Author Topic: 5D3 Dual-ISO RAW Video Test Comparisons (night time)  (Read 10314 times)

DeafEyeJedi

  • Hero Member
  • *****
  • Posts: 3411
  • 5D3 | M1 | 7D | 70D | SL1 | M2 | 50D
5D3 Dual-ISO RAW Video Test Comparisons (night time)
« on: December 04, 2013, 10:26:26 PM »
Magicians -

Here I came up with a thought of taking Dual ISO RAW Video recordings of outdoors by the City Hall/Fire House in Los Angeles, CA while waiting for my step-daughter to finish her Girl Scout meeting. This was taken at night for test purposes!

I've also decided to push the boundaries to the next level by doing three sets of different Dual's (100/1600, 400/3200 & 800/6400) just for the heck of it and to see how it compares in its performance.

The entire testing was done with ETTR, 1/48th, 14mm T3.1

Workflow as usual - Raw2dng (0.13)/ cr2hdr (1.6)/ ACR (8.2)/ Quicktime (7)

Enjoy and THANKS to A1ex & ML crew!

Dual-ISO (100/1600)

Dual-ISO (400/3200)

Dual-ISO (800/6400)

Dual-ISO (100/1600) <-- (watch what happens when the car pulls in reverse to park..its rather strange!)

Dual-ISO (400/3200)

Dual-ISO (800/6400)

Dual-ISO (100/1600)

Dual-ISO (400/3200)

Dual-ISO (800/6400)
5D3.113 | 5D3.123 | EOSM.203 | 7D.203 | 70D.112 | 100D.101 | EOSM2.* | 50D.109

a1ex

  • Administrator
  • Hero Member
  • *****
  • Posts: 12564
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #1 on: December 05, 2013, 03:40:50 AM »
A good comparison IMO would be 100 vs 100/400, 100/800, 100/1600 and 100/3200, all color-graded exactly in the same way. You can also use a denoising software on ISO 100.

When watching a test like this, one should be able to decide what settings to use, when setting X works better than Y and so on. This is why I think it's important to use exactly the same grading on all videos (otherwise you are just comparing apples with oranges).

To avoid flicker, make sure you have the same black and white levels on all your DNGs (check with exiftool); cr2hdr doesn't take care of that.

To avoid the "crawling" artifacts, tweak the FPS. IIRC, at 25.000 fps it's OK.

swinxx

  • Hero Member
  • *****
  • Posts: 606
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #2 on: December 05, 2013, 07:02:01 AM »
interesting

on the first shot iso100/1600 and iso 800/6400 are flickering..
400/3200 seems brighter than 800/6400.. isnt that strange?

@ alex
how can i apply noise reduction only on the iso100 lines?
greets.

DeafEyeJedi

  • Hero Member
  • *****
  • Posts: 3411
  • 5D3 | M1 | 7D | 70D | SL1 | M2 | 50D
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #3 on: December 05, 2013, 07:05:23 AM »
Point taken, A1ex! I actually had that in mind and went with the same XMP's for each scene (so three separate XMP's were used - One for each SCENE) in order to try and keep everything the same except for the Dual ISO's.

However, I've tried researching on ExifTool forums in attempt to learn and understand how to use ExifTool through Terminal. I am very well aware of how to use Terminal since I have jailbroken a bunch of Apple devices over the last several years and still do.

Still yet, I can't manage to get to see my Black and White levels on my DNG's in Terminal...I guess I'm just not sure what's to type in to navigate through my DNG files -- so I feel a little insecure to admit this to you. Thoughts?

Would it be reasonable enough for me to take down the tests that I've just posted online and do another testing that is proper according to your statements from earlier (which makes sense to me) after all.

Thanks for suggesting for me to change my FPS to 25 fps in order to avoid the "crawling" artifacts.

I'm still puzzled and not sure why it flickers and sometimes it doesn't? Even if it cr2hdr doesn't take care of that?

Appreciate your time & comments, A1ex!

P.S. Have you read my private message that I sent to you awhile ago?
5D3.113 | 5D3.123 | EOSM.203 | 7D.203 | 70D.112 | 100D.101 | EOSM2.* | 50D.109

a1ex

  • Administrator
  • Hero Member
  • *****
  • Posts: 12564
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #4 on: December 05, 2013, 07:08:29 AM »
Should be something like this:

to check:
exiftool -BlackLevel -WhiteLevel foo.dng

to set the same value for all files:
exiftool -BlackLevel=8192 -WhiteLevel=50000 *.dng

(you can take the values from the first file, or some sort of average)

PM didn't arrive.

DeafEyeJedi

  • Hero Member
  • *****
  • Posts: 3411
  • 5D3 | M1 | 7D | 70D | SL1 | M2 | 50D
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #5 on: December 05, 2013, 07:23:28 AM »
so from my understanding is that I type in "exiftool -BlackLevel=2048 -WhiteLevel=15000 *.dng" for each DNG's file or if I were to type that in terminal right now and it should automatically make all my DNG's in the same value from here on out with using Raw2dng/cr2hdr conversions?

Am I on the right path?

Thanks again for your time.
5D3.113 | 5D3.123 | EOSM.203 | 7D.203 | 70D.112 | 100D.101 | EOSM2.* | 50D.109

a1ex

  • Administrator
  • Hero Member
  • *****
  • Posts: 12564
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #6 on: December 05, 2013, 07:35:35 AM »
Yeah. 2048/15000 are typical values for regular raw DNGs, but dual ISO DNGs are 16 bit (not 14), so the values are around 8192 and 50000.

The flicker issue is because cr2hdr was designed for photos and handles each frame without looking at the others. The same happens with ACR with certain controls.

DeafEyeJedi

  • Hero Member
  • *****
  • Posts: 3411
  • 5D3 | M1 | 7D | 70D | SL1 | M2 | 50D
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #7 on: December 05, 2013, 07:50:49 AM »
So I just typed that in my terminal and then I closed the terminal. Restarted my laptop. I plan on taking one of my RAW files and start the process all over again with raw2dng (0.13)/cr2hdr (1.6) [were you able to figure out the problem with the recent update for raw2dng yet?]

As of now -- I continue to use cr2hdr 1.6 and NOT 1.9 for DUAL-ISO Raw DNGs only -- is this still the correct way?

Although, I am still using the latest cr2hdr 1.9 for Dual ISO photos only and they still look great to me.

Also should I try to avoid certain controls in ACR or does that not matter anymore since I 'fixed' the issue regarding Black/White levels with ExifTool through Terminal.

Thanks again bud!
5D3.113 | 5D3.123 | EOSM.203 | 7D.203 | 70D.112 | 100D.101 | EOSM2.* | 50D.109

a1ex

  • Administrator
  • Hero Member
  • *****
  • Posts: 12564
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #8 on: December 05, 2013, 09:50:39 AM »
I have no idea what's the difference between 1.6 and 1.9, I only tag them with a changeset number. Try with the windows executable under Wine.

DeafEyeJedi

  • Hero Member
  • *****
  • Posts: 3411
  • 5D3 | M1 | 7D | 70D | SL1 | M2 | 50D
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #9 on: December 05, 2013, 10:48:10 AM »
Well here's Take 2...

I did it all over again with raw2dng/cr2hdr after adjusting the Blacks/Whites values in ExifTools through Terminal. (Thanks to A1ex!) Apparently it did help even out the flickering.

However, I do still see minimal flickering along with noisy aliasing but again I used the same XMP file for all three of them in ACR (8.2).

I will try and redo the rest when necessary...I am off to bed for now.

Enjoy and keep believing in Magic!



5D3.113 | 5D3.123 | EOSM.203 | 7D.203 | 70D.112 | 100D.101 | EOSM2.* | 50D.109

DeafEyeJedi

  • Hero Member
  • *****
  • Posts: 3411
  • 5D3 | M1 | 7D | 70D | SL1 | M2 | 50D
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #10 on: December 06, 2013, 07:55:09 AM »
Should be something like this:

to check:
exiftool -BlackLevel -WhiteLevel foo.dng

to set the same value for all files:
exiftool -BlackLevel=8192 -WhiteLevel=50000 *.dng

(you can take the values from the first file, or some sort of average)

PM didn't arrive.

@a1ex -- sorry but I have to admit that after using Terminal last night and typing in the commands for Blacks/Whites values in exiftool is when I realized, after tonight, is the fact that it never did do anything in the first place. So now I'm not sure how I was still able to manage minimal flickering/artifacts in my 2nd Take.

Here's the info what it says on Terminal now as oppose to last night...

Sean-Jacksons-MacBook-Pro:~ SeanMichael$ exiftool -BlackLevel=8192 -WhiteLevel=50000 M03-1933.dng
Error: File not found - M03-1933.dng
    0 image files updated
    1 files weren't updated due to errors
Sean-Jacksons-MacBook-Pro:~ SeanMichael$ exiftool -BlackLevel=8192 -WhiteLevel=50000 *.dng
Error: File not found - *.dng
    0 image files updated
    1 files weren't updated due to errors
Sean-Jacksons-MacBook-Pro:~ SeanMichael$

*grunts*
5D3.113 | 5D3.123 | EOSM.203 | 7D.203 | 70D.112 | 100D.101 | EOSM2.* | 50D.109

a1ex

  • Administrator
  • Hero Member
  • *****
  • Posts: 12564
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #11 on: December 06, 2013, 08:34:40 AM »
Indeed, a computer program does what you tell it to do, not what you want it to do.

If you are in the bedroom and you ask it to wash your dishes, it will look up the dishes in the bedroom and will not find any ;)

RenatoPhoto

  • Moderators
  • Hero Member
  • *****
  • Posts: 1509
  • 5DM3 / 7D
Re: 5D3 Dual-ISO RAW Video Test Comparisons (night time)
« Reply #12 on: December 06, 2013, 01:12:16 PM »
Sean-Jacksons-MacBook-Pro:~ SeanMichael$ exiftool -BlackLevel=8192 -WhiteLevel=50000 M03-1933.dng
Error: File not found - M03-1933.dng
    0 image files updated
    1 files weren't updated due to errors

I dont know it works on Mac, but in windows you have to use the exif command in the directory where the dng are found.
http://www.pululahuahostal.com  |  EF 300 f/4, EF 100-400 L, EF 180 L, EF-S 10-22, Samyang 14mm, Sigma 28mm EX DG, Sigma 8mm 1:3.5 EX DG, EF 50mm 1:1.8 II, EF 1.4X II, Kenko C-AF 2X