mpg2 color correction?

[sanlyn]

Darker is correct. Those settings adjust IRE black levels to match DVD sources. Darker = NTSC black (North America), lighter = match the higher production IRE black levels for many foreign technologies so that those recordings play with correct black levels thru North American and similar players. I think lighter is also there for some types of video cameras that use a higher-IRE base.

Tyhe best way to preserve detail in extreme lights or darks is do it in YUV before any further processing. The during-capture solution is to use a good proc amp (PA-1 is a favorite) or an AVT-8710 outboard TBC that has basic bright/contrast controls (the latter ain't the best way, but the AVT's rudimentary controls are clean and free of artifacts). You can always tweak later. Capturing with an incorrect color system/IRE base will create more problems than just "brightening" your video.

[jagabo]

Depending on your capture card, you might be able to set up a real time levels monitor. I've used a graph built with GraphStudio and opened it with DirectShowSource("name.grf") in AviSynth and added Histogram() or VideoScope(), then played that AVS script with a media player. I used GraphStudio to access the capture filter's proc amp settings while the video was playing. That gave real time feedback and made it easy to adjust the proc amp.

[sanlyn]

I don’t know why the LD/DVD colors look so much lighter on my PC’s monitor. I don’t expect them to look exactly the same, but the difference is quite a bit. Both have been adjusted using various test patterns so they should not be too far off. Maybe it’s just the inexpensive PC monitor.

Many factors are responsible for the differences. Monitor calibration isn't the only cause, but if you've tried to calibrate a monitor with test patternws (whether the monitor's cheap or not), it's not a very effecti9ve way of doing it. I drove myself crazy for years trying to do it that way (some would say that mentally I'm still not over it). This is the only way to do it:
http://www.tftcentral.co.uk/reviews/eye_one_display2.htm .
Last time I look at Amazon, these were selling for less than half price.

This thread has graphs showing the difference between using eyeball and test patterns to calibrate an inexpensive monitor, compared with using the proper tools: https://forum.videohelp.com/threads/335402-VHS-capture-critique?p=2083260&viewfull=1#post2083260

But monitor calibration, well, different story altogether. Plenty of info about that on the web.

[Mike99]

sanlyn -

I really don’t do any critical photo or video work so I never invested in any hardware monitor calibration equipment.

For my HDTV I’ve used the Avia calibration regular DVD, the THX black level calibration logo on a standard DVD, and also an HD test pattern recorded from one of the HD channels. They all seem to give me slightly different black level settings. I basically compromised between them and made sure I could see some detail in the dark shadow areas.

I previously looked into calibration equipment but did not want to spend the money. I‘ll have to check out stuff again.

[sanlyn]

True, monitor calibration can get sticky (and expensive), but the playback differences in Windows are the big issue and could have any number of causes. Frankly, I get some videos that display oddly with every media player on my XP machines. Most of those players, including PowerDVD, etc., are really antiquated, not up to the level of modern HDTV's. IMHO all of those players handle colorspaces and interlacing differently, and some do it more poorly than others. I usually end up playing almost everything in VirtualDub or VirtualDubMPEG2, converting to RGB32 with an AviSynth script instead of letting VDub do it, then resizing the VDub window. The only media that seem to play "normally" thru any Windows media players are commercial DVD's (not that they all look so great with those players, either, but at least they all tend to "look alike"). It's been a long time since I simply played a tape or DVD directly onto my PC. I'll have to try that and see what happens after all the changes over the years.

One thing I don't have a problem with: clips that I capture with VDub's capture window and videos recorded on my DVD recorder and transferred as MPEG's to my PC's seem to play normally thru every app. Go figure.

[jagabo]

I usually end up playing almost everything in VirtualDub or VirtualDubMPEG2, converting to RGB32 with an AviSynth script instead of letting VDub do it, then resizing the VDub window.

Try VLC with the output device set to Windows GDI. VLC will convert to RGB (rec.601) before sending video to the graohics card.

[Mike99]

I was just given 2 different DVDs of the video that I’ve been playing with. They are similar in quality but not exact. I don’t know their video sources but they play 91 mins vs. 94 mins, so that makes me suspect they were sped up from a 25fps PAL source.

But they are both sharper than the PAL VHS conversion I already had, and one is almost as sharp as the LD. This makes me think they did not come from a PAL VHS source. So I don’t know what to think about where they came from.

Their colors are also pretty good. Not identical to each other, but good. They also exhibit a bit of that color pumping similar to the PAL VHS conversion..

If you look at the short clip I previously provided, watch the scene where the grandmother walks in to see the boy. That LD version shows she has a black dress. All the other versions show a dark blue dress with black cuffs & buttons. You cannot even see the buttons on the LD version. Attached is a clip from one of the new DVDs. While this new DVD may lack some sharpness compared to the LD, it actually shows more image content such as the buttons.

One of the new DVDs (DVD#1) shows considerable blending when I pause the video during some action animation scenes. It sort of looks like a double exposure. I’m presuming this is a result of frame rate conversion, possibly from PAL to 29.97fps. (all these DVDs are 29.97fps). But the other new DVD (DVD#2) does not exhibit this dramatic blending. Could that mean DVD#2 might have come from a film print? If so, why would it be 4% faster?

[jagabo]

That clip shows blending of the chroma channels. Interlaced YV12 was treated as progressive YV12 at some point. It was probably filtered with VirtualDub which has that problem. The colors also drift -- probably captured with an ATI 650/750 capture device which has an overactive macrovision detection circuit. There's not a lot of motion in that clip but otherwise it appears to be normal 3:2 pulldown.

[sanlyn]

I wouldn't know how jagabo recognizes chroma blending (not that I disagree, though, I just don't know how to tell). Many colors differ from the LD I saw (the front apron now has a cyan cast, but it just might be correct because other collars and cuffs look white). In the other versions gramma's dress is a thick black, here it's blue-green and at least has some detail for a change -- even very dark clothing has some level of detail. If that dress color is the real color, the wardrobe designer should be arrested: that color doesn't do anything for the overall look of the scene, doesn't blend well with the other hues and keeps pulling your (my) eyes away from the "action"...assuming that what I see in the m2v are the original colors.

Shucks, I thought "digital" was supposed to avoid this kind of confusion.

[intracube]

Mike; What's the name of the film? Have you ever spotted it broadcast on TV before?

[Mike99]

jagabo -

Here's a few short clips of an animated scene.
The first one (Laser) is from the Japanese LaserDisc >DVD.
The second one (PAL) is from the PAL VHS > DVD.
The third one (DVD1) is from the first of the newly acquired DVDs. Original video source is Unknown.
The fourth one (DVD2) is from the second newly acquired DVD. Original video source is Unknown.

sanlyn -

The LaserDisc version is the only one where grandma’s dress looks black & has no buttons. The other 3 all show the dark bluish color & with buttons, so I’m presuming that is correct.

I thinks DVD2 is the best compromise.
The LD has poor-to-normal color, but is sharp.
The PAL VHS has normal-to-over saturated color and is low-rez.
It’s almost a toss-up between the 2 new DVDs. I’d probably have to watch them all the way through to see if anything strange shows up. DVD1 has more “blending”, but I did not notice any problem with this on my initial viewing. However I like the colors of DVD2 better & it appears a bit sharper than DVD1. At this point DVD2 is the winner.

[Mike99]

Mike; What's the name of the film? Have you ever spotted it broadcast on TV before?

Take a look at message #45.

I heard is was on the Disney Channel in the USA several years ago, but I never saw this. And supposedly a few years ago on TV in Australia and the UK.

[sanlyn]

Mike; What's the name of the film? Have you ever spotted it broadcast on TV before?

Take a look at message #45.

I heard is was on the Disney Channel in the USA several years ago, but I never saw this. And supposedly a few years ago on TV in Australia and the UK.

The movie is Song Of The South, Disney's beautiful Uncle Remus classic. Not allowed in the USA, Land of the Free and Home of the Brave, because it's about old-time black folk tales and has black characters and a beloved black story-teller and some of the most amusing and enlightening allegories and character studies ever found in American literature -- therefore making it Politically Incorrect in the U.S.A.

[Mike99]

Darker is correct. Those settings attempt to adjust for videos that use different IRE bases.

I’m a bit confused.

If “Watching a Japanese NTSC analog signal on an American TV would result in too much contrast with crushed blacks”, then why would I not choose the Lighter setting to "adjust for videos that use different IRE bases"?

I think the Darker does look better, but don't understand why. Shouldn't the Lighter adjust for 0 IRE?

[jagabo]

Well, it seems like "lighter" would be the correct setting when playing an NTSC-J disc. Capture something with some definite blacks from an NTSC-US disc then an NTSC-J disc with the same settings. Compare the black levels using something VideoScope() or Histogram().