essential for such a large sensor and F2.0, so maybe there’s a reason for the focus system being designed the way it is.

Whilst I find it extremely accurate in good light and acceptably quick for most kinds of work (whereas the X Pro 1 was unacceptably slow), it is pretty bad when shooting sports or in low light, bordering on unusable. The RX1′s mediocre AF system has also got stick from one of the main targets of the camera, street shooters and they have a point but one of the greatest of all time Henri Cartier-Bresson never needed it – really this depends what style of street photography or photojournalism you’re doing. If you’re going to point and hope, the AF will let you down. If you’re going to lie in wait, think ahead to what will happen then the AF will work just fine as long as the scene is not moving very busily and unpredictably.

Fly by wire manual focus on the RX1 is not the disaster it is on some other cameras like the X100 (though the X100S improves the feel). The manual focus ring is responsive and you can rack quickly and precisely like you can on a manual lens from the 70′s. That’s progress!

I recommend turning off the automatic manual focus assist and instead assigning this to one of the two function buttons when you want to use it. The live view feed has so much moire and aliasing it acts like peaking is enabled which is just as well since the peaking function of the camera bafflingly only works when in magnified focus assist! On the RX1′s little brother the RX100 and ALL OTHER CAMERAS peaking works as it is meant to work.

Electronic viewfinder

The attachment OLED-EVF is a very useful add-on but a mixed bag performance wise.

Electronically it is very good with a nice high resolution OLED display. Great colour and contrast. It seems to be identical in performance to the OLED in the NEX 7 and the TruFinder attachment for the NEX 5N. It is slightly higher priced than that but has a metal housing and can be angled up or down.

I generally enjoy using it but it has one major issue and one smaller quirk.

The optics are poor, with a lot of smearing when your eye is not directly straight with the viewfinder. Your eye needs to be so straight to avoid the soft edges that most of the time you find yourself concentrating more on getting your eye straight than on actually composing a shot. The optics should have been made larger to compensate for the edge smearing issue, but the compact size of the EVF meant Sony didn’t do this.

Now the quirks – the diopter on the side of the EVF can easily be changed by accident, the proximity sensor does not allow automatic switching on eye contact between the camera LCD / EVF and the EVF has no securing screw like the one for the NEX 5N. This is both a blessing and a curse. It means the RX1 features a ‘normal’ hotshoe not a proprietary attachment like the NEX 5N and NEX 7, whilst maintaining the smart electronic functionality.

With the superb OLED EVF attached you can focus manually very quickly whilst composing the full frame without needing to use any focus assists thanks to the terrible moire in the live view video stream

Build quality

If not Leica standard the RX1 is exemplary for a compact camera aside from two areas. Although the lens is solid and feels excellent, the focus ring is plastic. The only plastic part on the entire lens aside from the Zeiss badge is the most important part of the lens. That makes little sense.

The battery / card door is plastic and again that seems a shame given that the entire body is a sleek carved block of alloy cold to the touch – even the lens cap is metal!

The flash has an all alloy construction and intelligent design, so no expense spared there unlike on a lot of other compacts with their pop up plastic stalks.

Puzzling stuff

There’s a few things lacking which are on the much cheaper NEX and Alpha cameras and rival mirrorless cameras. A few quirky things.

I’ve already mentioned the slower AF than NEX cameras, well there’s also no in-body stabilisation like on Alpha cameras. Sony owns all this technology, why isn’t it on their $2800 RX1? The lens lacks the optical image stabilisation of OSS NEX lenses and the articulated screen mechanism from the equally thin NEX 5R body is missing altogether. Still an underrated feature. Unlike the Fuji X100S the camera lacks a built in ND filter despite rather needing one with such a fast lens and a leaf shutter maxing out at 1/2000 when shooting at faster apertures than F4. The battery is the same as the one in the tiny RX100, with this camera packing a full frame sensor you’d think they’d have put a bigger battery in there to match. Size shouldn’t be the problem here as the tiny Nikon V1 has a smaller body but packs in the battery from the Nikon D800!

OK so maybe not quirks but major flaws but the images are just so good and the camera just so nice to use, that complaining about these quirks is a lot like abandoning a Ferrari 458 Spider because there’s a scratch on the passenger door.

Yes Sony should work on these areas for future models and some of them I don’t see an excuse for. Stuff like the lack of built in ND filter – probably no room between the lens and sensor and hard to get the quality, are forgivable. Other things like peaking not working over the full image is however a bit less justifiable and should be looked at in a firmware update as should video quality which needs to be drastically improved (or rather fixed – I consider it broken).

Conclusion

Aside from the disappointing video quality, the RX1 is a gem. I’ve been looking for a stealthy street camera with that Henri Cartier Bresson magic for a long time and now I finally have it, without resorting to the expense of a Leica system.

I had hoped the Fuji X Pro 1 would be that camera, but the cripplingly slow AF, terrible manual focus feel and chunky body resulted in it going on eBay. Nor does that camera have the full frame look of the RX1 and outright image quality.

I’ve lost count at the number of times I’ve picked up my iPhone as a camera due to convenience and the fact you have it with you, and shared a photo taken with it. Well it is time that stopped. From now on I’m carrying my RX1 around. If it had a phone on the back, it’d be even better!

I’m kidding but the point still stands. You’re more likely to pick the RX1 and shoot stills with it than a 5D Mark III. As for crop sensor cameras, every time I used one for stills in the back of my mind something bothered me – that temptation to lug my full frame camera around purely because the quality was unmatched. Now not only does the RX1 match the look of the 5D Mark III it betters it. The RX1 gives you an image quality reason for using it not just a convenience factor or size advantage. Imagine no longer having to carry a bulky DSLR and set of primes, let alone a huge medium format camera, tripod and lens onto a fashion shoot.

Medium format quality in a compact camera? It had to happen some day I suppose.

Hurray for technology!

Pros

• It gets the important stuff so incredibly right
• Leica M image in a compact
• Over 14 stops dynamic range
• 24MP sensor has even higher resolution in reality thanks to lack of OLPF filter D800E style
• Superb low light performance
• Zeiss lens easily worth almost a third of the cost of the camera
• Rich colour
• Superb JPEG engine
• Solid build quality and excellent handling on the most part
• Versatile choice of focal length
• Incredibly small for a full frame camera and currently unique on the market with no rivals
• Almost completely silent leaf shutter
• Stealth factor very attractive for pros and photojournalists – discretely small, dark, silent and not intimidating
• Characterful images feel unique to the RX1, especially when shooting wide open at F2
• Attractive design gets admiring glances
• High quality LCD and useful brightness setting for sunlight visibility
• Attachment EVF has good resolution and contrast (though poor optics)
• Accurate AF system (though not that fast)
• Convenient USB charging (though no separate charger in the box for your $2800)
• Useful auto-ISO in M mode allows exposure compensation dial functionality even with full control over shutter and aperture
• Snappy menus and responsive controls
• Capable built in flash
• Actually useful scene modes! Handheld Twilight produces genuinely better JPEGS in low light, Sweep Panorama mode is class leading

Cons

• Heavy line skipping in 1080p, worse video quality than the little $600 RX100 (simply unacceptable aliasing and moire, very electronic feel to video)
• All manual focus assists disabled in video mode (no magnification and no peaking)
• Peaking only works whilst zoomed in, defying the whole point of having peaking function
• No uncompressed or clean HDMI output
• No optical or in-body image stabilisation and electronic SteadyShot for video heavily crops the image (down to APS-C!)
• Live view quality in P,S,A,M stills modes lower than in movie mode
• Purple fringing sometimes quite evident wide open at F2.0
• Poor battery life well short of peers (notably Nikon V1, GH3, OM-D E-M5 or a DSLR)
• No built in EVF and add on has issues – poor viewing angles, smearing, cannot switch between main screen and EVF with proximity sensor, no hot-shoe securing mechanism
• AF not up to standard set by Micro Four Thirds cameras or Nikon 1 series especially in low light and video mode
• No articulated screen makes tripod work uncomfortable
• Maximum 1/2000 shutter speed at fast apertures (due to leaf shutter) makes shooting at F2.0 in daylight marginal without ND
• No built in ND filter like Fuji X100S
• SD card slot placement not optimal (in battery compartment rather than side)
• Focus ring is plastic and doesn’t feel quite as solid as the rest of the camera
• Confusing and unnecessary separation of stills and videos in playback mode

The post Sony RX1 Review appeared first on EOSHD.com.

Download sample 2K DNG 24p frame from the hacked 5D Mark III and see the image quality for your own eyes!

Recently Magic Lantern had the breakthrough discovery of beautiful pin sharp 2K sensor feed in raw format on the 5D Mark II and 5D Mark III.

This completely turns our understanding of DSLR video image quality on its head. Why is the video quality so far from what the camera is really capable of?

Disclaimer: Some of the following represents my thoughts and understanding of how camera technology works and some of it may be speculative as I am not a Canon engineer!

As I understand it the sensor does one sampling run to provide a feed that is used for both live view, video and HDMI. Magic Lantern has gained access to that image and it is very nice, lovely 2K full frame 14bit raw straight from the sensor via the camera’s buffer memory. The image processor (DIGIC) then has to produce an image for the live view LCD (approx. 640×480 or 1MP) and a second output for compressed 1080p video from that one high quality feed.

Alex of Magic Lantern told me DNG was ported from the Canon CHDK project and the camera can successfully write DNG frames to the buffer at 24fps to 30fps in 2K mode, which we now have access to (up to 30 frames at a time) in silent stills burst mode which is now working with Magic Lantern. Steps and research are being taken to see what the feed can be used for aside from silent stills, 30fps burst mode and timelapse with no mechanical shutter movement…

What this shows is somewhat revelatory – that the sensor isn’t the problem when it comes to moire or the general poor soft image quality in video mode.

So what is going on?

UPDATE: Sam (you may know him from the Flaat picture profiles) thinks the sensor is still line-skipping to produce the 2K raw image. He has run tests on 22MP 5D Mark III raw stills in Photoshop using the most basic nearest-neighbour downsampling to mimic 3:1 line-skipping on the sensor the end result looks very similar to the 2K raw output in Magic Lantern. In the DSLR community we have traditionally tended to blame line-skipping for soft resolution and moire in video mode, but maybe that isn’t the whole story. Sam’s tests seem to show that image quality doesn’t suffer too much with it.

The 14bit raw is turned into compressed 8bit 4:2:0 but rather than use all the native resolution of the 2K raw frames, the image processor is playing all kinds of tricks, at one point even downscaling to 1904 and then back up to 1920 again in the case of the 5D Mark III. Why does it do that?

Canon’s Chuck Westfall and Tim Smith say line skipping on the sensor was necessary to get down from the native sensor resolution of 21MP+ to 2MP, implying this results in moire and aliasing. However the 2K output of the sensor is much cleaner without the same moire issues as the final 1080p video, and resolution is much closer to 1080 lines than the mushy 5D Mark III 1080p is.

The softer image with moire and aliasing appears to be introduced by the image processor further (and crudely) resizing the 2K raw image, which seems unnecessary. All the camera should need to do is crop to 16:9 and compress the data, not reduce the resolution dramatically further like it is doing. It could be that in order to provide a simultaneous VGA live view feed and 1080p to the SD card, the 2K data has to be downrezed to a middle ground between 640×480 and 1920×1080, then upscaled back to 1080p but I’m purely speculating.

I asked Alex of Magic Lantern today how feasible it would be to get continuous raw video recording from the 5D Mark III and our conversation went like this:

“I think 1920×720 is doable on the 1000x card.

“In 720p, the raw data is 1920×670 or something around this.

“I also believe the video and live-view image is coming from that [2K feed], since they are perfectly synchronized.

“[Canon's] image pipeline is probably optimized for fast math, not for high quality.

“They don’t do highlight recovery, they probably denoise too much… who knows?”

Disclaimer: I think it is important to keep expectations in check, nothing is confirmed with regards raw video and don’t forget to donate to Magic Lantern to keep the progress coming.

I will be trying the firmware hack later today.

UPDATE: I now have the burst mode silent pics feature running and sure enough I’m getting up to 30fps 2K raw from my 5D Mark III until the buffer fills up after only about a second. If the Magic Lantern developers can find a way to write this raw output at 24fps to the card despite the large data rates, even at reduced resolutions, we could have continuously recording raw video.

The post UPDATED: Canon 5D 2K raw feed – 1920×720 possible on 1000x card? appeared first on EOSHD.com.

Is it any better? Rather than spoil the surprise please do go ahead and watch the short test video. You can download the original file at Vimeo here.

The test setup was -

• Blackmagic HyperDeck Shuttle II external recorder
• Small HD DP6 monitor
• HDMI mode set to 24p
• Internal recording is 1080/24p IPB mode not ALL I. This is because I feel IPB has the better image quality.

Workflow in post was the usual one I do for the 5D Mark III.

Added sharpen filter at 20 in Premiere CS6 and lifted shadows on the street shots (there’s also an ungraded HDMI shot in there).

Lens was Contax Zeiss 50mm F1.4.

A nice bonus is that the focus magnification on the camera also outputs via the clean HDMI feed giving you a 3:2 (or maybe 4:3) 1080p crop mode. Detail in this mode is very good and it seems like roughly a 16mm crop of the full frame sensor. The aspect ratio makes it interesting for anamorphic lens users.

The HDMI output frame rates are 60i, 50i and 24p. There’s no 25p! The 50i is interlaced. That isn’t great news for PAL shooters. The actual frame rate in 24p mode is 23.976fps.

The HDMI output seems to have more of a green cast in the blacks than the internal recording.

The camera can record to the card whilst also recording via HDMI. When the Mirroring option is enabled in the menus, the camera will not display any icons or menus via HDMI and the LCD will remain on whilst an external recorder is attached.

Thanks Rod on the EOSHD Forum for this image

On the Zacuto EVF via the Blackmagic HyperDeck Shuttle, only a 1:1 mapping of the upper corner of the image was displayed and none of the presets or scaling options got me the full image. I swapped to the Small HD DP6 only to find that in 24p mode it has a tearing of motion in the upper part of the frame, and only works when the HDMI is set to Auto or 50i. But I need it to be set to 24p for the recorder! I need to do further research here to find out if it is a HyperDeck Shuttle output or monitor issue.

On whether it is really 4:2:2 colour sampled as Canon claim I have my doubts. I don’t see any difference to the internal recording to suggest 4:2:2 and I doubt the sensor is sampling 4:2:2 from high resolution in video mode like a C300.

For now think of it is as 4:2:0 wrapped in 4:2:2 ProRes, try keying it and see if it meets your expectations (it is still only 8bit of course).

The post Canon 5D Mark III uncompressed HDMI sample footage appeared first on EOSHD.com.

Iscorama anamorphic on the 1D C and my test footage at Vimeo

I shot with the Canon 1D C and Canon C100 Cinema EOS cameras recently with SlashCam in Berlin. The 1D C is the most Jekyll and Hyde camera I’ve ever used and a difficult camera to review. The 1D C is a marriage of beautiful 4K images and an unwilling partner who hasn’t shifted an inch to accommodate it.

High end DSLR stills are ahead of even the best digital cinema cameras for image quality on a per-frame basis. The Nikon D800 does 7K raw images with 14.5 stops of dynamic range. If ever there was an opportunity to marry Canon’s expertise in high speed stills shooting to a moving image, the 1D C is it, but what a shame it makes no concessions at all to video ergonomics.

Image quality

The 4K image of the 1D C is way better than the best 1080p cameras like the C100 or C300, both in terms of the astounding level of resolution and the film-like feel to the image. Despite being only 8bit it really does banish any feeling of ‘video’ especially in Canon LOG mode.

Here’s a 1:1 crop from 4K:

And the original composition!

Viewing the 4K images on my 2.5K Dell 2711 the detail really does pop and I don’t even have a 4K screen yet. I enjoyed having my eye drawn to tiny details. That doesn’t happen in the same way with 1080p.

Although when viewed at 1:1 the 4K image is not quite at the mark of true 4K, it is still a significant leap from the C300. 4K also gives you the ability to crop in post if shooting 4K and delivering in 1080p but that will further reduce the full frame look of your lenses.

Here it is a full 4K frame (click the image to download it) with my Iscorama 36 anamorphic on the Contax Zeiss 50mm F1.4 -

In low light it is a new benchmark, completely displacing the Canon C100, C300 and Sony FS100. ISO 3200 is like silk, ISO 6400 is perfectly usable and even ISO 12,800 can be worked with especially if downsampled from 4K to 1080p and some light noise reduction applied in post.

Here’s a 1:1 crop of ISO 12,800 on both cameras. The C100 has been upscaled to 4K (as you can see, it is no match):

With noise reduction turned off you get the usual DSLR style noise (bright speckles of red and green) rather than a nice film-like grain (as seen on the Ikonoskop and Blackmagic Cinema Camera). With it turned on you risk a more plastic looking image and a drop in detail which is all the more noticeable in 4K mode. My advice is to turn noise reduction off and apply where needed in post if shooting at ISO 6400 and above.

Here’s the character of the noise for your pixel peeping pleasure…

Is the image better than a Red Scarlet, Blackmagic Production Camera 4K or FS700 with 4K recorder? That’s what the 1D C should be compared to. Certainly the 1D C has one advantage over the Scarlet (and likely Blackmagic S35 offering) – much cleaner in low light. However there’s no 12bit, 10bit, raw, global shutter, compressed raw, HD-SDI, peaking, no cinema camera features at all really. Rolling shutter is also a real issue on the 1D C – quite severe in 4K – so the global shutter equipped Blackmagic Production Camera certainly has an advantage when it comes to shooting fast paced action or handheld sequences.

Parts of the image suffered from a banding in the roll off to highlights, where the camera would render a steep drop to a blown highlight with just 2 or 3 bands of colour. There was a nice gentle roll off to blacks from the mids, but again a stepping of shades when going from dark grey to black over larger areas of the frame.

This 1D C only has 24p in 4K mode, the firmware update is due later. Under Berlin’s street lights the only usable shutter speed was 1/50 or 1/100, any other resulted in an unusable flickering image.

Canon log mode had an impressive effect on dynamic range, saving highlights and shadows where a normal picture profile would crush and burn. There are a few ungraded Log shots in the sample video for reference. I could happily shoot in low light with Canon Log without a significantly nosier image.

You can see my Log and low light test in the second half of the sample video (top of the page).

Overall, I am totally in love with the 4K image from the 1D C.

The 4K market

Inevitably 4K is the next cinema standard. Manufacturers are jostling for position. At this stage 4K is more important for the industry than the consumer (most of whom don’t yet need it). In the early days of a new standard it is important to establish a presence as Blackmagic are doing so with the Production Camera and Resolve 10. A state of the art 4K workflow from acquisition to post for $4000. Amazing.

Red have already established themselves technologically and Sony seem to be taking the lead with the F5 and F55 in the broader commercial videography market.

Canon’s 4K efforts are -

• The C500 – very expensive 1080p camera with a tweak to allow 4K
• The 1D C – very expensive stills camera with a tweak to allow 4K

Both the C500 and 1D C are a toe in the water, and Canon can’t be fully satisfied with their efforts. It is not an all guns blazing approach like Sony, Red or Blackmagic, it feels more like a side project at the R&D department, which is a huge shame given the imaging prowess of Canon’s stills sensor technology in the 1D X and 1D C.

The 1D C is a frustrating camera and I want to fall in love with her, but she happens to be a pain. You can’t have a camera review without a car analogy too, so it’s like having a Ferrari engine in a hover craft. It should really be in a car. Did that work!?

On an affordable DSLR you don’t mourn too much about the snared cinematic beauty trapped inside a stills camera. It is what it is. For Canon to offer such a tantalisingly beautiful 4K image in such an ill suited form factor at $12,000 is an avoidable situation because they could easily have put this sensor and 4K image in a C100 form factor for $6000 and owned the professional video world like they do stills with the 1D X. Sadly a huge added value with minimal overheads won the day.

Ergonomic problems

Using the 1D C on the same shoot as the C100 it becomes even more stark just how poor it is to use as a stills / video hybrid. Not a single concession has been made to the stills camera for video shooters apart from the additional 3 recording formats.

Unlike on most consumer DSLRs there’s no dedicated movie mode or record button, very much like the old 5D Mark II. The 1D C goes backwards from fundamental improvements Canon made with the 7D over the 5D Mark II way back 3 years ago – there’s no lever to switch live view mode between stills and video, instead the option is buried deep in the menus. This wouldn’t be such a problem if you could use the shutter release to grab a full frame raw still in the same mode you record 4K in but if you try this you get an error message. If you need to quickly grab a moment as a raw still, you have to exit live view altogether and use the optical viewfinder, or switch the live view mode to stills deep in the menus.

Canon log is also buried in the menus. When you eventually reach it, the Canon Log sub-menu contains 1 option – the option to turn it on! Only upon switching on Canon log does the reason for the sub-menu become clear – about 4 more parameters appear to fine tune stuff like saturation in log mode, then out you go through the cascading menus back to finally shoot. There’s no dedicated button to switch picture profile or quickly turn on or off Canon log!

Above: the form factor the 1D C’s image deserved but didn’t get

Setting ISO in live view mode obscures the frame with a list of ISO numbers on a darkened background. So to really see the effect of the exposure change you have to exit the ISO menu or half-guess at what the exposure looks like.

The start / stop button for video is assigned to a function button near the shutter release which is one of the least tactile and hardest to feel buttons on the camera, being so small and recessed. I recommend reassigning it to Set but there didn’t seem to be an option for this.

The focus assists for video – for there are none – come in the form of the usual stills magnification mode which every DSLR has, but unlike on a $400 T2i the button is placed awkwardly well away from any fingers and thumb in a bizarre location under the screen. Manually focussing really requires an external monitor, as shooting 4K requires precision sharpness but then you lose the DSLR stealth form factor and small size – so at over $12,000 why not just shoot with a proper cinema camera?

The crop factor in 4K mode is 1.3x and does not give either the full frame look to your lenses OR the Super 35mm look, as it is in-between. I’m used to building my lens collections around Super 35mm, Full Frame and Micro Four Thirds. APS-H is something else entirely and in 4k on the 1D C my Leica R 28mm feels something like a 36mm or 40mm and the character changes. This leads to inconsistency on a shoot with Super 35mm and full frame cameras, in terms of the look of lenses not just the field of view. On a full frame camera I have a beautifully gentle vignette on the Leica and yet it isn’t there on the 1D C in 4K.

1080p modes

The Super 35mm 1080p mode is sharp but I don’t consider it much of a selling point when you can get the same standard of Super 35mm 1080p from the C100 at $5500 without the ergonomic (and economic) struggle. In full frame 1080p mode the 1D C is identical to the $6000 1D X and the ‘new’ 1080/60p doesn’t seem very new when you realise it looks like upscaled 720/60p, which the much cheaper 1D X and 5D Mark III already have. Sneaky.

Downsampling 4K Canon Log to 1080p in post however results in a very cinematic image which definitely has the edge on the C100 and C300.

Battery capacity and screen

Most of my 1D C shoot involved me kneeling or crouching, or looking at a low battery indicator. Although the 600D has an articulated screen, and the GH3 has an articulated screen with weather sealing, the 1D C makes no concession to the video shooter here either despite being 12x more expensive than the GH3! As a result it is horrible to use bare bones, requiring constant kneeling or even lying down to be able to see the LCD when shooting on a tripod. Battery life is poor in video mode. Although almost a quarter of the camera body is taken up by a battery almost as wide as a 2.5″ SSD drive, it rates at just 2450mHa compared to the Nikon V1′s battery of 1900mHa and 5200mHa for the Canon C300 battery.

Above: my ultra small $400 Nikon V1 (seen here with Red Epic) has a similar capacity battery to the 1D C. What gives?

The Nikon V1 is a tiny mirrorless camera, the whole body is almost the size of the 1D C’s battery and yet manages to fit a camera in there somehow. Whilst we’re on the subject of space – the 1D C could easily have been designed to fit a 5200mHa Canon C100 battery and a 480GB Red mag sized SSD without increasing the size of the camera or dropping the DSLR form factor. A DSLR battery and CF card is all well and good for stills, to not upgrade these aspects for movie recording at $12,000 is not going to endear Canon to their customers very much.

Post

The 1D C is designed to fit into existing DSLR workflows (both post and production ones) and needs extensive rigging on a set to become more usable. That for me defies the point of having a DSLR in the first place – the stealth factor, the small size, the convenient marriage of stills and video in one device. Hollywood must surely have moved on from working around DSLR ergonomics now they can rig up the small Sony F55.

On common 600x compact flash cards the 4K mode randomly stops recording. On 1000x speed cards, your bank manager randomly calls during a shoot. On a 64GB 1000x CF card you can fit just 16 minutes of 4K footage, and archiving a year’s worth of footage becomes almost as expensive as archiving raw but with none of the benefits.

The 4K MJPEG is designed to fit existing workflows for Canon DSLR or Cinema EOS footage, but even on powerful Mac Pro hardware and CUDA I couldn’t get smooth playback in Premiere CS6. Changing playback resolution to 1/2 or 1/4 didn’t help, which leads me to believe the bottleneck is the drive, but the files were being streamed of a RAID 0 array of 3! The poor performance of MJPEG could be due to Premiere selecting the wrong mode for playback as it is not used to coping with MJPEG at such high resolutions (maybe the wrong playback buffer size). One mooted solution is to transcode to ProRes 4K but then your storage requirements bloom to even more unmanageable proportions. It has been suggested that simply changing the file extension to MPG (MPEG) to force Premiere to handle MJPEG in a better way, can give smoother results – I will try this on my next shoot.

It is clear at this point that the 1D C is not going to be a C100 or C300 replacement for those wanting 4K as it doesn’t even touch the Blackmagic Production Camera for ergonomics let alone Red or Sony. Consumer mirrorless cameras like the Panasonic GH3 and DSLRs have better video ergonomics for well under $1000.

Conclusion

I’m left in mourning for Canon’s early 4K efforts. For a start the price makes them an incredibly risky ownership prospect. In the case of the C500 would you blow $25,000 on a big petrol hungry car knowing it would cost you a lot to run and depreciate by $10,000 in the coming 12 months? The 1D C is also at risk of huge depreciation. I know many who can afford the asking price, but not many willing to risk the depreciation for 4K they don’t really need yet.

If the 1D X sensor can do 4K video, just put it in the 1D X and sell loads! I don’t understand Canon’s business strategy for charging $12,000 for a small quantities of a barely modified camera based on one that is relatively mass produced. The establishing of a new standard it is an opportunity to create space for one’s self and really push aggressively for market share, because there’s a whole television and film industry to feed, a billion dollar industry just beginning to build around and invest in 4K content. It is important to get there first and in the right way. The 1D C is not the right way.

The same flawed approach has been taken with the C500 which is almost identical to the C300 and requires an external recorder. Canon do not have the software or post production products to do an end-to-end workflow like Blackmagic for $25,000 let alone $4000. The fan on the C500 is too loud and the cabled external recorder isn’t an integrated solution like on the much cheaper Sony F5.

The Canon 1D X is designed to be held to the eye and the sensor is a very fast one, enabling sports shooters to capture full resolution stills at up to 14fps and “hey look we can also do 4K, charge double!”, but it has worse ergonomics for video than the $700 Nikon D5200 consumer DSLR, less video optimisations than 2009′s GH1 and all for that bargain price of $12,000 in the form of the 1D C. I’m a big believer in the end result (the image) being what counts. Though I was always happy to work around the shortcomings of DSLRs at $1000, to be working around them at $12,000 seems obscene.

Pros wishing to mix stills and video on one shoot will be frustrated that the 1D C does convergence so poorly. You might be better off with a separate 1D X and a Blackmagic Production Camera for the same money. It isn’t that the DSLR form factor hinders video on the 1D C, it is that absolutely no attention has been paid to making this particular DSLR form factor work better! Zero!  It is purely a stills camera, designed for photographers not filmmakers, that happens to record video to compact flash cards. 4K or not, that’s a lazy implementation of video considering the asking price.

The 1D C follows a familiar pattern of late for Canon. A great story, stifled by a cynical product strategy, coasting into our consciousness on a wave of good will from Hollywood. By current form I’ll be getting my 4K acquisition solutions from Blackmagic Design or Sony.

Pros

• Lovely end results in 4K, silky smooth and amazing detail
• World leading low light performance at high ISOs
• A flagship stills camera and same robust build quality as the 1D X
• Sharp in 1080p with Super 35mm crop (but the C100 has that for a lot less money)
• Ready to shoot out of the box (though painful to use bare bones)
• Internal 4K recording to common media and no need for external solution aka C500 (though media is expensive)

Cons

• Absurdly high price and huge potential for depreciation
• 4K only real selling point over half priced 1D X, but benefit of 4K questionable for audience and clients in 2013. By the time it becomes an industry standard the 1D C may well be obsolete!
• Zero handling concessions to video, not even a video record button on a $12,000 movie camera!!
• Live view activation required every time you turn on the camera
• No dedicated movie mode
• 4K mode has an odd crop factor, nowhere near the full frame look to lenses
• Very poor rolling shutter performance especially in 4K mode
• Mushy 1080/60p image quality looks like upscaled 720p
• Poor battery run time and low capacity considering physical size
• Cannot quickly shoot stills whilst in 4K recording mode
• Live view stills / movie mode hidden deep in the menu, no dedicated lever like on the 5D Mark III
• Video quality in full frame 1080p mode barely better than a $3000 5D Mark III, vertically smeared
• Expensive media (1000x 150MB/s 128GB CF minimum standard) and many cards required
• 4K requires card speeds above 100MB/s – only a little more optimisation would have enabled use of cheaper 600x 95MB/s cards
• No form factor excuse for not using small customised SSD mags like Odyssey 7Q / Red
• 8bit disappointing to see at $12,000, banding an issue under ISO 400 especially in Canon LOG mode
• Canon log buried deep in the menus
• No direct access to picture profiles like on consumer DSLR
• ISO menu obscures entire exposure / composition
• Poor feel to m.Fn button near shutter release (default for activating movie recording)
• Buttons, even the on / off switch impossible to see in the dark
• Illumination of top and rear info panels turns off whilst recording
• Menu system designed for stills shooters (lots of pages of AF!)
• Focus assist not designed for video shooting, no peaking
• Awkward positioning of magnified focus assist, more apt for playback mode than shooting
• Magnified focus assist too slow to move around the frame
• Awkward to use on a tripod for bare bones DSLR stealth shooting – no articulated monitor, stealth factor disappears when you add monitor
• Requires firmware update for 25p (and only came about after months of requests)
• Not nearly enough changes to the 1D X to justify doubling of the price
• No high frame rates (HFR) in 4K, not even 30fps
• Poor 4K editing performance in Adobe Premiere CS6 due to inefficient MJPEG codec and possibly compatibility issues

The post The Canon 1D C review appeared first on EOSHD.com.