Andrew Reid

  DXOMark also measured the 1D C (1D X sensor) at 11.8 stops total so something is wrong with that Canon LOG 12.5 stop test. You said yourself you can't get more dynamic range than the sensor provides in the raw data!   There's also a fundamental error of logic in the way you're comparing raw to MJPEG. You are suggesting that from the 11.7/11.8 stop raw 14bit sensor data, the 14bit raw video contains less dynamic range than compressed 8bit JPEG with LOG curve. Sorry but that makes absolutely no sense at all. By definition the raw data has the optimal dynamic range, and no matter what curve you apply to the JPEG, you lose quality when you compress it to 8bit.

It still has a better form factor than the FS100!

  Interesting. How would it impact dynamic range though? If you look at the dreadful DR in the H.264, I think a lot of that is due to 8bit compression.

Thread cleaned. I agree no more off topic stuff about Bloom. It's wrong to criticise religion.

This test was prepared jointly in cooperation with Rudi at Slashcam - here's his take on it (in German / in English) In the battle of the 1080p cameras, the game has changed. Here's how the 5D Mark III in raw recording mode compares to the best 1080p output from the Canon C300 and 1D C. [url=http://www.eoshd.com/content/10475/canon-1d-c-vs-5d-mark-iii-raw-and-c300-gh2-resolution-comparison]Read the full article here[/url]

Yeah and those same Alexa crews carry a generator and Arri suns 100m down the road, spending countless dollars to do so, then go and shoot ProRes. Madness.

  Yes that's actually what it is, raw still quality.   1080p is sampled from the full sensor area.   The 1:1 crop modes at up to 3.5K are a crop of the CR2 raw stills (every pixel read from a crop of the sensor).

  Wow 50D is looking very impressive, better than 600D.   Ssshhh don't tell ebay!!

They probably put the lag in there on purpose so you'd find the C300 more attractive.

The second video is genius. Very funny and original.

The BMCC definitely has more room to recover the highlights and when they do burn, the roll off is very smooth and film like.   In the blacks though, it is noisy.   The 5D3 has less potential for highlight recovery by about 2 stops and when they burn, they stand out.   The blacks though are silky and clean.   So the cameras both need a different approach. I expose the BMCC to the right to bring up the shadows, but I expose the 5D to the left to preserve the highlights.   James Miller has a good eye. Enjoyed his footage. I did a flatter grade on my footage today, and actually prefer the colour and contrast in the original. It is just so good out of the sensor, it doesn't really need that much work in post unless you are going for a very specific stylistic look.   Of course the potential for saving a shot which isn't exposed correctly is enormous with the BMCC and 5D3 raw compared to normal DSLR footage.

Teehee!!

  Why are you saying this on a forum full of filmmakers. We know. Stop stating the bleeding obvious.   How in any way does the story detract from the camera. They are complimentary. It is like saying to an actor not to bother acting as well as they possibly can because the script matters most.

  Because it isn't simple maths. It's hard maths.

Show isn't over for the Pocket Cinema Camera.   I suppose it is too hard to ask Mark to read anything properly.

  Turmoil! Disaster! Drama! Raw for free.   What a hardship. Better image quality than a $15,000 camera on your $3000 DSLR and the first ever full frame camera to record uncompressed raw internally to compact flash card.   FOR FREE.   I don't call that turmoil I call it a miracle.   You're trying to justify a preconceived purchasing decisions I'm afraid. The Blackmagic cameras whole selling point - ONLY selling point - was the image and raw workflow. Nobody bought a Blackmagic for the smaller sensor, heavy body, poor ergonomics and slipped delivery dates. We held out for the image. Now we have it in a DSLR, along with a ton of other important features - stills, better physical controls, HDMI, full frame sensor, low light performance and all else offered by Magic Lantern aside from raw like peaking, high frame rates, zebra, manual audio, etc.   Magic Lantern is reliable. I can shoot all day at 1080p to my fastest CF cards without a hitch. Battery run times are similar to when shooting H.264! The rig is much lighter than my Blackmagic with external battery.   I own both. I cannot for the life of me keeping the BMCC EF.   Blackmagic still have unique selling points for me though.   A) The Micro Four Thirds mount camera B) 4K and global shutter on the Production Camera C) Super 16mm sensor on the Pocket Cinema Camera D) Resolve 10   But as for the BMCC EF - it's been nice. It's been lovely. But show's over.

Does anybody actually know what the FS700 raw output looks like? Too early too call it on that one. What about pricing? FS700 is $8000 + $2000 minimum for the recorder (Odyssey 7Q) plus codecs have to be bought individually on top of that.   I'm calling it $11k minimum, close to the F5 price so you may as well get the F5.   All still very very far from $3000 and a free raw upgrade isn't it?   Don't forget how good the sensor is in the 5D Mark III. Very very clean, probably native 2500 ISO, 12 stops DR, not to mention full frame. I think people are underestimating it because they haven't downloaded any DNGs and have just watched the clips on YouTube.

  The problem with MJPEG is the CPU power to debayer and convert raw to JPEG inside the camera. That takes a hell of a lot of processing power. The easiest thing for the camera to do is raw! Not H.264. Not MJPEG. Not ProRes. Not anything. Just raw.   Saying the camera could do 2k 60fps in MJPEG makes no sense at all. The CPU would not do it. Also the sensor mode only goes to 60fps with a vertical resolution of around 670.   With the newest builds we are seeing a 16MB/s improvement in buffer write performance to the card and that makes 1440x540 possible at 60fps. Pretty nice. If you want to add a massive overhead to this lovely raw performance, the quickest way to do that is to convert o MJPEG in camera... It ain't gonna happen!

I'd say there's nothing in that F35 shot that the 5D Mark III can't match in raw.

But it shoots raw!   Perspective.

"Point to the ML hack"   Oh you are generous. Raw over laggy shitty looking HDMI is worth one point?   You're not easily pleased are you?

I agree, good to know the info. Thanks Miguel.   It is good to bear this in mind with any bayer sensor. Red is interpolating to get 4K resolution, whilst the Sony F65 over-samples from 8K to do true 4K and true 4:4:4 colour.   With the anamorphic mistake - totally put my hands up to this one. In my haste to get on with shooting, I assumed I was getting 2.39:1.   It is actually 2:1 and the horizontal resolution should be 2560 not 3000. This is quite handy actually as it perfectly matches the native resolution of my Dell display.   Here are the other anamorphic resolutions for 4:3 and 3:2...   2x anamorphic from 4:3 (1770x1280) = 3540 x 1280 (2.66:1) 2x anamorphic from 3:2 (1920x1280) = 3840 x 1280 (3:1)   1.5x anamorphic from 4:3 (1770x1280) = 2580 x 1280 (2:1) 1.5x anamorphic from 3:2 (1920x1280) = 2880 x 1280 (2.35:1)   Do those aspect ratios look right? I'm not too sure.   I'll re-upload the video and it should look perfect.

  I think 2580 x 1280 looks about on the money, which makes sense as 1720 x 1.5 = 2580!     I got 3000 x 1280 by assuming the Iscorama would give me 2.39:1 from 4:3. Seems it does not.   Will have a further look into this and likely re-upload the video tomorrow.   Cheers guys.

As you said it youself you can do raw to RGB with 444 color sampling. Sure there's some interpolation but it isn't inaccurate to call it 444.   Same as calling the Epic 4K, 4K. It isn't REAL 4K. But it's still 4K. Get what I mean?   Also you don't actually know how the sensor is sampling 1080p from the full 6K resolution. It could be oversampling and downsampling from a higher resolution to achieve the 1080p raw for all you know. As you said, the way to get great luma and chroma res is adding more resolution and downsample. The image is so good that could well be happening on the 5D's sensor.   If you'd like floating point precision in terms of facts and figures from me, I'd be the first to admit I am not a mathematician but a filmmaker.   There's some stuff I'm still working out as I go, we learn every day as filmmakers and my blog is part of me learning as I go, sharing as I go. I think the anamorphic aspect ratio needs to be checked too by peer review, since 1920x820 is 2.39:1 and that is how I derived 3000x1280 - it looks however like the anamorphic isn't quite giving us that wide aspect ratio from 4:3 like I expected. Maybe it should be 2.5K x 1280. All to be tried! Still early days.

You only get 16:9 from 4:3 with a 1.33x anamorphic lens.   The ones from the film days are 2x anamorphic. Iscorama is a 1.5x stretch.   Update: corrected the video, should be 2560x1280 from the 1.5x stretch of the Iscorama.