VideoHelp Forum
+ Reply to Thread
Results 1 to 9 of 9
Thread
  1. Member
    Join Date
    Jan 2014
    Location
    Cairo
    Search PM
    Hello

    I need the filters settings that make photo number 1 same photo number 2
    I use MeGUI and StaxRip to encode the video


    Photo number 1 :

    https://forum.videohelp.com/images/imgfiles/PJsJ6mW.jpg


    Photo number 2 :

    https://forum.videohelp.com/images/imgfiles/3Ym2oMO.png
    Last edited by JoJoJoJo; 8th Feb 2014 at 02:59.
    Quote Quote  
  2. Ouch. That's insane. Did he use too much tanning cream? But if you must, somewhere around Tweak(sat=2.3, hue=-8, cont=1.33, bright=-10).
    Quote Quote  
  3. Member budwzr's Avatar
    Join Date
    Apr 2007
    Location
    City Of Angels
    Search Comp PM
    This is what Vegas comes up with as a "color match"

    Quote Quote  
  4. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    I'm afraid we need a little information here. Where did the top "good" image come from? Where did the ugly bottom image
    come from? How are you processing these videos? How did the "bad" image get so screwed up? Why do you use huge
    bold multi-colored screaming fonts?

    I don't know that anyone can make both images "look alike" exactly. It appears that Vegas tries to create an "averaging"
    histogram, which is not the way it's done -- especially when the two images are so different. In the "corrected" version
    above, saturation far exceeds that in the original and both people appear to have terminal liver disorders.

    I'm not sure what you mean by "make this one look like that one", especially when the "good" copy looks a lot better but
    could still use some work. Maybe it depends on the scene''s mood; the woman looks great, but the guy looks as if he
    needs a good meal. I don't know, maybe the scene is supposed to look that way. In any case I don't think you'll
    get an exact match.

    I'll explain what I did below, even if you won't understand a word of it and would never attempt such a thing yourself.
    You probably won't like it, because I don't think you really want the bad one to look like the good one. I think you really
    want something else, because most "average" viewers would prefer something closer to the badly distorted
    bottom image. But we'll see...

    First, begin by assuming that magic buttons won't work. Next, go on at a later time to get familiar with basic color
    theory, valid video luma and chroma levels, learn what gross over saturation looks like, learn what over juiced ugly
    contrast looks like, then learn to use RGB color controls, pixel value readers. and histograms. If you don't learn to use
    any of this stuff, you'll never find an "auto" filter that will work on ugly distorted screwed-up video for you.

    Below, the original image (you couldn't get a small video sample for us?):
    Image
    [Attachment 23407 - Click to enlarge]


    Above, the original is badly oversaturated, darks are dense and very nearly crushed out of existence by pumping
    contrast like a bandit.

    Below, Step 1 was to reduce saturation by 33% with the ColorMill filter. This allows one to see
    immediately that the original is too green, even if the woman's blue scarf glows like neon and bleeds off the screen:
    Image
    [Attachment 23408 - Click to enlarge]


    Below: Steps 2, 3 and 4 used three different layers of VirtualDub's gradation curves. Each layer is a tweak of the one
    above it. Not an exact match: the scarf is a little too blue, my version is a bit brighter than the original "good" image,
    and there's more blue in the walls. At least the people don't look ill or freaky. They seem to look human. None of
    this work does anything about the oversharpening or the banding and clumpy-grain compression artifacts in the
    "bad" image. Those will look far worse when the final video is encoded.
    Image
    [Attachment 23409 - Click to enlarge]


    The filters used were ColorMill and three chained copies of the gradation curve plugin. To determine their filter
    values I first took pixel readings from vital areas in the "good" version of the posted images. The "bad" image was
    mounting in AfterEffects, where I used a combination of ColorFinesse color wheels and curve filters until I could
    make most areas have the same or near pixel values that I had taken from the "good" image. I then used
    the ColorFinesse white level and black level controls to tweak those areas so that blacks looked black and
    skin tones looked a little more healthy. I then translated those filter values into the VirtualDub versions of
    those same filters, which are ColorMill.vdf and Gradation.vdf. Those two plugins and a free pixel sampler
    desktop app called Csamp are included in the attached .zip file. I also included a .vcf file from VirtualDub that
    contains the filter configuration and settings that I used for the new image. If you have the .vdf files loaded
    in your VirtualDub plugins folder, you can use that .vcf file to load those filters and the "bad" image directly
    into VirtualDub. This will allow you to see how the filters were set up in detail. With the image showing in
    VirtualDub, you can turn the filters off and on, one by one, to see their effect.

    One doesn't really need ColorFinesse for this work, it just helps to speed up the process. The pixel reader
    tells you the values that you basically want to correct for: you just slide the filter controls to get those effects.
    Work on the brights, then the darks, then the middle values. If you don't know what any of those terms
    mean, then....well, you really should know this stuff if you want to improve your videos. Otherwise, live with
    whatever comes out of the box. I takes time, I know. Watching it, you can learn a lot.
    Last edited by sanlyn; 19th Mar 2014 at 07:38. Reason: images posted don't look like the originals.
    Quote Quote  
  5. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    I'm afraid we need a little information here. Where did the top "good" image come from? Where did the ugly bottom image
    come from? How are you processing these videos? How did the "bad" image get so screwed up? Why do you use huge
    bold multi-colored screaming fonts?

    I don't know that anyone can make both images "look alike" exactly. It appears that Vegas tries to create an "averaging"
    histogram, which is not the way it's done -- especially when the two images are so different. In the "corrected" version
    above, saturation far exceeds that in the original and both people appear to have terminal liver disorders.

    I'm not sure what you mean by "make this one look like that one", especially when the "good" copy looks a lot better but
    could still use some work. Maybe it depends on the scene''s mood; the woman looks great, but the guy looks as if he
    needs a good meal. I don't know, maybe the scene is supposed to look that way. In any case I don't think you'll
    get an exact match.

    I'll explain what I did below, even if you won't understand a word of it and would never attempt such a thing yourself.
    You probably won't like it, because I don't think you really want the bad one to look like the good one. I think you really
    want something else, because most "average" viewers would prefer something closer to the badly distorted
    bottom image. But we'll see...

    First, begin by assuming that magic buttons won't work. Next, go on at a later time to get familiar with basic color
    theory, valid video luma and chroma levels, learn what gross over saturation looks like, learn what over juiced ugly
    contrast looks like, then learn to use RGB color controls, pixel value readers. and histograms. If you don't learn to use
    any of this stuff, you'll never find an "auto" filter that will work on ugly distorted screwed-up video for you.

    Below, the original image. It's over saturated, with dense dark nearly crushed nelow zero-black and contrast
    pumped-up like a bandit.

    Image
    [Attachment 23416 - Click to enlarge]


    Below, Step 1 was to reduce saturation by 33% with the ColorMill filter. This allows one to see
    immediately that the original is too green, even if the woman's blue scarf glows like neon and bleeds off the screen:
    Image
    [Attachment 23417 - Click to enlarge]


    Below: Steps 2, 3 and 4 used three layers of VirtualDub's gradation curves. Each layer is a tweak of the one
    above it. Not an exact match: the scarf is a little too blue, my version is a bit brighter than the original "good" image,
    and there's more blue in the walls. At least the people don't look ill or freaky. They seem to look human. None of
    this work does anything about the oversharpening or the banding and clumpy-grain compression artifacts in the
    "bad" image. Those will look far worse when the final video is encoded.(Left-click to enlarge)
    Image
    [Attachment 23418 - Click to enlarge]


    The filters used were ColorMill and three chained copies of the gradation curve plugin. To determine their filter
    values I first took pixel readings from vital areas in the "good" version of the posted images. The "bad" image was
    mounting in AfterEffects, where I used a combination of ColorFinesse color wheels and curve filters until I could
    make most areas have the same or near pixel values that I had taken from the "good" image. I then used
    the ColorFinesse white level and black level controls to tweak those areas so that blacks looked black and
    skin tones looked a little more healthy. I then translated those filter values into the VirtualDub versions of
    those same filters, which are ColorMill.vdf and Gradation.vdf. Those two plugins and a free pixel sampler
    desktop app called Csamp are included in the attached .zip file. I also included a .vcf file from VirtualDub that
    contains the filter configuration and settings that I used for the new image. If you have the .vdf files loaded
    in your VirtualDub plugins folder, you can use that .vcf file to load those filters and the "bad" image directly
    into VirtualDub. This will allow you to see how the filters were set up in detail. With the image showing in
    VirtualDub, you can turn the filters off and on, one by one, to see their effect. Note that most images in the
    forum viewer appear darker than they do in any other viewer I'm familiar with.

    One doesn't really need ColorFinesse for this work, it just helps to speed up the process. The pixel reader
    tells you the values that you basically want to correct for: you just slide the filter controls to get those effects.
    Work on the brights, then the darks, then the middle values. If you don't know what any of those terms
    mean, then....well, you really should know this stuff if you want to improve your videos. Otherwise, live with
    whatever comes out of the box. I takes time, I know.
    Last edited by sanlyn; 19th Mar 2014 at 07:39.
    Quote Quote  
  6. @sanyln: May be I'm missing something, but aren't you adjusting the wrong image?
    From the looks of it he wanted help to adjust the normal looking image to an over-saturated one.
    From a quick look it looks like you are adjusting the over-saturated one to look more normal.
    Quote Quote  
  7. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    That's what I think I said JoJoJoJo wanted. Anyway I'm too tired to fix the typing because of the O.P.'s huge
    JPG that overfills the page on my laptop screen. The O.P. can use any contrast and saturation filter to make all of
    his videos as ugly as he wants. What's so complicated? That's the wonder of digital processing; it's so easy to
    make your movies look uglier than VHS could ever hope to be. The trick for those of us who aren't blind is fixing
    what the half-blind are doing. I understand that ugly is what many people want. I just don't understand why. Perhaps
    they are afflicted with visual disorders ? ? ? JoJoJoJo never answered my question of how the ugly one was obtained.
    If he already has the ugly one, why deface #1 to look like it? Perhaps he has entered some sort of contest.

    I'd put a smiley here but I have to scroll all the way to the other side of this stretched window to get there. I'll
    just hit Alt+S. I don't know the keyboard shortcut for a smiley.
    Last edited by sanlyn; 19th Mar 2014 at 07:39.
    Quote Quote  
  8. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    #2 is also missing details. It's softer.
    Lots of posterizing too.

    FYI: Big fonts and big images is a turn-off to most forum users. (Me included.)
    Want my help? Ask here! (not via PM!)
    FAQs: Best Blank DiscsBest TBCsBest VCRs for captureRestore VHS
    Quote Quote  
  9. Member
    Join Date
    Jan 2014
    Location
    Cairo
    Search PM
    Sorry for delay in reply because I have some problem prevented me to reply

    Thank you all for your help
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!