VideoHelp Forum
+ Reply to Thread
Page 1 of 2
1 2 LastLast
Results 1 to 30 of 35
Thread
  1. I am encoding a few of my DVDs to H.264 (x264) and want keep the original/correct video width/height/aspect ratio using HandBrake. Have I selected the correct settings as per the images?

    Video: MPEG2 Video 720x480 (4:3) [Video]
    Image
    [Attachment 51549 - Click to enlarge]


    Video: MPEG2 Video 720x480 (16:9) [Video]
    Image
    [Attachment 51550 - Click to enlarge]


    Video: MPEG2 Video 720x576 (4:3) [Video]
    Image
    [Attachment 51551 - Click to enlarge]


    Video: MPEG2 Video 720x576 (16:9) [Video]
    Image
    [Attachment 51552 - Click to enlarge]
    Quote Quote  
  2. aBigMeanie aedipuss's Avatar
    Join Date
    Oct 2005
    Location
    666th portal
    Search Comp PM
    dvd mpeg-2 is a non-square pixel format. there is no 1 to 1 transfer to square pixel format h264. some players might display what you created ok but others won't.
    --
    "a lot of people are better dead" - prisoner KSC2-303
    Quote Quote  
  3. Member netmask56's Avatar
    Join Date
    Sep 2005
    Location
    Sydney, Australia
    Search Comp PM
    If you are not also shrinking the file size ie file size in = file size out then you could use MakeMKV.
    SONY 75" Full array 200Hz LED TV, Yamaha A1070 amp, Zidoo UHD3000, BeyonWiz PVR V2 (Enigma2 clone), Chromecast, Windows 11 Professional, QNAP NAS TS851
    Quote Quote  
  4. Originally Posted by netmask56 View Post
    If you are not also shrinking the file size ie file size in = file size out then you could use MakeMKV.
    Just trying to keep the DVD source image as original as possible so everything untouched, size does not matter.

    Originally Posted by aedipuss View Post
    dvd mpeg-2 is a non-square pixel format. there is no 1 to 1 transfer to square pixel format h264. some players might display what you created ok but others won't.
    What will MakeMKV do to the non-square pixels? From what I can tell it looks like it stripped it 1:1 from the DVD, this should then resize accordingly to 4k TV/Monitor? It puts it in the mkv container and leaves the codec as MPEG-1/2 can this be set to H.264? Also can MakeMKV exclude chapter markings? I'd rather not use MKVToolNix as it lowers the bitrate slightly even if chosen No Compression
    Last edited by ss2012; 18th Jan 2020 at 21:53.
    Quote Quote  
  5. Originally Posted by ss2012 View Post
    What will MakeMKV do to the non-square pixels? From what I can tell it looks like it stripped it 1:1 from the DVD,
    It will only remux the video and audio streams into an MKV container. It doesn't change anything else, the quality isn't changed.

    Originally Posted by ss2012 View Post
    this should then resize accordingly to 4k TV/Monitor?
    It depends on the player. Some will follow the MPEG 2 aspect ratio flags, some will not.

    Originally Posted by ss2012 View Post
    It puts it in the mkv container and leaves the codec as MPEG-1/2 can this be set to H.264?
    No, the whole point is not to change the video in any way. If you reencode to h.264 you will lose some quality. Maybe not visibly if you use sufficient bitrate and know what you're doing.

    Originally Posted by ss2012 View Post
    Also can MakeMKV exclude chapter markings? I'd rather not use MKVToolNix as it lowers the bitrate slightly even if chosen No Compression
    MakeMKV does not lower the video or audio bitrate. The MKV container has less overhead than VOB or MPG -- that's why the file come out a little smaller. The same will happen with MakeMKV.
    Quote Quote  
  6. What will MakeMKV do to the non-square pixels?
    Nothing, there's no transcoding involved to all parameters stay the same.

    From what I can tell it looks like it stripped it 1:1 from the DVD, this should then resize accordingly to 4k TV/Monitor?
    The original quality is 100% preserved so it's the best you can get. Any decent standalone player should recognize the aspect ratio (4:3 or 16:9) and display the picture accordingly.

    It puts it in the mkv container and leaves the codec as MPEG-1/2 can this be set to H.264?
    If you want H.264 you want transcoding, which implies a loss of quality, which can be practically negligible or significant depending on the quality settings. Usually with H.264 encoding the size can be reduced by half with no noticeable quality loss, but it depends on the content, and the level of tolerance / scrutiny of the viewer (some people apparently still think that a 700MB Xvid DVD rip is good enough, as there are millions of them still floating around, and still being released I suppose in 2020). Those are two different approaches. If you want a real backup, just extract the whole DVD as ISO or MKV ; if you want a smaller file for convenience, transcode it to H.264 with a -crf setting (Constant Rate Factor) that seems satisfying to you in terms of quality / size compromize (usually -crf 20 is considered very good, and on average it produces a ~2GB file from a 2h 720x480 movie ; a smaller -crf value will preserve the quality better but produce bigger files, while a higher -crf will produce smaller files but may introduce a noticeable quality loss).

    I'd rather not use MKVToolNix as it lowers the bitrate slightly even if chosen No Compression
    Where did you see that “No compression” setting ? As far as I know MKVToolNix never compresses the streams but only muxes them into a MKV container. The MKV container is said to have less “overhead” than other containers, especially MPEG/VOB, meaning, the size of the structures of the container itself (header, index and whatnot) is small, which must explain the difference in size. But the actual streams (video, audio, subtitles if any) should be identical.


    EDIT : “jagabo” was quicker ! I hadn't seen the post above while typing mine, at least there are few potentially confusing discrepancies...
    Quote Quote  
  7. Originally Posted by abolibibelot View Post
    I'd rather not use MKVToolNix as it lowers the bitrate slightly even if chosen No Compression
    Where did you see that “No compression” setting ? As far as I know MKVToolNix never compresses the streams but only muxes them into a MKV container. The MKV container is said to have less “overhead” than other containers, especially MPEG/VOB, meaning, the size of the structures of the container itself (header, index and whatnot) is small, which must explain the difference in size. But the actual streams (video, audio, subtitles if any) should be identical.
    Image
    [Attachment 51555 - Click to enlarge]



    Thanks for the response @jagabo and @abolibibelot.

    I will go with re-encoding to H.264 as Plex is transcoding the MPEG-1/2 anyway.

    I would just like to confirm I have selected the correct dimensions for each video source in the images to keep it as close to original or the correct width/height? I will play around with the -crf.
    Quote Quote  
  8. No good idea here since I usually let the program, like Handbrake, handle the encodes and aspect ratios.

    One site notes: https://www.vegascreativesoftware.info/us/forum/what-are-the-correct-pixel-aspect-ratios--39808/
    "
    Jřran Toresen wrote on 3/28/2005, 4:32 AM
    Various applications reports different pixel aspect ratios for PAL 4:3 and 16:9 DV video. For example, Vegas define the pixel aspect ratio PAR as 1.0926 and 1.4568 for PAL DV 4:3 and 16:9 video respectively. VirtualDub reports equal numbers expressed in fractional form: 59:54 and 118:81. Adobe Premiere Elements and Pinnacle Liquid Edition define / report the pixel aspect ratio PAR as 1.067 and 1.422 for PAL 4:3 and 16:9 video respectively.

    The calculation of the latter numbers is straightforward and easy to understand. Define the Display Aspect Ratio DAR as the ratio between the width and the height of the physical display (TV).

    (1) DAR(4:3) = 4 / 3 = 1.3333
    (2) DAR(16:9) = 16 / 9 = 1.7778

    Define the Frame Aspect Ratio FAR as the ratio between the number of horizontal and vertical pixels in each video frame. The Frame Aspect Ratio is identical for both standard (4:3) and widescreen (16:9) video.

    (3) FAR(PAL) = 720 / 576 = 1.25

    The Pixel Aspect Ratio PAR for PAL 4:3 and 16:9 video can be calculated as the ratio between the Display Aspect Ratio and the Frame Aspect Ratio, given the definitions above.

    (4) PAR(PAL 4:3) = DAR(4:3) / FAR(PAL) = (4/3) / (720/576) = 1.067
    (5) PAR(PAL 16:9) = DAR(16:9) / FAR(PAL) = (16/9) / (720/576) = 1.422

    These numbers are the pixel aspect ratios reported by Adobe Premiere Elements and Pinnacle Liquid Edition."

    ....

    In any case, a simple rip to MKV using MakeMKV to retain the original video quality, or a decrypted rip to ISO to retain the original video quality and DVD menus, are the two easy ways to get a 1:1 copy of the main movie without degredation in quality. Another way is a decrypted rip of VOBs to MPEG-2 (no encoding, only format change) using MPG2CUT2.

    Since you're not using KODI or another media server that handles ISO/VOB natively, you'll need to convert to MP4 since Plex doesn't even support MPG AFAIK.
    The basic Handbrake default settings for 480 30p, adjusted to not crop unless letterboxed and to CFR to match the original source, and a quality setting of about 18 (whether H.264 or H.264 Quicksync), Level High/5.2, will generally get you a decent/nice rip without too much work.
    (Naturally, if you set the auto encodes to pass-through rather than encode to AAC, you'll retain the highest quality audio. Who knows whether Plex will support it all - PCM, Dolby, DTS etc.)

    Not that much work if it's only a few dozen DVDs, but a ton of work once you reach the hundreds+ - at that point, switching to Kodi, etc that natively support ISO rips is the best move since there's no re-encoding necessary, so that saves on CPU processing time.
    Quote Quote  
  9. Originally Posted by ss2012 View Post
    Originally Posted by abolibibelot View Post
    I'd rather not use MKVToolNix as it lowers the bitrate slightly even if chosen No Compression
    Where did you see that “No compression” setting ? As far as I know MKVToolNix never compresses the streams but only muxes them into a MKV container. The MKV container is said to have less “overhead” than other containers, especially MPEG/VOB, meaning, the size of the structures of the container itself (header, index and whatnot) is small, which must explain the difference in size. But the actual streams (video, audio, subtitles if any) should be identical.
    Image
    [Attachment 51555 - Click to enlarge]
    That has nothing to do with the audio or video. If you hover over the line you'll see the context help:

    Image
    [Attachment 51559 - Click to enlarge]


    Currently only certain subtitle formats are compressed with the zlib algorithm.
    Quote Quote  
  10. I don't have time to read this whole thread, but if you're trying to encode a 720x480 0.909 PAR DVD to a square pixel format, I would use 656x480. I've done that dozens of times over the years.
    Quote Quote  
  11. I don't have time to read this whole thread, but if you're trying to encode a 720x480 0.909 PAR DVD to a square pixel format, I would use 656x480. I've done that dozens of times over the years.
    Why 0.909, isn't it supposed to be exactly 4:3, hence a 0.888 pixel aspect ratio ? Or is it the “ITU” ratio, with a few extra pixels left and right beyond the actual picture ?
    Quote Quote  
  12. Originally Posted by abolibibelot View Post
    Why 0.909, isn't it supposed to be exactly 4:3, hence a 0.888 pixel aspect ratio ? Or is it the “ITU” ratio, with a few extra pixels left and right beyond the actual picture ?
    It's because of the discrepancy between the ITU spec and the MPEG 2 spec. MPEG 2 is very clear. Only the DAR is flagged (and of the four possible values only 4:3 and 16:9 are valid for DVD), and the DAR refers to the entire frame (unless overridden by a sequence_display extension). The ITU spec puts the DAR in a 704x480 or 704x4x576 portion of the 720 pixel wide capture (actually those are approximations, the nearest even value). Most DVDs made from analog video sources are captured by ITU compliant devices and the difference between the two specs is ignored. In my experience DVD players are schizophrenic about this, following the ITU spec at the s-video and composite outputs, but upscaling the full frame at the HDMI outputs.

    If you're a real stickler for technical correctness you have to examine the video and look for hints that it's an ITU cap : fuzzy video (ie, analog tape source), ~8 pixel wide pillarbox bars left and right, head switching noise at the bottom of the frame, hard pulldown, field blending from a PAL/NTSC conversion, etc.

    But the reality is that nobody can really see the ~2 percent AR difference so it's not worth worrying about.
    Quote Quote  
  13. Thanks for all the replies, still a little confused about what dimensions to select in Handbrake but because I have about 488 documentaries to rip I tested MakeMKV and it plays via Plex to my TV without transcoding however only in the browser on a laptop or mobile it transcodes.

    MakeMKV: About 3 min per documentary
    HandBrake 3 hours 20min per documentary
    Quote Quote  
  14. MakeMKV: About 3 min per documentary
    HandBrake 3 hours 20min per documentary[/QUOTE]

    No idea about the aspect ratio part since handbrake can usually handle that fine automatically on 4:3 or 16:9 sources for me.
    I only need to specify automatic crop when it's 16:9 letterboxed to get rid of the black bars on top and bottom when 16:9 is shown in a 4:3 frame.

    ....

    3 hours+ for a handbrake encode!?!?!?!
    "My God- Jim!!" - to use a Star Trek quote.
    Video Tab and set the encoder to H.264 QSV Q15 to Q18, high, 5.2, quality. You should be able to encode at 200+ fps on any modern Intel CPU that supports Quicksync video for super-fast mp4 encodes. Like under 15 minutes per dvd.

    Most times, the default H.264 encoder gets you the highest quality with the smallest file size, but as long as you can accept a slightly larger file size, the Intel QS encoder performs very well.
    Quote Quote  
  15. Originally Posted by johnmeyer View Post
    I don't have time to read this whole thread, but if you're trying to encode a 720x480 0.909 PAR DVD to a square pixel format, I would use 656x480. I've done that dozens of times over the years.
    656 loses about 10% of the resolution versus 720, and since dvds are already low resolution, this further degrades the image quality vs retaining the original dimensions of the digital image from the dvd.

    PCs etc have no trouble at all upscaling from 720x to fit higher resolution monitors and tvs, and does it better with higher resolution source files (720 better than 656 better than 320 etc).
    Quote Quote  
  16. Originally Posted by babygdav View Post
    No idea about the aspect ratio part since handbrake can usually handle that fine automatically on 4:3 or 16:9 sources for me.
    I only need to specify automatic crop when it's 16:9 letterboxed to get rid of the black bars on top and bottom when 16:9 is shown in a 4:3 frame.

    ....

    3 hours+ for a handbrake encode!?!?!?!
    "My God- Jim!!" - to use a Star Trek quote.
    Video Tab and set the encoder to H.264 QSV Q15 to Q18, high, 5.2, quality. You should be able to encode at 200+ fps on any modern Intel CPU that supports Quicksync video for super-fast mp4 encodes. Like under 15 minutes per dvd.

    Most times, the default H.264 encoder gets you the highest quality with the smallest file size, but as long as you can accept a slightly larger file size, the Intel QS encoder performs very well.
    lol I believe it was because I set the Encoder Preset to 'Very Slow', Medium should be okay?

    Yeah with these documentaries I'm not sure if I should use auto crop as the sides keep adding black bars on the sides when the scene changes so it looks odd. With Auto crop it minus the width/height but I put it back to the source size should I not?

    Image
    [Attachment 51592 - Click to enlarge]

    Image
    [Attachment 51593 - Click to enlarge]
    Quote Quote  
  17. If you don't want to deal with the custom calculations and crop for each disc (likely small variations), then custom crop to 0 all around and encode as you've set before to 720x480 with custom par settings. It's a LOT of discs, and I've seen borders vary a lot across discs.

    ...

    Instead of MAIN profile, HIGH.
    Also, constant frame rate, not variable.

    ...

    You're using the default h.264 (x264) encoder, so with 2 pass, it'll be slow no matter what.
    Active turbo first pass - generally works well for slow, high quality encodes, but faster by a bit. (Still hours)

    ...

    Otherwise, use the H.264 (qsv) encoder option with two pass off (the Quality slider mostly controls the Quicksync encode quality, so slide to 15-18. Encoder Preset then will be Quality, 5.2, High.)
    Using quicksync reduces the 3+ hour encode to less than 15 minutes. (Only if you have an Intel cpu - sorry amd users.)

    Image
    [Attachment 51594 - Click to enlarge]
    Image
    [Attachment 51595 - Click to enlarge]
    Image
    [Attachment 51600 - Click to enlarge]


    For example, if I was going to rip a 16x9 video, I'd use settings like this, with Quicksync to accelerate the encode. (Notice for the 49 minute video, it's reporting a quick 9 minutes to RIP and encode!
    Last edited by babygdav; 21st Jan 2020 at 03:19.
    Quote Quote  
  18. Edit: you need 'Intel Skylake' and its greyed out in Preferences/Video

    Its a bit odd as I don't have QuickSync

    Image
    [Attachment 51603 - Click to enlarge]
    Quote Quote  
  19. 656 loses about 10% of the resolution versus 720, and since dvds are already low resolution, this further degrades the image quality vs retaining the original dimensions of the digital image from the dvd.
    It's debatable. With the same bitrate (if setting a higher CRF value to compensate) and ~10% less pixels, the compression quality can be better.

    Yeah with these documentaries I'm not sure if I should use auto crop as the sides keep adding black bars on the sides when the scene changes so it looks odd. With Auto crop it minus the width/height but I put it back to the source size should I not?
    That's the kind of things that made me wary of using Handbrake long ago. When I did I made sure to set the crop values to 0, unless I knew that a particular video would benefit from being cropped by specific values, which I had duly verified. I'm not sure if I understood correctly what you meant here, but no, you shouldn't use auto crop, and generally speaking, unless the whole video has black bars of the same size all along, disabling all cropping should be preferred.
    Once I criticized the conversion someone had made of a series of documentaries on the Monty Pythons broadcast on television, using Handbrake and relying on its default auto crop feature. Such documentaries alternate between contemporary footage shot in 16:9 aspect ratio and archive footage shot in 4:3 aspect ratio (with black bars on the sides when it's done right — which is certainly not as odd as stretching the frame to the 16:9 ratio, or cropping it vertically by 25%, or a mix of both, as can be seen in many programs featuring butchered archives), and Handbrake, unless its analysis method has vastly improved since then, usually chooses to crop to the size of the 4:3 frame. So it looked like this :
    Click image for larger version

Name:	MP_AlmostTheTruth_3 [réencodé avec HandBrake, 'cropping' sauvage].mkv 42'26 vlcsnap-2013-11-26.png
Views:	94
Size:	1.03 MB
ID:	51601
    and should have looked like this :
    Click image for larger version

Name:	monty_python_toute_la_verite_ou_presque 3-6 {eMule}.flv 44'36 vlcsnap-2013-11-26-22h34m51s69.png
Views:	96
Size:	556.8 KB
ID:	51602
    And the resolutions were different between the episodes... Despite the fact that his version was in better quality than the other one (which had been downloaded from the channel's “replay” service), this egregious defect made it unwatchable. (Yet the dude was adamant that he had done so on purpose, to get a good quality / size compromize at the expense of the “modern” shots, considering that the archive images were the most important — even though the 16:9 content included many still pictures as the one shown on the screenshots above.)
    Quote Quote  
  20. Using Quicksync reduces the 3+ hour encode to less than 15 minutes. (Only if you have an Intel cpu - sorry amd users.)
    Wow, that's impressive... But how does the quality and size compare to a 3h encode with x264 ?
    Quote Quote  
  21. here is a 5 minute sample of the vob that I'm dealing with.

    From the doco it seems VFR is a better choice than CFR?

    Even though I selected VFR Media Info says Constant
    Quote Quote  
  22. Originally Posted by ss2012 View Post
    Edit: you need 'Intel Skylake' and its greyed out in Preferences/Video

    Its a bit odd as I don't have QuickSync

    Image
    [Attachment 51603 - Click to enlarge]
    Preferences > Video > Allow use of the Intel Quicksync encoders (check/on).

    ....

    If you have an nvidia card, enable nvidia nvenc encoding. Try and compare to see what works best for you.
    Last edited by babygdav; 21st Jan 2020 at 05:32.
    Quote Quote  
  23. Originally Posted by abolibibelot View Post
    Using Quicksync reduces the 3+ hour encode to less than 15 minutes. (Only if you have an Intel cpu - sorry amd users.)
    Wow, that's impressive... But how does the quality and size compare to a 3h encode with x264 ?
    Quality "The eighth annual MPEG-4 AVC/H.264 video codecs comparison showed that Quick Sync is comparable to x264 superfast preset in terms of speed, compression ratio and quality (SSIM);[5] tests were performed on an Intel Core i7 3770 (Ivy Bridge) processor." And, the Quality has improved since the 3rd generation cpus.

    https://en.m.wikipedia.org/wiki/Intel_Quick_Sync_Video

    Size-wise, qsv encodes can be the same size as x.264 encodes, but will be worse in quality (at the best encode settings). Or, equal quality but the qsv encode will be slightly larger in file size.

    On the latest Intel cpus that have H.265 Quicksync, simply switching to the better encode format allows for better quality encodes at the same file size vs x.264 h.264 encodes...But far faster.

    ........

    Now, what this all really asks is
    IS encoding MPEG-2 dvd rips to H.264/h.265 going to give me ACCEPTABLE encodes?

    NOT ALWAYS!

    If the dvd source was filmed in digital, mastered in digital, like the latest 4K movies, there will be essentially no noise or artifacts of note on the dvd video. As crisp and clean as you can get, so any h.264/h.265 encode will likely give you a great result.

    If the dvd source was filmed in analog cameras, mastered on take, dropped to dvd digitally, you'll see lots of analog noise and artifacts in the original. (E.g. Many concerts from the 80, 90, 2000s) ANY attempt to try and convert will likely result in crap files at the usual expected compression rates (E.g. 2GB h.264 encode of a 4gb mpeg-2 master). Qsv, h.264 x.264, h.265... None will be able to handle the noise well - You'll either lose lots of detail and get a crap encode, or get a huge (often bigger than original) encode that still isn't nice looking.

    Mpeg-2 to h.264 on these types of dvds are not recommended - keep the video untouched in mpeg-2 instead (or vob, or iso).

    Trust me on this - if qsv can't get a nice encode at quality 12-15 (normally 15-18 is fine), this tells you not to bother encoding. Keep in mpeg-2. I've tried all sorts of settings and encoders on such older dvds, and nothing works better than simply keeping it in mpeg-2 format as a straight disc rip.

    ...

    Naturally, if you've got an interlaced source, it's even trickier and tougher to get a nice h.264 encode, so again, if qsv at the 12-15 quality setting with default handbrake interlace detection and deinterlacing settings isn't working well, don't bother encoding - keep it in the original format.

    ...

    Most modern discs rip fine. And qsv is a godsend in bluray rips. But older discs might not survive the encode.
    Quote Quote  
  24. see the comparison here of makemkv vs handbrake. I'm finding it hard to tell which has lost detail, maybe someone with keener eyes can tell.
    Quote Quote  
  25. Originally Posted by babygdav View Post
    Originally Posted by johnmeyer View Post
    I don't have time to read this whole thread, but if you're trying to encode a 720x480 0.909 PAR DVD to a square pixel format, I would use 656x480. I've done that dozens of times over the years.
    656 loses about 10% of the resolution versus 720, and since dvds are already low resolution, this further degrades the image quality vs retaining the original dimensions of the digital image from the dvd.

    PCs etc have no trouble at all upscaling from 720x to fit higher resolution monitors and tvs, and does it better with higher resolution source files (720 better than 656 better than 320 etc).
    Yes, 656 is about 10% fewer pixels than 720.

    But no, it does not lose 10% of the resolution.

    Why?

    Because the 720 encode uses rectangular pixels and the 656 encode uses square pixels. There is zero information in the long horizontal direction of the rectangular pixel and therefore no information (resolution) to lose.
    Quote Quote  
  26. aBigMeanie aedipuss's Avatar
    Join Date
    Oct 2005
    Location
    666th portal
    Search Comp PM
    dvd spec 720x480 can either be 4:3 or 16/9. so when converting to square pixel formats like h264 you should choose 640x480 for 4:3 material or 854x480 for 16/9. they are easy, encoder friendly and close approximations of the mpeg-2 sizes.

    abolibibelot - your post #19 is a 16/9 source and needs to be re-encoded that way. you did it to 4:3 that's why a 1/3 of the video frame is missing.
    --
    "a lot of people are better dead" - prisoner KSC2-303
    Quote Quote  
  27. Originally Posted by ss2012 View Post
    see the comparison here of makemkv vs handbrake. I'm finding it hard to tell which has lost detail, maybe someone with keener eyes can tell.
    The differences can be subtle.
    I usually take screen shots at 1:1 of source and output.
    Open them in a photo editor like Photoshop or gimp.
    Layer them together in one file.

    Apply an invert/negative filter to one layer, the top one.

    What will result is the difference between the two, seen clearly.

    ...

    Note that single frame comparisons does not show differences in time - This often causes worse artifacts like edge strobing, aliasing, etc , than you see in a single frame.

    Also, if you have fine text or details like a hair, it's more obvious. Detail in the sand of a beach shot. Etc
    Quote Quote  
  28. Yes, 656 is about 10% fewer pixels than 720.
    But no, it does not lose 10% of the resolution.
    Why?
    Because the 720 encode uses rectangular pixels and the 656 encode uses square pixels. There is zero information in the long horizontal direction of the rectangular pixel and therefore no information (resolution) to lose.

    Huh?

    An ntsc source material like broadcast, dv tape, vhs, dvd, are all targeted originally to be displayed on a 4:3 aspect tv using a 720x480 digital frame size (rectangular pixels, stretching automatically occurs during display).

    All computers think in square pixels, so a 4:3 aspect ratio image is squished to a 1:1 aspect ratio 640x480 frame size. When a 1:1 640x480 is displayed on a monitor, no squishing occurs, but if fed to a tv, it is stretched to fit the 4:3 aspect ratio tv set instead of the 1:1 computer monitor.

    ....

    The stretching of ntsc source material or computer generated source material when displayed outside the native display format is automatic and always occurs.

    What shouldn't is the native resolution and detail stored in each pixel until display.

    For example, a ntsc source video with 1-pixel wide vertical color stripe, followed by another color, and another all the way across encodes 720 different colored stripes.

    Reducing that resolution by shrinking the horizontal resolution to 6xx DOES NOT RETAIN ALL original 720 colors. The stretching and squishing of pixels to fit the screen aspect ratio doesn't change the fact you're now missing almost 100 colors.

    Now, there's some stretching and squishing artifacts occurring when displaying ntsc source material on a pc, or computer graphics on a tv - BUT, talking about ntsc source material, all 720 colors should be present in any encode or rip in the pc file when displayed 1:1 without any aspect ratio adjustments.

    https://helpx.adobe.com/premiere-pro/using/aspect-ratios.html#Frameaspectratio
    Last edited by babygdav; 21st Jan 2020 at 14:18.
    Quote Quote  
  29. Originally Posted by aedipuss View Post
    dvd spec 720x480 can either be 4:3 or 16/9. so when converting to square pixel formats like h264 you should choose 640x480 for 4:3 material or 854x480 for 16/9. they are easy, encoder friendly and close approximations of the mpeg-2 sizes.
    There's also 4:3 letterboxed, in which case a crop of the top and bottom black bars helps reduce encoded file size since there's no video information of importance there. 1:1 pixel size depends on the crop and aspect ratio of the letterboxed video.

    (Bluray supports 1:1 in addition to 4:3 and 16:9)
    Quote Quote  
  30. Originally Posted by babygdav View Post
    An ntsc source material like broadcast, dv tape, vhs, dvd, are all targeted originally to be displayed on a 3:2 aspect tv...
    Do you just make this stuff up? I doubt that in the whole history of television manufacture there's ever been a 3:2 (1.5:1) television sold. In the old tube days they were all 1.33:1; these days they're (mostly) 1.78:1.
    When 640x480 is displayed on a monitor, no squishing occurs, but if fed to a tv, it occurs to fit the 3:2 aspect ratio tv set instead of the 4:3 computer monitor.
    Nonsense.
    Quote Quote  



Similar Threads

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