Menu

Show posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Show posts Menu

Messages - itsskin

#26
Ok, it did not hang this 2 times, but result is strange. Please have a look:
How keyframes looked


What script did:


Log:
https://dl.dropboxusercontent.com/u/17452116/Debug.zip
#27
You are awesome :D Working now. int... var... who cares :D
Please give me some some time to run it with some test material.
#28
An error occurred while running a startup script named 'BridgeRampingScript'.  It may not be compatible with this version of Bridge; you can check for available updates by selecting the Updates command from the Help menu.  The script has been disabled, but can be re-enabled in the 'Startup Scripts' section of Preferences.

Error in /Users/itsskin/Library/Application Support/Adobe/Bridge CC/Startup Scripts/BridgeRampingScript.jsx
Line 736: int timeout = 30;//30 seconds
Illegal use of reserved word 'int'
#29
Great! Let me deflicker some other timelapse. Be back with the results.
Thank you!
#30
Script still stops for me randomly.
XMP is being written correctly. It just stops in the middle of the process. If I press "cancel" after it hangs I get something like "resize undefined" if I remember correctly.

EDIT: It says "ReferenceError: resize is undefined"
#31
So any ideas how to fix this? Can I help?
#32
What I did last time was deleting all non-keyframe XMP files, syncing crop value form first keyframe to all others, Ramp all values and run deflicker process.
#34
Here is the debug log. I changed the sequence and now it hangs in other place.
https://dl.dropboxusercontent.com/u/17452116/Debug%20log.zip

I don't know how to go through the code step by step. If log won't help, probably you'll have to tell me how to do this.
#35
Hi!

I have a new bug here with version from the website.
On every 3rd iteration, it is stuck at one file before the keyframe. Suppose it's my one special sequence, but still:



Update: Running script in 2 iterations "batches" works fine.
#36
Thank you again :)

Here is another example of what script can do. It was shot earlier in "manual ettr" :) Adjustments were made in 1EV stops. I'm super happy with the result.

#37
Quote from: Joachim Buambeki on November 05, 2013, 09:30:12 PM
Please try "Metadata->Read Metadata from files" and see if that works.

After this step LR5 is stuck updating. Importing is stuck on "importing".

I prefer LR to edit keyframes and copy\paste settings. It's just my main tool overall.
#38
No error messages, but something is still going wrong:


First white frame XMP file: https://dl.dropboxusercontent.com/u/17452116/_08B8132.xmp
The first white is just after second keyframe. Other non-whites are keyframes also.

Managed to Ramp Multiply selecting less files using keyframes as first and last. Running deflicker process on all files selected now.

UPDATE: Deflicker worked on 1700 images like a charm with great results.

UPDATE 2: Seems like Lightroom can not read .XMP files created. Just doing nothing importing files, or updating from xmp metadata. Unreadable XMP+RAW https://dl.dropboxusercontent.com/u/17452116/Archive.zip
#39
Tried to Ramp All on 1700 images, got this error:



Found the problem: 1 folder was selected within the sequence.
#40
Tested the new version and no bugs so far. Thank you. Great work!!!
#41
Quote from: dmilligan on November 03, 2013, 09:37:31 PM
Nice!

Can you send me the debug info again? I'll see if I can fix it.

I replicated the issue and now it looks different. Second frame is OVERexposed. But still seems because of first keyframe is ignored. Keyframe is +1 EV, second is +6.

Initial sequence with 4 keyframes:


Ramp All:


Settings:


Result:


Iterations log: https://dl.dropboxusercontent.com/u/17452116/10%20Iterations.rtf
#42
If the last bug would be addressed when scrip ignores the first keyframe as per my previous post - it would be 100% awesome.
That's what I managed to get with current version:
#43
It worked! Really smooth transitions now. Except small glitch with 1st keyframe and second after it.
keyframe is +0.9 EV, second frame is +0.22 EV. But after this it's flawless. Thank you!

#44
In case PM didn't work (can't see any sent folder) I'll post it here too:


*** Iteration 1 ***
keyframe 108B8326.CR2: 201
keyframe 108B8340.CR2: 193
108B8327.CR2: 0.1546200767843ev (200.764705882353 - 197)
108B8328.CR2: 0.18067945827999ev (200.529411764706 - 195)
108B8329.CR2: 0.21202748185588ev (200.294117647059 - 192)
108B8330.CR2: 0.05213793341141ev (200.058823529412 - 202)
108B8331.CR2: 0.06745294370828ev (199.823529411765 - 200)
108B8332.CR2: -0.03123530744516ev (199.588235294118 - 206)
108B8333.CR2: 0.05064978091786ev (199.352941176471 - 200)
108B8334.CR2: 0.00853158892045ev (199.117647058824 - 202)
108B8335.CR2: 0.14161791784376ev (198.882352941176 - 192)
108B8336.CR2: 0.14326958944278ev (198.647058823529 - 191)
108B8337.CR2: 0.14499630687368ev (198.411764705882 - 190)
108B8338.CR2: 0.06662311624361ev (198.176470588235 - 195)
108B8339.CR2: 0.20509297546206ev (197.941176470588 - 185)
keyframe 108B8359.CR2: 108
108B8341.CR2: -0.21309330529631ev (155.5 - 168)
108B8342.CR2: -0.14492432038395ev (153 - 162)
108B8343.CR2: -0.1103221428348ev (150.5 - 158)
108B8344.CR2: -0.07466635013191ev (148 - 154)
108B8345.CR2: -0.13249812473191ev (145.5 - 155)
108B8346.CR2: -0.09706680509338ev (143 - 151)
108B8347.CR2: -0.06049204923087ev (140.5 - 147)
108B8348.CR2: -0.06276926647353ev (138 - 145)
108B8349.CR2: -0.00426795118219ev (135.5 - 140)
108B8350.CR2: 0.4452787722787ev (133 - 118)
108B8351.CR2: 0.50003196481029ev (130.5 - 114)
108B8352.CR2: 0.55728057295068ev (128 - 110)
108B8353.CR2: 0.61724619877515ev (125.5 - 106)
108B8354.CR2: 0.62414957439629ev (123 - 104)
108B8355.CR2: 0.63092798851693ev (120.5 - 102)
108B8356.CR2: 0.66657285856446ev (118 - 99)
108B8357.CR2: 0.6935730813898ev (115.5 - 96)
108B8358.CR2: 0.22152619213017ev (113 - 111)
#45
It is keyframe, yes. I have 4 of them. First and Last images of the sequence and 2 in the middle to handle some WB and other adjustments transition. So 8341 should have been matched to 8340 keyframe. Instead it's going to opposite direction.
#46
Yes. keyframe is 0 EV and next one is -1.2 EV.
#47
Here is what I'm talking about.
Check 8340 (keyframe) and 8341 transition:

This is not deflickered sequence. ETTR decided to go from 5 sec in 8340 to 4 sec in 8341.

This is what happens with defliker after after 1, 5, 10 iterations. All the same result:

Script just continue to underexpose 8341, instead of adding exposure.
#48
I'm running script on sequence of 600 images and it randomly can stuck "Processing xxxxx.CR2". Both in CS6 and CC. Any ideas how to avoid this?
#49
Quote from: dmilligan on October 31, 2013, 03:13:49 PM
Exactly, they're keyframes, the deflicker algorithm does not touch them, it uses them as a reference to make the images around them try to look the same. That's the whole point. When you mark a frame as a keyframe you are saying 'I want the other frames to look like this'.  The algorithm measures the specified percentile on the histogram of the keyframes, this gives a grayscale color level 0-255, for example 127. It then tries to match the histogram percentile of all the frames inbetween the keyframes to this target that is interpolated linearly (ramped) from one keyframe to the next. For example: frame 0 and 30 are keyframes. frame 0 measures the percentile at 100, frame 30 measure the percentile at 130. The script will then try to make the percentile of frame 1 at 101, frame 2 102, frame 3 103 and so on.

I guess I wanted to say that I get flicker around keyframes. Anyway, I'll play a little bit more with it and report back. Thx for clarification!
#50
Quote from: Joachim Buambeki on October 31, 2013, 01:55:30 PM
From my perspective it is working exactly as it should.
By adjusting the exposure of the keyframe and the deflickering adjust the surrounding frames to that target - if you don't touch your exposure this is your target. Just open all the keyframes in ACR and bring the exposure/brightness close to each other.

My keyframes looks like skipped in deflicker algorithm...