Raw video on 5DMK2

Started by olik, May 15, 2013, 08:19:46 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

avasarin

Quote from: Kabuto1138 on May 18, 2013, 06:04:25 AM
Hey guys,

Is there a recent MODULE? hast that been changed in the last 24-36 hours for the 5d2?

thanks

Don't know if in the latest nightbuild there's something useful!
5dmk2 - Nikkor set(20-28-35-50-105mm), Vivitar 85-105 - BMPCC
http://www.frabiatofilm.com
https://www.facebook.com/stringssenzatempo

cmac

Quote from: avasarin on May 18, 2013, 09:34:56 AM
Don't know if in the latest nightbuild there's something useful!

I'm somehow missing the "M" menu in the nightbuilds... Is it just because of me or there is something not generally implemented?

And another thing - It seems that when I do 4 recordings and then exit the session - only the last one is something that I can process with raw2dng - others are giving me: This ain't a lv_rec RAW ... Should I fix each file by adding the footer of the last one to them?

edit: after copying the last 10 raws from the last file to the previous batch, they are processed normally, but then there is a strange thing. All my testas came out greenish... Will keep testing.

bronson

I'm a bit lost now with the last builds for 5d mark 2...
The May16 evening build caused me a lot of invalid raw files. I try to rewrite a correct footer, but some resolutions are not present in the May15 build.
So if someone could compile and upload the last corrections of the raw_rec or lv_rec in the module and a last corrected build it will be great..

The official nighty build was very cool before raw video. Perhaps things are not totaly unified now between 5d2 5d3 600d etc...
I hope we could play with new working things

Many thanks to all involved.

andyshon

Heres some 5D2 test footage, shot in the wild at 1880x720 with the May 16th build.

Thanks to a.d. for the footer fix, and the new build. I shall give it a whirl.

https://vimeo.com/66442959

Dunc101

Hi,

Have any of you had experience with getting the playback function on the mark IIi to work on mark 2?

I was thinking of trying it out?

Cheers
D

olik

Quote from: Dunc101 on May 18, 2013, 12:14:39 PM
Hi,

Have any of you had experience with getting the playback function on the mark IIi to work on mark 2?

I was thinking of trying it out?

Cheers
D

works for me on build 17th (check first post). It is in black and white, only works for the last clip and kind of slowmotion but it does work!

core_32

First of all thank you ML!

Quote from: andyshon on May 18, 2013, 11:32:34 AM
https://vimeo.com/66442959
This is really beautiful!
Is it true that 5dmk3 doesn't have that moire/aliasing issues 5dmk2 has?

Anyone thinking of selling 5dmk2 to buy 5dmk3???

Africashot

Quote from: core_32 on May 18, 2013, 12:33:30 PM
First of all thank you ML!
This is really beautiful!
Is it true that 5dmk3 doesn't have that moire/aliasing issues 5dmk2 has?

Anyone thinking of selling 5dmk2 to buy 5dmk3???
I am still hoping that buying a Mosaic Engineering filter will do the job...
ML 5D2 & T3i

andyshon

Quote from: core_32 on May 18, 2013, 12:33:30 PM
Is it true that 5dmk3 doesn't have that moire/aliasing issues 5dmk2 has?

The 5D3 may be slightly better but both cameras are going to suffer badly with aliasing, and no filter will really help without decimating the resolution. You're only recording from about 1 in 9 of the pixels on the chip. Thus almost 90% of the image is falling down the holes in between, and worse still the R,G and B source pixels being used for each output pixel will be some considerable distance apart.

So details will drop in and out of existence as they move across the sensor. Monochromatic edges will have serious colour fringing and bright details will cycle through the rainbow as they move. All very familiar to anyone who's used scientific high speed cameras with no low pass filter. That said, if you deal with this in post on a shot by shot basis, then you must be able to do a better job than Canon do in camera. ACRs colour noise reduction handles the colour artefacts remarkably well and I believe the moire reduction is pretty impressive for stills.

But this is all throwing some light on why DSLR video generally looks so terrible. Canon have really done a pretty amazing job getting rid of most of the colour fringing and detail flickering, in camera, in real time. Just wish they gave us the choice to turn that 'feature' off!

mvejerslev

My first test shots. Its like having a hole through to another dimension!

https://vimeo.com/66414746
5D Mark II, PC

pholler

Hey guys,

i just received the Lexar 1000x for RAW video. I am a bit disappointed. I thought i could get 1880x1250 with the best card on the market. Well i made some tests now with global draw off. The best resolution possible with the best card on the 5D2 seems to be 1880x840 with 63 MB/s writing speed for continuos filming.
The card should do better. Shouldn't it? Could anyone achieve a better resolution than that?

I used the May 17th build.

Yoshiyuki Blade

Quote from: pholler on May 18, 2013, 03:47:26 PM
Hey guys,

i just received the Lexar 1000x for RAW video. I am a bit disappointed. I thought i could get 1880x1250 with the best card on the market. Well i made some tests now with global draw off. The best resolution possible with the best card on the 5D2 seems to be 1880x840 with 63 MB/s writing speed for continuos filming.
The card should do better. Shouldn't it? Could anyone achieve a better resolution than that?

I used the May 17th build.

As far as I can tell, that's pretty much within expectations for me. Some tests on the 5D2 showed about 65 MB/s, which is probably the limit of the camera. 1880x1250 requires over 90 MB/s which is pushing it even for the 5D3.

pholler

Quote from: Yoshiyuki Blade on May 18, 2013, 03:52:52 PM
As far as I can tell, that's pretty much within expectations for me. Some tests on the 5D2 showed about 65 MB/s, which is probably the limit of the camera. 1880x1250 requires over 90 MB/s which is pushing it even for the 5D3.
I thought it's gonna be more on the 5D2. The CF-Benchmark gives me 79 MB/s with 2048k-Buffer. The RAW-files are quite big, so these 79 MB/s should be realistic during RAW Video, shouldn't they?


Yoshiyuki Blade

Quote from: pholler on May 18, 2013, 04:08:51 PM
I thought it's gonna be more on the 5D2. The CF-Benchmark gives me 79 MB/s with 2048k-Buffer. The RAW-files are quite big, so these 79 MB/s should be realistic during RAW Video, shouldn't they?

You did the benchmark in playback mode, which makes it read/write faster than it actually does. Try again with the liveview open and global draw off. Turn off any auto power off functions if it shuts off before the benchmark finishes.

This will actually be useful to know since we haven't seen too many benchmarks of the 5D2 with a 1000x card, but the results I've seen so far have been low and I'd expect similar results.

pholler

The benchmark keeps on crashing in LV-mode, GD off, performing the last test. So i cannot upload a screenshot. The writing speed with maximum buffer (16 MB) seems to be crucial. At this buffer level i can achieve 64,7 MB/s. Seems like you are right. So buying a faster card than 800x makes no sense on the 5D2. Which on the other hand means there is no change of recording full HD RAW vids on the 5D2 at the moment.

I just received the Lexar from Amazon. As a fast card does not give me any benefit i should send it back. Too bad.

pholler

I just managed to get the screenshot. Here is the benchmark for the Lexar 1000x, 5D2, LV on, GD off:

Kabuto1138

How about the Toshiba 1066x, has anybody tried that one yet?

silvertonesx24

Quote from: pholler on May 18, 2013, 04:08:51 PM
I thought it's gonna be more on the 5D2. The CF-Benchmark gives me 79 MB/s with 2048k-Buffer. The RAW-files are quite big, so these 79 MB/s should be realistic during RAW Video, shouldn't they?



How'd you get such good speed here? I can't break 65MB/s on my 32gb 1000x Komputerbay.

abaeza

Max Rex out of a 800x Lexar 16GB Card is 1880x524 limited to around 4GB per take. Lot's of noise. Any solution?

Yoshiyuki Blade

Quote from: pholler on May 18, 2013, 04:56:36 PM
I just managed to get the screenshot. Here is the benchmark for the Lexar 1000x, 5D2, LV on, GD off:

Indeed, it's kind of unfortunate that this seems to be our limit. But at the very least we can manage something close to a 2.35:1 ratio.

When I inquired a1ex about this limit, he says it's not safe to assume anything, but that was a few days ago: http://www.magiclantern.fm/forum/index.php?topic=5247.msg34263#msg34263

I'm not sure if all signs point to a hardware bottleneck just yet, but I haven't heard any good news about this as of yet.

olik

I think with a 1000x card the resolution of 1880x787 should be doable, giving us a nice 2.39:1 aspect ratio and close enough to full HD.

JackDaniel412

Quote from: olik on May 18, 2013, 06:26:17 PM
I think with a 1000x card the resolution of 1880x787 should be doable, giving us a nice 2.39:1 aspect ratio and close enough to full HD.

With my SanDisk Extreme Pro 16gb 90 MB/s i record continuosly at 1880x768, with 1000X you can do 2.35 aspect ratio (1880x800).

silvertonesx24

Low light, source lights, anamorphic test. Heavy grading which holds up so well- the RAW footage was pretty much all tungsten. password is rawtest.

5dm2, 1000x 32gb Komputerbay, 1880x860, May 15th firmware. Really looking forward to a new compile as this firmware ruins half of my footage with magenta blocks, but I've had no luck with May 16 and 17 firmware and the raw2dng converting problem.


pholler

Quote from: silvertonesx24 on May 18, 2013, 06:07:04 PM
How'd you get such good speed here? I can't break 65MB/s on my 32gb 1000x Komputerbay.
I made a mistake. I did the benchmark in Playback, not in Liveview, mode. I redid the benchmark afterwards in Liveview and i couldn't break 65 MB/s either.

silvertonesx24

Quote from: pholler on May 18, 2013, 08:53:23 PM
I made a mistake. I did the benchmark in Playback, not in Liveview, mode. I redid the benchmark afterwards in Liveview and i couldn't break 65 MB/s either.

I don't know a whole lot about the internal workings of the camera, but to me that looks like it means that the controller isn't as limiting as initially thought.

My ideal resolution personally would be 1920x1280.