VideoHelp Forum
+ Reply to Thread
Results 1 to 21 of 21
Thread
  1. Hey All,

    Like so many others, I thought I would try to convert some old VHS and 8mm tapes to digital recently. I have dusted off my late 90's Sony crt , JVC HR-S9500u, Canopus ADVC-110 and have gotten started with VirtualDub2 for the work. I have been reading quite a few posts about the nice quality one can get using QTGMC for deinterlacing the video but am quite at a loss as to how to get started with this.

    Anyway, I am wondering if any of you with some experience with this might be able to help get me started or point me in a good direction. I'm on a Win 10 system and so far the test captures that I have made look good (for VHS) outside of being a slightly darker and warmer on the lcd w/dv playback than what shows on the VHS direct to crt. Would like to clean the image up a bit more if possible.

    Any thoughts would be greatly appreciated.
    Quote Quote  
  2. Recently an app was released to help in using QTGMC for those not using AviSynth already. I know nothing about this app and have never used it. I can't recommend or not recommend it.

    https://forum.videohelp.com/threads/397026-Qtgmc-GUI-A-simple-encoder-for-your-Qtgmc-s...ighlight=qtgmc
    Quote Quote  
  3. There is no reason to use QTGMC. Any deinterlacing always degrades the video. Your display device can handle the deinterlacing.

    Instead of worrying about deinterlacing and QTGMC, work instead on learning the ideal settings on your JVC HR-S9500u to get the best possible capture. Make sure you used the "Edit" function, and turn off all of its fancy digital "restoration" circuitry.
    Quote Quote  
  4. Thanks to both of you for your replies.

    Despite the combing observed in the original digital copy, it does look sharper overall than a deinterlaced version. By what I have read I was of the opinion that QTGMC was some kind of miracle filter....perhaps not. I would still be interested to see what type of quality it can produce at some point as it has gotten much praise.

    Fwiw - I have been using the s-vid connector with the TBC, video calibration and Digital R3 on in the function menu for my tests. I have certainly noticed a difference between the S-vid and composite output with respect to certain intense color patterning. The image drawing is smoother in these areas and with what looks like more accurate color in others, like skin tones. It seems more natural to the way we see by taking the noise / interference information away. Sorry, not sure how to describe it. Anyway, if there is a better way to deal with those issues in editing software, I would prefer that as I can dial in the amount of reduction I would like and not have it baked in. For now, I'll do more tests with everything turned off and see how it goes.
    Quote Quote  
  5. Originally Posted by johnmeyer View Post
    There is no reason to use QTGMC. Any deinterlacing always degrades the video. Your display device can handle the deinterlacing.
    Does the display device handle de-interlacing without degrading the video?

    If the video has to be de-interlaced some time, why does it not make sense to use the best method possible, especially when it helps with compression. Aside from QTGMC's de-interlacing generally making the video more compressible than other deinterlacing methods, you can encode as progressive instead of interlaced.

    http://www.chaneru.com/Roku/HLS/X264_Settings.htm#tff

    tff
    Enable interlaced encoding and specify the top field is first. x264's interlaced encoding uses MBAFF, and is inherently less efficient than progressive encoding. For that reason, you should only encode interlaced if you intend to display the video on an interlaced display (or can't deinterlace the video before sending it to x264)


    Some de-interlacers (Yadif comes to mind) output the original scan-lines untouched when bob-deinteracing, so nothing is degraded from that perspective.QTGMC has lossless and semi-lossless modes. Nobody would use filters to subjectively improve the quality if it always degraded it, and QTGMC is just another type of filter.

    Standard Definition,
    The zip file below includes a fairly low quality source file, a version de-interlaced with QTGMC, and another de-interlaced with Yadif (something closer to hardware de-interlacing quality). Have a look and judge for yourself. For the QTGMC version, I enabled it's own noise filter, so there's probably some fine detail lost due to the denoising, but that's what happens with any denoising. Generally though, QTGMC's noise removal to blurring ratio is better than most dedicated noise filters.

    Deinterlacing Examples.zip (44.1MB)
    Quote Quote  
  6. The ONLY reason to deinterlace is if you have to re-size or do some other restoration which requires it. It otherwise should be avoided because it most definitely degrades the video. Pretty much any player can deinterlace, and certainly all TV sets do it, and they all now do it well.

    If the OP is seeing combing, then they simply have forgotten to turn on deinterlacing for their playback device. For instance, when using VLC, a popular computer playback software program, you have to turn on deinterlacing (which you can set as the default).

    The other reason to not do deinterlacing is that it takes a LOT of time, if you have a large video inventory, and also most neophytes screw it up. Ever wonder why you see so many screwed up videos online, in this forum, over at doom9.org, and in dozens of other places on the Internet?
    Quote Quote  
  7. I agree with hello_hello. The video is going to be deinterlaced somewhere along the line, even if it's the TV or player that does it. For viewing purposes you want to use the best deinterlacer available. For most material that's QTGMC. It's better than any TV's deinterlacer I've seen. Sure, keep a copy of the interlaced original in case better deinterlacers come along in the future. But for viewing now and in the near future use QTGMC. One exception is DVD production. DVD doesn't support progressive 60p or 50p. In that case you're better off leaving your video interlaced.
    Quote Quote  
  8. Originally Posted by johnmeyer View Post
    The ONLY reason to deinterlace is if you have to re-size or do some other restoration which requires it. It otherwise should be avoided because it most definitely degrades the video. Pretty much any player can deinterlace, and certainly all TV sets do it, and they all now do it well.
    As it turns out, I've also been discussing this subject in another thread. The sample I was using isn't great quality, but encoding it as interlaced seems to have quite an adverse effect. The encode doesn't de-interlace as well as the source.
    There's a screenshot here (second screenshot in that post). The source is included in the zip file attached to my post above. If you can encode it as interlaced and have it de-interlace nicely, please let me know what I'm doing wrong, and if you can encode it as interlaced and have the result look as good as the QTGMC sample when it's decoded, I'll have to totally rethink my current view of reality.

    Originally Posted by johnmeyer View Post
    The other reason to not do deinterlacing is that it takes a LOT of time, if you have a large video inventory, and also most neophytes screw it up. Ever wonder why you see so many screwed up videos online, in this forum, over at doom9.org, and in dozens of other places on the Internet?
    I've seen problems such as interlaced video encoded as progressive, or interlaced video that's been resized etc, and probably more often than I've seen video screwed up as a result of de-interlacing.
    Last edited by hello_hello; 12th May 2020 at 13:07.
    Quote Quote  
  9. I appreciate your additional feedback, links and screenshots hello_hello & jagabo. I'll consider all of what has been posted and continue to experiment / read-up.

    Cheers
    Quote Quote  
  10. Originally Posted by hello_hello View Post
    I've seen problems such as interlaced video encoded as progressive, or interlaced video that's been resized etc, and probably more often than I've seen video screwed up as a result of de-interlacing.
    I agree that QTGMC does a great job deinterlacing. However, lots of people use lesser tools, including throwing out fields, etc. Then, with QTGMC, it does denoising as well as deinterlacing, and denoising can (and does) introduce all sorts of its own artifacts, depending on the settings used.

    This is a "newbie" forum, and all of us, when we first start out, find very creative ways to screw up a video. Therefore ...

    "Less is more." -- Mies van der Rohe
    Quote Quote  
  11. I guess some of that would apply if this weren't a thread about using QTGMC, and you hadn't said "there is no reason to use QTGMC".

    I'd still like to know if interlaced encoding can introduce artefacts that make an encoded version harder to de-interlace, and therefore not de-interlace as well as the source, and therefore be a good reason for de-interlacing first. I suspect the very few times I have tried interlaced encoding that's been the case to some degree, and it appears to be for the sample I mentioned previously.
    Quote Quote  
  12. Originally Posted by hello_hello View Post
    I guess some of that would apply if this weren't a thread about using QTGMC, and you hadn't said "there is no reason to use QTGMC".
    There is no reason to deinterlace -- whether with QTGMC, or with something else-- when the goal is simply to watch the result.

    I have already stated why, but to make it clear: it won't look any better than simply letting the display or media player do the deinterlacing.

    Yes, I know that QTGMC is regarded as the "best" deinterlacer, but that really isn't the point (and may not be true). The point is that modern displays and players can do just as good a job, and in the case of deinterlacers which have access to the display hardware (e.g., a TV set), they can do things that you can't do in software.

    When using AVISynth, deinterlacing is only needed when applying temporal filters or re-sizing, and even when doing those things, there are other alternatives, many of which produce better results, or at least don't produce the nasty artifacts that are inherent in the QTGMC architecture.

    You can look at my extensive posts and tests over at doom9.org doing motion interpolation test with MVTools2, and showing the not-at-all-subtle artifacts which result. Since QTGMC is nothing more than a script, and since its relies entirely on MVTools2 to create a progressive version of a video without losing horizontal resolution, having advanced knowledge of those artifacts, and knowing that they not only cannot be avoided, but are absolutely going to happen, is what makes me so negative about using this tool.

    It is the same reason why, when preparing 15, 16, or 18 fps silent film for display that I usually advise to NOT use motion interpolation and instead either run the film at those native speeds, if the display permits, or use the old-fashioned telecine approach. You still get artifacts, but they are predictable and never weird or ugly.
    Quote Quote  
  13. Thats ideal World,
    but actually person who makes proper QTGMC content because he wants to and does it without major screw up, enough of bitrate, actually dodges a bullet of constant playback fail performances, encode failures, improper handling of interlaced content , tff, bff mix ups , apps do not handle interlaced content well at all, people later on will not do anything to make content progressive. That is some ideal case for all of us here, but not a regular person that discovers content later on. To draw a red line and make it progressive because one wants to and does it properly is not a bad thing at all.

    Both footage's , original capture avi and progressive mp4 should reside in same directories. As a back up anyway. But chances that someone will do something with that interlaced avi is minimal in real World.
    Quote Quote  
  14. Originally Posted by johnmeyer View Post
    [I have already stated why, but to make it clear: it won't look any better than simply letting the display or media player do the deinterlacing.
    Rubbish. Depending on the source, QTGMC can, and often does, look better than player de-interlacing.

    You should work on your quoting of posts, as you seem unable to quote and answer the question I've asked at least twice now.
    Why is it not likely encoding as interlaced will create artefacts that make the encoded video harder to de-interlace than the original, and therefore not de-interlace as well? I even supplied a little slice of mpeg2 from a DVD with my earlier samples where I'm sure that's the case. You could use it as an example to show me why I'm wrong.

    Originally Posted by johnmeyer View Post
    When using AVISynth, deinterlacing is only needed when applying temporal filters or re-sizing, and even when doing those things, there are other alternatives, many of which produce better results, or at least don't produce the nasty artifacts that are inherent in the QTGMC architecture.
    Well you've covered virtually every one of my encodes with temporal filters and resizing, but are you saying temporal filtering is okay unless it's a temporal filter that also de-interlaces, because that'd be what QTGMC is. What are the alternatives for better results with interlaced video? Are they not only better, but can also be used without effecting the quality of de-interlacing on playback.

    Originally Posted by johnmeyer View Post
    You can look at my extensive posts and tests over at doom9.org doing motion interpolation test with MVTools2, and showing the not-at-all-subtle artifacts which result. Since QTGMC is nothing more than a script, and since its relies entirely on MVTools2 to create a progressive version of a video without losing horizontal resolution, having advanced knowledge of those artifacts, and knowing that they not only cannot be avoided, but are absolutely going to happen, is what makes me so negative about using this tool.
    I've used QTGMC often enough to base my opinion on the result of it's de-interlacing, not the result of unrelated motion interpolation tests. It relies entirely on MVTools2 to create a progressive frame? That seems like something of an exaggeration. MVTools2 isn't used to bob the video. It's one of the tools used to clean up the progressive frames after they've been created though.
    And yes, there are occasions where QTGMC can produce interpolation type artefacts, but as a general rule it doesn't produce any I can see. Feel free to look at the sample I uploaded and point out the interpolation artefacts I've failed to notice though. If they absolutely happen it should be easy to take a screenshot of the QTGMC encode and a screenshot of the source being de-interlaced by your player and point out all the terrible things QTGMC did to the picture and where your player's de-interlacing did a better job.
    Quote Quote  
  15. Originally Posted by hello_hello View Post
    You should work on your quoting of posts, as you seem unable to quote and answer the question I've asked at least twice now.
    Too snarky. I no longer have interest in continuing with you.
    Quote Quote  
  16. Originally Posted by hello_hello View Post
    If they absolutely happen it should be easy to take a screenshot of the QTGMC encode and a screenshot of the source being de-interlaced by your player …….
    How does one take screenshots of a TV screen? I think that today's TVs have excellent deinterlacers (bobbers), keeping in mind that 1080i is still widely used for TV broadcasting. Can people see deinterlacing artefacts? Just wondering.
    Quote Quote  
  17. Originally Posted by johnmeyer View Post
    Originally Posted by hello_hello View Post
    You should work on your quoting of posts, as you seem unable to quote and answer the question I've asked at least twice now.
    Too snarky. I no longer have interest in continuing with you.
    You wouldn't look at the sample and tell me why my opinion is wrong, and pointing out you're ignoring a valid question is hardly being snarky. Claiming otherwise is just as a way to permanently avoid doing both.
    Quote Quote  
  18. Originally Posted by Sharc View Post
    Originally Posted by hello_hello View Post
    If they absolutely happen it should be easy to take a screenshot of the QTGMC encode and a screenshot of the source being de-interlaced by your player …….
    How does one take screenshots of a TV screen? I think that today's TVs have excellent deinterlacers (bobbers), keeping in mind that 1080i is still widely used for TV broadcasting. Can people see deinterlacing artefacts? Just wondering.
    I assumed he'd use a software player with his video card de-interlacing. Surely a video card should be able to match a TV for de-interlacing quality.
    I have no idea if people generally notice 1080i de-interlacing artefacts. If they can, most would probably assume "low quality" without knowing why. I thought 1080i was mostly used for sport, and I watch almost none if it, but if that's the case there's probably a reason why it's not used much for other types of broadcasts.
    Quote Quote  
  19. Originally Posted by jagabo View Post
    I agree with hello_hello. The video is going to be deinterlaced somewhere along the line, even if it's the TV or player that does it. For viewing purposes you want to use the best deinterlacer available. For most material that's QTGMC. It's better than any TV's deinterlacer I've seen. Sure, keep a copy of the interlaced original in case better deinterlacers come along in the future. But for viewing now and in the near future use QTGMC. One exception is DVD production. DVD doesn't support progressive 60p or 50p. In that case you're better off leaving your video interlaced.
    hello jagabo
    what would be the best tool or method to reduce noise?
    Quote Quote  
  20. Originally Posted by Sharc View Post
    Originally Posted by hello_hello View Post
    If they absolutely happen it should be easy to take a screenshot of the QTGMC encode and a screenshot of the source being de-interlaced by your player …….
    How does one take screenshots of a TV screen? I think that today's TVs have excellent deinterlacers (bobbers), keeping in mind that 1080i is still widely used for TV broadcasting. Can people see deinterlacing artefacts? Just wondering.
    I see them all the time
    Quote Quote  
  21. Originally Posted by Takashi View Post
    what would be the best tool or method to reduce noise?
    If you're already using QTGMC try adding its built in noise reduction feature:

    Code:
    QTGMC(EZDenoise=1.0, DenoiseMC=true) # higher value for more noise reduction
    It's not a great noise reducer and I wouldn't use it much above 1.0 or 2.0. Other NR filters you can try are SMDegrain(), TemporalDegrain(), MCTemporalDenoise() -- all motion compensated (and slow). Some simpler ones are HQDN3D(), fft3dFilter(), etc.
    Quote Quote  



Similar Threads

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