50D Raw video

Started by Andy600, May 22, 2013, 03:40:57 PM

Previous topic - Next topic

0 Members and 2 Guests are viewing this topic.

50Deezil

Quote from: Andy600 on October 25, 2013, 12:04:12 AM
I just did a quick, unscientific test and the max I can get it to show is 11.4EV at ISO 100 & 200 (which proves the Dx0 mark ISO charts) using center weighted average metering.

That's good to know.  I did try to provide as much good supporting information as I could find.

a1ex, I don't understand why you think it's nonsense to describe a cameras Dynamic Range as "shifted to highlights or shadows" when all i'm trying to describe is the results of Dynamic Range tests where they determine the mid gray point as 50% luminance and measure the number of stops below mid gray and above.  Some sensors have more DR in the shadows and others in the highlight range, thus my saying "shifted to".  2 sensors could have the same DR overall but not have the same distribution of stops above or below mid gray.   Its just a manner of speaking not a technical term.  The point being that when establishing a mid point as a point of reference you can determine how much a camera sees into the shadows or highlights.  Saying a sensor is "shifted" one way or the other is relative to 50% luminance, it may have been clumsy but it's hard to shorten it any other way IMO.

a1ex

All Canon DSLR sensors are linear throughout the entire range. All of them have a hard clipping point (white). The "distance" from that white point to the darkest shadow that can be captured with a SNR of 0 dB is the dynamic range.

Mid gray is where Canon meter is calibrated. You can (and should) ignore it when shooting raw. Learn to expose to the right and don't look back.

soundwizard99

Is it me or does the raw video from the Canon 50d look better than the video from the Blackmagic pocket cinema camera. I own both cameras and I'm extremely disappointed with how soft the BMPCC video is compared to the 50d, almost to the point where I may get rid of the BMPCC if the raw upgrade doesn't beat out the 50d. I honestly think that Blackmagic dropped the ball on this camera.


Andy600

Quote from: soundwizard99 on October 25, 2013, 05:47:27 PM
Is it me or does the raw video from the Canon 50d look better than the video from the Blackmagic pocket cinema camera. I own both cameras and I'm extremely disappointed with how soft the BMPCC video is compared to the 50d, almost to the point where I may get rid of the BMPCC if the raw upgrade doesn't beat out the 50d. I honestly think that Blackmagic dropped the ball on this camera.

I would wait and see what the BMPCC looks like with raw before deciding. I've seen the pocket up against the 7D (raw) and it holds up well even with prores.
Colorist working with Davinci Resolve, Baselight, Nuke, After Effects & Premier Pro. Occasional Sunday afternoon DOP. Developer of Cinelog-C Colorspace Management and LUTs - www.cinelogdcp.com

pags

I'm using the Komputerbay 32gb 1000x, I'm receiving what appears to be "dead pixels" or at least that what they appear to be.
Some "dead pixels" are pink and some are just black. Is it my camera or card??

1%

Probably you have dead pixels... I have them on every sensor except I think EOSM.

pags

Quote from: 1% on October 25, 2013, 06:21:51 PM
Probably you have dead pixels... I have them on every sensor except I think EOSM.

I did a few small tests and i don't receive the dead pixels every time i shoot. Sometimes they are their and sometimes they aren't at all.
I did a test last week ranging from iso 100 - 1250, and the dead pixels where in every iso. I did the same test today ranging from iso 100 -1250 and they weren't in ANY shots.
could it still be my camera or CF card?

I have also tried another transcend 32gb 400x with the same tests and haven't received any dead pixels.
i've done 3 tests with my Komputerbay card and dead pixels were random in each shoot.
I've done 1 test with the Transcend card and no dead pixels.
It Appears to be the card, but i need to do further tests to be 100% sure.

rockfallfilms

Quote from: pags on October 25, 2013, 06:33:16 PM
I did a few small tests and i don't receive the dead pixels every time i shoot. Sometimes they are their and sometimes they aren't at all.
I did a test last week ranging from iso 100 - 1250, and the dead pixels where in every iso. I did the same test today ranging from iso 100 -1250 and they weren't in ANY shots.
could it still be my camera or CF card?


If you could post some photo examples, then it would be easier to help you.

1%

It happens randomly, esp bad at high ISO + long power on time. And yes, they move around/come and go.


pags

Quote from: 1% on October 25, 2013, 06:56:04 PM
It happens randomly, esp bad at high ISO + long power on time. And yes, they move around/come and go.

Is there anything I can do to help prevent dead pixels? Anything in proproduction or anything i just need to be cautious of?

1%

Shoot at lower ISOs? Thats all I can think of, its why I think we need dead pixel/finding removal... that thing is 1px big in the actual raw file but it grows when you debayer. Its even worse when you pan and you have all this purple "dust" on your shot.

pags

Quote from: 1% on October 25, 2013, 07:15:34 PM
Shoot at lower ISOs? Thats all I can think of, its why I think we need dead pixel/finding removal... that thing is 1px big in the actual raw file but it grows when you debayer. Its even worse when you pan and you have all this purple "dust" on your shot.

Thanks for the advice and feedback 1%!
Is dead pixel/finding removal currently in the works or line up?

tmte


Mirco

This is my first post here so hello fellow 50D owners and ML users  ;D

I think (this is untested as I haven't encountered the problem  ;)) a relatively simple POST production 'fix' for the dead pixel issue, if you have Adobe After Effects, would be to use the clone tool to have the dead pixel area replicate the area on one side of it. This isn't a true fix but may help mask the problem somewhat.
Though this would become very labour intensive if the dot is moving, it wouldn't prove difficult if the dot is stationary in the frame.



dlrpgmsvc

Quote from: Mirco on October 26, 2013, 09:29:35 AM
a relatively simple POST production 'fix' for the dead pixel issue, if you have Adobe After Effects, would be to use the clone tool to have the dead pixel area replicate the area on one side of it. This isn't a true fix but may help mask the problem somewhat.
Though this would become very labour intensive if the dot is moving, it wouldn't prove difficult if the dot is stationary in the frame.

Do it propagate to all the subsequent frames automagically ? Or do you need to do this frame-by-frame for every frame ?
If you think it's impossible, you have lost beforehand

Wlad81

What is the algorythm of calculating the size (in bytes) of the one minute video file depending on the video resolution? Is that, for instance, (1592*1062*14*24*60)/8?
Canon EOS 5D Mk III + Canon 24-105 F/4 L IS USM + SanDisk Exreme Pro 64 GB (SD, ML Nightly.2021Feb07.5D3113) + SanDisk Extreme Pro 128 GB (CF).

dsManning

Quote from: dlrpgmsvc on October 26, 2013, 02:08:31 PM
Do it propagate to all the subsequent frames automagically ? Or do you need to do this frame-by-frame for every frame ?

Not sure for AE, but for Adobe Camera RAW, just use the clone tool on one frame, and synchronize all frames (making sure clone is checked in the popup).  Hit Apply, and wait a bit for it to process.  Fixed a panning wide shot with a small sensor dust in the sky using this.  VERY hard to notice unless you are really looking for it, and with panning motion in the shot, you are not focusing on the sky so much.  For a still shot it would be almost impossible to tell.

Of course, you can get unlucky and get a spot of changing colours or lighting that you may have to go back and fix a span of a few seconds worth of frames, but just go back to that section and reclone a different area.

araucaria

Quote from: Wlad81 on October 26, 2013, 07:35:49 PM
What is the algorythm of calculating the size (in bytes) of the one minute video file depending on the video resolution? Is that, for instance, (1592*1062*14*24*60)/8?
That multiplication looks allright.

dickson

I've had 2 issues with my 50d since installing the latest build. When changing the iso the lvf goes to a kind of flashing white, and as far as I can see all I can do is turn the camera off. Also, while I can record much longer clips than before at top resolution, if they're too long (over 3 gigs, I think) I get a file that won't open in raw2dng. I have to stop it at about 50 seconds. And at about 60 seconds the lvf just shuts off and that's it.

I guess that's more than two issues. But god knows I'm not complaining about getting 50-second raw clips!

D.L. Watson

Love the new build. Very clean and professional. I like how it says how many seconds I've recorded. Nice.

Only issue I've seen, and maybe it's been addressed in the forum someplace else, but when I go to 5x mode, the preview starts glitching. Anyone else have this issue and is there a solution?

Thanks in advanced.

Took my 50D up to Crater Lake a few months ago with some really old dirty lenses.

See my portfolio of work at www.dlwatson.net

D.L. Watson

See my portfolio of work at www.dlwatson.net

maxotics

Quote from: Wlad81 on October 26, 2013, 07:35:49 PM
What is the algorythm of calculating the size (in bytes) of the one minute video file depending on the video resolution? Is that, for instance, (1592*1062*14*24*60)/8?

I did some calcs for all resolutions, for the EOS-M, what it can do is in green.  However, the higher numbers up to the high 70s (MBS) should apply to the 50D

http://www.magiclantern.fm/forum/index.php?topic=8825.msg82948#msg82948

I would use (1592*1062*(14/8)*24*60)

1% sent me this information, on how he does it for the camera messages (he thought my numbers close enough for my purposes):

Frame size * frame rate is what it does.

int frame_size_padded = (res_x * res_y * 14/8 + 4095) & ~4095;
frame_size_real = res_x * res_y * 14/8;
frame_size / 1000 * fps

It does some calculations with the buffer slots too so thats why you're coming up with different numbers.

funkysensation

Hey guys,

hereĀ“s a test of my new Slider with 50D Raw.


Infos: http://550draw.blogspot.com/

D.L. Watson

Hey folks, just wanted to give you a quick look at the differences when using Neat Video on your RAW 50D footage. It's a great and powerful noise reduction plugin and in my experience works much better cleaning and preserving details in the shadows than denoiser.

See my portfolio of work at www.dlwatson.net