Low light test of 5D Mark III raw vs H.264

[JHines]

Why are you so mad, man? I live in the U.S. They're available everywhere.

 

Not mad at all. 

[HurtinMinorKey]

I think a low-light test between the c300 and the hacked 5D would be fun.

[Ilias Giarimis]

If you looks at the actual raw files the DR advantage for the BMCC is clear.

 

If we have raw black frames it's easy to mesure the DR of BM vs 5DIII.

 

BTW if 5DIII uses all the pixels of the 16/9 sensor crop (no skipping) in video mode then the expected DR at 1920X1080 is about 1 stop more than DxO's measured DR of 11.74 stops for Canon at 8Mp. 

[Hugo]

If we have raw black frames it's easy to mesure the DR of BM vs 5DIII.

 

BTW if 5DIII uses all the pixels of the 16/9 sensor crop (no skipping) in video mode then the expected DR at 1920X1080 is about 1 stop more than DxO's measured DR of 11.74 stops for Canon at 8Mp. 

 

 

I'm curious, how do you mesure DR using black raw frames? whats the process.

[Hugo]

If you looks at the actual raw files the DR advantage for the BMCC is clear.

 

I took a good look at both raw DNG files.. and I wouldnt say BMCC has a clear advantage.. they look pretty even to me.. If BMCC has better DR, which is possible.. then it's very hard to tell with these DNG.

[HurtinMinorKey]

If we have raw black frames it's easy to mesure the DR of BM vs 5DIII.

 

BTW if 5DIII uses all the pixels of the 16/9 sensor crop (no skipping) in video mode then the expected DR at 1920X1080 is about 1 stop more than DxO's measured DR of 11.74 stops for Canon at 8Mp. 

 

Correct me if I am wrong, but I thought each photosite has the same dynamic range for what it can measure in terms of signal. And if you are imputing DR by using multiple photosites, more photosites would mean more DR, not less, as your statement implies. 

[HurtinMinorKey]

And let the record state that I'm cheering for the 5D3. I'll never put more hours into a camera than I did with my mkii.

[Sean Cunningham]

And let the record state that I'm cheering for the 5D3. I'll never put more hours into a camera than I did with my mkii.

 

Of course, there'd be no need for BMD camera users (or concerned citizens) to rain on these guy's 5D's parade if 5D users hadn't been repeatedly showing everyone their "O" face every five minutes, making absurd declarations and moving through the five phases of grief in repetition.  This place and everywhere else DSLR related has become a splash zone of premature emissions.

[Andrew Reid]

Really puzzled by your lack of enthusiasm for 5D raw Sean.

 

For once the hype is totally justified.

[HurtinMinorKey]

I agree Andrew's latter point. 

 

The improvement between the 5D2 and 5D3 was marginal. By all appearances, this hack single handedly bumped the 5D ahead a generation (or two). I have no doubt it will put the on-board codec in the c300 to shame in terms of overall image quality (in most settings). 

 

How many times have so many people gotten this big of an upgrade, for free? It's certainly worthy of some hype.

 

As long as it doesn't end up fragging people's sensors(no reason to think that it will).

[mtheory]

5D raw video is in 240dpi resolution. It isn't just raw...it's Retina RAW. :D

[Ilias Giarimis]

I'm curious, how do you mesure DR using black raw frames? whats the process.

 

Open the black frame DNG with RawDigger http://www.rawdigger.com/ , go to "preferences" and disable "subtract black" and then read the values under sigma (σ). The displayed values are the stdev of the black frame which stdev is the definition of noise in engineering.

This way you have the "read noise" or else the "noise floor".

Then find in exif data (exif button in RawDigger) the "White Level" value and subtract from this the exif "Black Level" or better subtract the avg value (rawdigger display) of the inspected channel. This way you have the "Max Level"

Dynamic range is Max Level/Noise floor. This is the so called "engineering DR" for a single pixel.. it's what DxO measures as "Display DR" and what BlackMagic reports as DR.

 

To make a fair comparison we have to compare at the same output size. So if the DR of 5dIII is for 1920X1080 (2073600 pixels) output and BM is 2432X1366 (3322112 pixels) we have to normalize the later for 1920X1080 output ..

DR(BM1080) = DR(BM) + log2(sqrt(3322112/2073600)) = DR(BM) + 0.34 stops ..

 

http://www.dxomark.com/index.php/Publications/DxOMark-Insights/Detailed-computation-of-DxOMark-Sensor-normalization

 

Correct me if I am wrong, but I thought each photosite has the same dynamic range for what it can measure in terms of signal. And if you are imputing DR by using multiple photosites, more photosites would mean more DR, not less, as your statement implies. 

 

You are correct about the implication of more pixels as shown a few lines before in this message, ... but wrong at interpreting my message.

 

I mean that if there is no skipping the above normalisation for final size is OK and by resampling/binning 8Mp to 2Mp gives +1 stop DR. But if there exists any line or pixel skipping we have to account for the real number of used pixels.

 

If for example, due to skipping, only half of the pixels contribute to the final image we have to normalize 4Mp to 2Mp and the differense will be 0.5 stops.

 

So we agree isn't it ?? More pixels with given DR result in more DR when we refer to the same final size ..

[blackrat]

I'm curious, how do you mesure DR using black raw frames? whats the process.

Shot something that blows out all channels and measure the signal recorded (WP=white point)and then shoot a frame with body cap on and measure the noise of that (RN=read noise) and find the BP=blackpoint (avg value of the black frame):

log2((WP-BP)/RN)=DR

 

You also might want to normalize. You could just apply the factor to RN so for instance say you had a 21MP cam and got 6ADU RN then to compare at 8MP basis you  do (8/21)^2 * 6 and get like 3.7 or so ADU RN and then you just use 3.7 above instead of 6.

[blackrat]

Correct me if I am wrong, but I thought each photosite has the same dynamic range for what it can measure in terms of signal. And if you are imputing DR by using multiple photosites, more photosites would mean more DR, not less, as your statement implies. 

Max signal stays the same but averaging many photosites to a few means noise floor goes down and then plug in and DR goes up. Trading spatial resolution away in the process.