Canon 5DMk2 versus BlackMagic Pocket Cinema 4K

[honza123]

You can compare RAW-video data from Canon 5DMk2 and BlackMagic Pocket Cinema Camera 4K

Canon 5D Mk2 + Tamron SP 28-70mm F/2.8:
ISO 100, Shutter: 1/55, Aperture 5.6
RAW (Reddeercitys) 3008x1080, 10bit, 23,978fps

BlackMagic Pocket Cinema Camera 4K + Olympus M.Zuiko 12-100mm F/4
4K, ISO 100, Shutter: 1/50, Aperture 5.6, 23,978fps

Download from:
http://www.uschovna.cz/zasilka/KTMDZCM9XFINZ6MB-8T4/

My first experience with Blackmagic camera ....

[domo94]

I would love to actually see what you're talking about...

This is what shows up.

[honza123]

 
Sorry, I don't have these old video-files anymore.
I have to shoot new comparison videos (5DMk2 X BPCC4K)

[honza123]

Try this:
http://www.uschovna.cz/zasilka/MB24ETX734LZ2HRH-X42/

Canon 5D Mk2 + Tamron SP 28-70mm F/2.8:
ISO 100, Shutter: 1/48, Aperture 8
RAW (Reddeercitys 2.6k_5d2_4-3-2019_crop_rec) 2560x1200, 10bit, 23,978fps

BlackMagic Pocket Cinema Camera 4K + Olympus M.Zuiko 12-100mm F/4
4K, ISO 100, Shutter: 1/50, Aperture 8, 23,978fps

[yourboylloyd]

Try this:
http://www.uschovna.cz/zasilka/MB24ETX734LZ2HRH-X42/

That doesn't work either

[theBilalFakhouri]

YouTube or Google Drive are a good choices to upload

[domo94]

Yeah idk what the heck is uschovna lol

Or even upload to vimeo, bro. Don't complicate yourself

[honza123]

Canon 5D Mk2 + Tamron SP 28-70mm F/2.8:
ISO 100, Shutter: 1/48, Aperture 8
RAW (Reddeercitys 2.6k_5d2_4-3-2019_crop_rec) 2560x1200, 10bit, 23,978fps

orig
https://drive.google.com/file/d/1nHVNpsfM7KS2CLPjtTkAqeJQAYgpGP6a/view?usp=sharing
preview
https://drive.google.com/file/d/1BB9-vITesGorFlskjaZxvnRVg-DY4mVF/view?usp=sharing


BlackMagic Pocket Cinema Camera 4K + Olympus M.Zuiko 12-100mm F/4
4K, ISO 100, Shutter: 1/50, Aperture 8, 23,978fps

orig
https://drive.google.com/file/d/183pXDeUDvpVm6nE9Q_1AfqHA-LIMrXfL/view?usp=sharing
preview
https://drive.google.com/file/d/1I0AFYeR0d5lSE3-lFCst12y4Ktd3m4iw/view?usp=sharing

[Dmytro_ua]

I really like Blackmagic here.

[domo94]

I guess the focal length is throwing me off, but yeah I like the BMPCC more, actually. Funny enough.

[dfort]

Comparing apples and oranges? The 5D2 is a 10+ year old still camera that had video added to it as an "extra" feature while the Blackmagic 4K pocket camera is designed to shoot video. In addition, you're comparing Prores 422 LT 4096x2160 in log colorspace to MLV raw 10-bit 2560x1200.

Just for fun I put the "originals" (had to process the MLV first) into DaVinci Resolve and did a rough match on the color. The old 5D2 held up quite well against the new BMPCC 4K. Of course in practice the newer camera is probably easier to work with and the post workflow much simpler to manage.

[domo94]

Dfort, I agree apples to oranges. I suppose the settings weren't even enough for a fair comparison.

I am unable to view the sample you provided.

I am still heavily considering a 5d mk III for photo/video purposes.

[theBilalFakhouri]

@dfort

Your video is not working "This video isn't available" wrong link ?

[dfort]

Oops - forgot to press the "publish" button -- here's a link to the video.

https://youtu.be/XmEFWiluLeA

Of course once it is on YouTube any subtle differences are wiped out.

[domo94]

Oh wow never mind. This was for the mk 2. I'm clearly illiterate.

Wow this is a crazy comparison.

[JADURCA]

Did almost same color grading for both and crop BlackMagic one to a close focal length for an easy compare presentation. Wich one you think is Canon and BlackMagic?

https://www.dropbox.com/s/okd14a8wpq5b4d4/Canon%205DMk2%20versus%20BlackMagic%20Pocket%20Cinema%204K.mp4?dl=0

[Tyronetheterrible]

First one, BMPCC4k, second one Canon?

[JADURCA]

First one Canon, second BlackMagic 

[honza123]

Another "test" - apples and oranges with ISO 1600:

The Apple
Canon 5D Mk2 + Tamron SP 28-70mm F/2.8:
ISO 1600, Shutter: 1/48, Aperture 4
RAW (Reddeercitys 2.6k_5d2_4-3-2019_crop_rec) 2560x1200, 10bit, 23,978fps

https://drive.google.com/file/d/10U3pXPJei-eSRYfqcGHKbUWm4aM6O2gf/view?usp=sharing


The orange
BlackMagic Pocket Cinema Camera 4K + Olympus M.Zuiko 12-100mm F/4
4K, ISO 1600, Shutter: 1/50, Aperture 4, 23,978fps

https://drive.google.com/file/d/1dRoExLRxlZSqcYdAnNmIcarTg_U67DIO/view?usp=sharing

I love apple 

[dfort]

Great comparison. Here's what I'm seeing:

BMPCC4K


5D Mark II


Noticed I didn't do that great of a job matching the two cameras. BMPCC is in a flat log colorspace and 5D2 is raw (no colorspace) and matching two completely different cameras/formats isn't easy. Still, they look pretty good from here.

However, let's not kid ourselves. The BMPCC4K is a larger image and the old 5D2 is showing quite a bit of noise.

BMPCC4K


5D Mark II

[domo94]

The two snapshots uploaded make it seem like the 5d mk ii is better in my eyes, anyways. Did you get em backwards?

[dfort]

Did you get em backwards?

Click on the images and it will take you to Flickr where you can see the file names and even download the full sized images if you want to do some pixel peeping.

[dariSSight]

Question @dfort, would the 5D Mark II image look more competitive and have reduced noise if we compared FHD 3x3 1856x1248 vs the BMPCC4K. We are really in the Vortex when we compare 2.6K Crop vs 4K crop sensor of noise, when the real comparison is how much scaling can the 5D Mark II ML FHD 1856x1248 take vs BMPCC4K 4K crop sensor. To make a long story short, What is the pixel durability of a FULL FRAME Sensor Pixel VS a Super 35 Crop Sensor? I've may have missed this discussion in the quality of a FULL FRAME ML FHD RAW sensor pixel durability.

[Deadcode]

First of all, this is an insanely good result from 5D2 in 2019.





On the other hand it's ageing. The crispness (1080p timeline!), the noise, the rolling shutter (jello with the 5D).. it's a no comparison against a cinema camera. The colors are still flying but on this example i prefer de BMPCC4K in terms of color too.

[dfort]

Question @dfort, would the 5D Mark II image look more competitive and have reduced noise if we compared FHD 3x3 1856x1248 vs the BMPCC4K.

No. The comparison here with the 5D2 using raw 10-bit 2560x1200, 1x1 sampling should be more competitive than 3x3 sampling. In addition, you'll get aliasing with 3x3 sampling. Things will get much better when (if?) we can get 1x3 sampling on the 5D2.

To make a long story short, What is the pixel durability of a FULL FRAME Sensor Pixel VS a Super 35 Crop Sensor? I've may have missed this discussion in the quality of a FULL FRAME ML FHD RAW sensor pixel durability.

Pixel durability? Had to Google that one and found some phone drop tests. What are you talking about?

As far as the full frame vs. a Super 35 crop sensor -- the BMPCC 4k has a Micro 4/3 sensor. The 5D2 in 2560x1200, 1x1 sampling mode is probably using even less real estate than Micro 4/3. The Micro 4/3 crop factor is 2.0, right? I believe that crop_rec setting is something like a 3x even though you need to put the camera in 5x zoom mode. In any case, you aren't using the full sensor.

However, this does bring up a good point. Why does a full frame sensor have a more pleasing "look" than the smaller sensor cameras? The cine lens makers have been jumping on the full frame bandwagon lead by the camera manufactures -- Alexa LF, the RED Monstro, Sony VENICE, Panavision DXL2 and -- ok, let's get back to the ML universe. The new crop_rec settings will give you greater resolution but there's something about the look of a full frame sensor that goes beyond pixel peeping.

[dariSSight]

Thanks for the quick response @dfort, What I mean by Durability of FuLL Frame Sensor's Pixel is that a full frame Video Pixel is so much bigger then a Micro 4/3 sensor aka Super 35 crop sensor, so by every FULL FRAME sensor pixel being bigger it gives room for better In Depth Field and less noise. But if like you say Magic Lantern is using less sensor real estate then a Micro 4/3 sensor to achieve 2K, 3K or 4K, that will answer a big question because it means we get more resolution but a much worse image. It might make sense to battle the aliasing factor with 3x3 sampling then get more resolution with a smaller crop factor of the sensor then a Micro 4/3 Sensor. My study say the goal if you decide to minimize video presentation to the spec wars, is to have a FULL FRAME Sensor with 4K recording from a Full Frame Sensor. I was just wondering if ML 5D2 FHD 1856x1248 is using the FULL FRAME Sensor? I know the most we can get from the 5D2 is 1880x1040 instead of 1920x1080, but is 1880x1040 FULL FRAME video recording and we just lack 40 pixel wide and 40 pixel high? This whole conversation is the hunt for what is the best image from ML RAW video from 5D2? Is it 10/12/14bit 2019Mar24 from Experimental, Bleeding Edge Experimental ReddeerCity's 2.6k April16's 2.6k_5d2_R_T_PreView_4-7-2019_crop_rec or is it still our trusty Feb 16, 2014 for MLV 2.0 and Oct 24 2013 for RAW 1.0? Thanks for any response, I just want to help communicate the best image from ML Raw.