VideoHelp Forum

Try DVDFab and download streaming video, copy, convert or make Blu-rays,DVDs! Download free trial !
+ Reply to Thread
Page 3 of 3
FirstFirst 1 2 3
Results 61 to 74 of 74
Thread
  1. Originally Posted by dellsam34 View Post
    I de-interlace first before encoding, As far as setSAR to 10/11, I've tried it on few software players, my iPhone and LG TV and so far the 4:3 DAR is produced perfectly.
    Good Information!!! I'll play with that on some of my stuff.
    Quote Quote  
  2. You'll find that some devices ignore Sampling Aspect Ratio flags and will play such videos at the Frame Aspect Ratio or sometimes just stretch the video to full screen.
    Quote Quote  
  3. Originally Posted by jagabo View Post
    You'll find that some devices ignore Sampling Aspect Ratio flags and will play such videos at the Frame Aspect Ratio or sometimes just stretch the video to full screen.
    So the 100% solution is to resize to 640 x 480 with 1:1 Pixel Aspect Ratio?
    Quote Quote  
  4. Dinosaur Supervisor KarMa's Avatar
    Join Date
    Jul 2015
    Location
    US
    Search Comp PM
    Originally Posted by GrouseHiker View Post
    So the 100% solution is to resize to 640 x 480 with 1:1 Pixel Aspect Ratio?
    Permanently resizing it to 640x480 should take the guess work out of whether or not the video will be properly displayed everywhere.
    Quote Quote  
  5. Originally Posted by GrouseHiker View Post
    Originally Posted by jagabo View Post
    You'll find that some devices ignore Sampling Aspect Ratio flags and will play such videos at the Frame Aspect Ratio or sometimes just stretch the video to full screen.
    So the 100% solution is to resize to 640 x 480 with 1:1 Pixel Aspect Ratio?
    It depends. If you're planning to make DVDs or BDs 640x480 is not a legal frame size. If you're making MKV or MP4 videos it's ok.
    Quote Quote  
  6. Originally Posted by GrouseHiker View Post
    Does this approach work on all players: Apple, PC, Smartphone, YouTube?

    For YT, you'd probably want to double rate deinterlace, and upscale square pixel 960x720 or 1280x720 (pillarbox), because 720 height is the lowest resolution that offers 50p/59.94p (assuming this is still 8mm home video)
    Quote Quote  
  7. Originally Posted by poisondeathray View Post
    For YT, you'd probably want to double rate deinterlace, and upscale square pixel 960x720 or 1280x720 (pillarbox), because 720 height is the lowest resolution that offers 50p/59.94p (assuming this is still 8mm home video)
    Ok - 960x720 (4:3) with empty space displayed at the sides or 1280x720 (16:9 pillarbox with purposely added bars at sides) - both of which result in square pixels. - Correct?

    Would the 720 vertical best be achieved by capturing 480 and resizing or capturing 720 from the beginning?

    ... and yes - Video8 for me at the present...
    Quote Quote  
  8. Originally Posted by GrouseHiker View Post
    Would the 720 vertical best be achieved by capturing 480 and resizing or capturing 720 from the beginning?
    SD video doesn't have 720 lines. If your drive/software allows 720 lines in the saved video it's resizing the 480 (or 486) line cap. You would be much better doing that in AviSynth after QTGMC.
    Quote Quote  
  9. Originally Posted by jagabo View Post
    SD video doesn't have 720 lines. If your drive/software allows 720 lines in the saved video it's resizing the 480 (or 486) line cap. You would be much better doing that in AviSynth after QTGMC.
    I understand that, but via experiment I captured 1440 x 1080 (Hi8 in this case on the eBay camera). There must be some type of hardware interpolation going on, but I guess you're saying quality will turn out better by capturing 480 or 486 and resize via software.
    Quote Quote  
  10. Originally Posted by GrouseHiker View Post
    via experiment I captured 1440 x 1080 (Hi8 in this case on the eBay camera). There must be some type of hardware interpolation going on
    What makes you think it's hardware? It's more likely just software. In any case, there's nothing magic about hardware. It's inferior to software on almost all counts except one: speed. Hardware (while capturing) has to work in real time. It's designed to perform some simple matrix algorithms quickly. Not for general computing.

    Originally Posted by GrouseHiker View Post
    but I guess you're saying quality will turn out better by capturing 480 or 486 and resize via software.
    Cleanup, QTGMC, nnedi3 will get you better results than any hardware or capture driver/software.
    Last edited by jagabo; 27th May 2020 at 22:46.
    Quote Quote  
  11. Member dellsam34's Avatar
    Join Date
    Jan 2016
    Location
    Paris Ca, 92345 Mexico
    Search PM
    Resizing is always a compromise, The resizing will always be fractional not integer so some lines get crashed, some get doubled, Why don't you leave that to the display that was designed to handle this kind of work.
    Quote Quote  
  12. Originally Posted by GrouseHiker View Post
    Originally Posted by poisondeathray View Post
    For YT, you'd probably want to double rate deinterlace, and upscale square pixel 960x720 or 1280x720 (pillarbox), because 720 height is the lowest resolution that offers 50p/59.94p (assuming this is still 8mm home video)
    Ok - 960x720 (4:3) with empty space displayed at the sides or 1280x720 (16:9 pillarbox with purposely added bars at sides) - both of which result in square pixels. - Correct?
    Yes

    Would the 720 vertical best be achieved by capturing 480 and resizing or capturing 720 from the beginning?
    480, then proper deinterlacing, proper upscaling will give you slightly better results


    Originally Posted by dellsam34 View Post
    Resizing is always a compromise, The resizing will always be fractional not integer so some lines get crashed, some get doubled, Why don't you leave that to the display that was designed to handle this kind of work.
    Upscaling was suggested for the YT case . Currently, that's the only way you get full temporal resolution (59.94p), otherwise you discard 1/2 the motion samples (29.97p) . The cutoff is 720 height. It's been like that for a few years for YT, I don't think it's going to change
    Quote Quote  
  13. Originally Posted by dellsam34 View Post
    Resizing is always a compromise, The resizing will always be fractional not integer so some lines get crashed, some get doubled, Why don't you leave that to the display that was designed to handle this kind of work.
    Good point for non-YouTube video. I'll do a repeat of your CD test video and post for family to view as a test.
    Last edited by GrouseHiker; 28th May 2020 at 17:12. Reason: clarified non-YouTube
    Quote Quote  
  14. Back to the original subject of this post... I did more research and this is the result. If I understand this correctly, it turns out the "color under" system of recording on tape significantly reduces the color resolution. For Video8, Hi8, and VHS, sampling at 720 resolution with 4:2:2 pixels produces a huge (372%) oversampling rate. For sampling an NTSC-conforming signal, the oversampling is only 6%.

    These are simply numbers and calculations based on Nyquist-Shannon. From what I have read the Nyquist-Shannon minimum sampling rate of 2x peak frequency is too low, and oversampling is required.

    The numbers are interesting, but I am NOT making any recommendations here!

    Image
    [Attachment 53589 - Click to enlarge]
    Quote Quote  



Similar Threads