50D Raw video

[rommex]

how come u make test on ISO 400 of canon 50d?
watch video above i made on ISO 3200
and thats even without ETTR stuff

I explained it in my post. Also note that usually DR is wider at lower ISOs.

And please stop this trolling. Obviously after all the forums and nicks that you've gone through, you've grown some hard forehead of yourself to ignore remarks to your meaningless posts.

I guess soon the time will come when we on the forum will have to address the moderator to keep this forum clean and acceptable for those who are hungry for 50D RAW info, rather than having to sift through someone's egotistic garbage...

[Monti]

I explained it in my post. Also note that usually DR is wider at lower ISOs.

And please stop this trolling. Obviously after all the forums and nicks that you've gone through, you've grown some hard forehead of yourself to ignore remarks to your meaningless posts.

I guess soon the time will come when we on the forum will have to address the moderator to keep this forum clean and acceptable for those who are hungry for 50D RAW info, rather than having to sift through someone's egotistic garbage...

yeah just become professional
and then complain about my posts people)

[VideoSauce]

Hey guys, got a quick question for yall

So recently I began optimizing my 50D for general shooting and took the advice from a poster upon being able to achieve focus easier using LV. He stated to use a combination of focus-peak and the zoom box alongside the digic setting "Cartoon Look" set to number "1". This made achieving focus super easy but recently the setting (though active) did not appear on my LV display regardless of whether or not I re-installed the latest version of tragic lantern 2.0 or disengaged and reengaged the digic setting. All of my other settings have remained the same and I am running a lexar 32gb x100 card.

So whats going on here?

Also I never got the chance to record while the cartoon look was applied, so is it applied to your recorded footage? if so I just assume you disengage the setting before shooting.
Thanks guys,

[a1ex]

Any warnings in the menu? When something doesn't work, it usually prints why.

[rockfallfilms]

I explained it in my post. Also note that usually DR is wider at lower ISOs.

And please stop this trolling. Obviously after all the forums and nicks that you've gone through, you've grown some hard forehead of yourself to ignore remarks to your meaningless posts.

I guess soon the time will come when we on the forum will have to address the moderator to keep this forum clean and acceptable for those who are hungry for 50D RAW info, rather than having to sift through someone's egotistic garbage...

+1 - Monti's comments are beyond tedious now.

[rockfallfilms]

Hi all, I recently was shooting on BMCC and decided to make comparison of BMCC vs 50D.
So I decided to evaluate the usable DR of both cameras.
This is not a scientific test -- just practical approach.

I was shooting both scenes in a ETTR way (without ETTR module), so that the whites are mostly overexposed.

Then I tried to push the shadows and see how much that can go until you have unacceptable noise (purely in my opinion)

I'm pretty sure that someone can make more sense from these tests, so welcome!

BMCC was shot at ISO 800.

I decided to shoot 50D at ISO 400. On the one hand, I wanted their ISOs somehow to match. On the other hand, ISO 800 of 50D would be too unfair to 50D, being a no low-light king.

The same 17-50 2.8 lens was used.


CONCLUSION: Based on the eyedropper tool info in the shadows I estimate that BMCC is 1 stop wider in its Dynamic Range.

As I said, this is a brief non-scientific test. Don't shoot me based on this. Any reasonable comments are welcome.

And yes, the original RAW files are found here: https://copy.com/oG6C31ghjeJ0

_{PS Monti, please relax and don't respond to this. You may create your own thread somewhere in the forum and post, post, post...}

Interesting test, you could probably push those 50D shadows even further and use neat video. Pretty amazing considering the price difference between the cameras.

I've got that Shot by shot book too!

[Andy600]

I'd like to see them both at ISO800. I know DR might be reduced even further on the 50D but it would be a more accurate indicator and show noise levels. Noise is manageable to some extent.

[rommex]

I'd like to see them both at ISO800. I know DR might be reduced even further on the 50D but it would be a more accurate indicator and show noise levels. Noise is manageable to some extent.

I could do that next time BMCC is in my hand, but I would disagree it could add accuracy to the test. 800 ISO is basic (native) sensitivity with BMCC. With Canon 50D it is ISO 100. I tend to think that you have to test them in these respective ISOs to be scientifically accurate. Correct me if I'm wrong.

[Andy600]

I could do that next time BMCC is in my hand, but I would disagree it could add accuracy to the test. 800 ISO is basic (native) sensitivity with BMCC. With Canon 50D it is ISO 100. I tend to think that you have to test them in these respective ISOs to be scientifically accurate. Correct me if I'm wrong.

So ISO 800 is the lowest setting on the BMCC?

TBH I'm not sure what constitutes native ISO on a DSLR but if it's it's lowest then the 50D would be nearer ISO 200. I'm interested to see them at the same ISO sensitivity mainly to see the difference in noise. I think you'd be pulling the shadows down rather than pushing them up and it should (in theory) be cleaner. At least that's the concept behind the DUAL ISO developments.

Having said that, it's remarkable how close the cameras actually are considering age and sensor development. I've been looking at the BMCC and pocket as a future investment but haven't yet seen anything that makes me want to jump on board. The Kineraw Mini looks to be a better option, albeit a little more expensive. Makes deciding on an upgrade to a 5D3 even harder but I'm in no hurry.

[akumiszcza]

True, so not viable for normal cinema use (big and bulky tripods, optical stabilizers, and great spaces, not viable for indoor use, and so on)

 
But this is exactly what I'm using raw video for. Low tripod, gimball head, tele lens and x3 magnification in camera — very nice for birds. Video is especially useful for identifying uncommon birds, rarities. Furthermore, even when I don't have tele lens with me, I use my kit 17-55/2.8 lens with magnification and it's enough for documentary. Some additional ML features are great for birding too.

[rommex]

Yes, Andy, BMCC has ISO 800 as its lowest.

And that's why you have to use ND filter more often than not on BMCC.
And this is basically how I would do the test ideally: clean lens on 50D and ND8 on BMCC.

I didn't get your part on pulling shadows down. Usually to increase the visible DR (and to approach the iconic filmic look) I push the shadows up. I guess everybody else does, no? )

And the noise (if pushed in ACR to Gamma 2.0) looks like this: BMCC is cleaner, 50D has surfacing the rectangular pattern. But both manageable to some extent, I agree.



But I do, I do second you -- looking at this difference, I don't see a 5-year gap between these cameras. Nor a $1500 gap )

[a1ex]

The noise pattern from 50D can be removed with dark frames. Take a look at the GoPro workflow, they used this method.

[Andy600]

...I didn't get your part on pulling shadows down. Usually to increase the visible DR (and to approach the iconic filmic look) I push the shadows up. I guess everybody else does, no? )

Of course. I must have left my brain in bed today 

[Andy600]

The noise pattern from 50D can be removed with dark frames. Take a look at the GoPro workflow, they used this method.

Dark frame subtraction works well for the 50D. I've experimented a couple of times by leaving the lens cap on for a second but this eats valuable recording time especially if you change ISO or change from non-crop to crop. Does this noise remain a constant? i.e. you can have a set of pre-made dark frames at different ISO values to use or should it be done every shot? I guess if you're shooting in crop there is the possibility that you're shooting with different parts of the sensor so the dark frames wont match.

I can see this is probably a stupid question but can this be achieved in-camera by not opening the shutter for a few frames at the start of recording? or does the shutter need to be open to produce a dark frame?

[a1ex]

I didn't do any extensive testing, so... no idea. Try using a dark frame from a few days ago and see if the pattern changes its place.

There are some tricks to record image data with shutter closed (e.g. quick AF mode, where I had to kill zebras), or on 5D3, by setting shutter blanking time to maximum. Look in the MLV thread, I've just described the data structure for storing dark correction info (nothing implemented yet).

[Andy600]

I didn't do any extensive testing, so... no idea. Try using a dark frame from a few days ago and see if the pattern changes its place.

There are some tricks to record image data with shutter closed (e.g. quick AF mode, where I had to kill zebras), or on 5D3, by setting shutter blanking time to maximum. Look in the MLV thread, I've just described the data structure for storing dark correction info (nothing implemented yet).

I'll have to make new ones to test with. Deleted the others.

On second thoughts, it's probably better to wait for .mlv because I would need to manually log the ISO for every shot as it's not currently saved to exif.

[rommex]

Speaking about the noise, sensors and everything:

DXO Mark have completed the sensor test of the new 70D, which shows that its sensor has gone a tiny step from both 7D and 50D.

In terms of color depth 70D is better than 50D by 3(!)%
In terms of dynamic range -- by 1.7%

And only in low-light performance it has advanced by substantial 33%

DXO Mark aren't by far an absolute truth, but they catch the main trend I think.
It gives both reasons to rejoice and to contemplate the Canon development strategy 

Proof link: http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/%28appareil1%29/895|0/%28brand%29/Canon/%28appareil2%29/619|0/%28brand2%29/Canon/%28appareil3%29/272|0/%28brand3%29/Canon

[Monti]

yeah if magic lantern wasnt existing
i would probably buy black magic cinema camera or pocket version for sure

also any developments in new DSLR without compact flash card doesnt interest me in any ways

i have to test that dark frames things
if anyone explain it little bit or send link?

[rommex]

pretty simple: http://en.wikipedia.org/wiki/Dark-frame_subtraction

[Andy600]

Dark frame subtraction is pretty straight forward.

Shoot a few frames with the lens cap on, remove it and continue your shot as usual.

When you process your raw file you'll have some dark frames. Add one of these (or a composite of several) to a layer above your footage and choose 'difference'. This should help clean up fixed pattern noise. The dark frame should be at the same ISO and be the same frame dimension as the footage you're shooting so you may need to shoot several dark frames at different ISO values and crop sizes.

You can also composite one dark frame with a copy of itself that's slightly blurred as a difference matte.

Dark frames are also useful for creating default, ISO based noise profiles in Neat Video.

[Monti]

hmm thats interesting
i should try that

but now this push me to idea

how about to implement dark frame in magic lantern settings

so that we can get automaticly clean images?

[Andy600]

hmm thats interesting
i should try that

but now this push me to idea

how about to implement dark frame in magic lantern settings

so that we can get automaticly clean images?

You obviously didn't read what a1ex wrote above 

[VideoSauce]

+1 - Monti's comments are beyond tedious now.

I have gone through everything and all text present is just the explanation text of what the setting does. I tried using the other digic effects including Cartoon Looks "2" and "3" and they all work. At this point I am at my wits end when it comes to this issue.

Also, if I may post outside the general topic here for a second. Do you guys believe through firmware modifications we can squeeze a few extra trimmings out of say a Hero 3 BE? Nothing crazy just forced CBR and a possible multiplier? Possibly a manual exposure mode? ISO selection alone would be phenomenal.

[Monti]

You obviously didn't read what a1ex wrote above 

seems like he is just on the way to do this function...

[paulforte]

So ISO 800 is the lowest setting on the BMCC?

No, 200 is. I've no idea why they say 800 is the "native" ISO (whatever that means). I shot with it at 200 and 800 and 800 was clearly grainier as you'd expect.