VideoHelp Forum
+ Reply to Thread
Results 1 to 23 of 23
Thread
  1. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Hi,
    I have a question regarding the frame height and I hope that someone has an answer for me. Window explorer reads my HD files' frame width and height as 1920x1072. However, Mediainfo reads the same files as 1920x1080. Which reported info is technically accurate?
    Here is the google drive link to one of those HD videos: https://drive.google.com/file/d/1HmZ1xPD-lTsCJg4eNBKFyUDp1b1GrFRx/view?usp=sharing
    I'd appreciate your helpful input to the question. Thank you in advance for your time.
    Quote Quote  
  2. Technically neither are correct, but mediainfo can be considered "less wrong" in this case

    It's actually stored as 1920x1088 . There is a standardized crop frame flag in AVC that cuts the bottom 8 pixels, so it displays as 1920x1080 - but it's actually 1920x1088. All "1920x1080" AVC videos are technically 1920x1088
    Quote Quote  
  3. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Originally Posted by poisondeathray View Post
    Technically neither are correct, but mediainfo can be considered "less wrong" in this case

    It's actually stored as 1920x1088 . There is a standardized crop frame flag in AVC that cuts the bottom 8 pixels, so it displays as 1920x1080 - but it's actually 1920x1088. All "1920x1080" AVC videos are technically 1920x1088
    Thanks for your response. Could you please download this HD video and examine it: https://drive.google.com/file/d/1Po3tvtHxYZq_A2lc9qy0fSGe50jysC1L/view?usp=sharing
    Both this one and the other are of the same VHS tape. The differences are the format I chose to record it with, and the (.mp4) video was recorded with 10-bit 4:2:2 instead of 8-bit 4:2:0 like the (.ts) video file. However, the (.ts) video file's frame height/ width reported by both Mediainfo and Window explorer are consistent. And the standardized crop frame flag in AVC didn't seem to apply in the (.ts) video. Please let me know. Thank you.
    Quote Quote  
  4. Yes, it's still technically 1920x1088. 8bit or 10bit AVC, "1920x1080" are really 1920x1088. Both have the crop frame flag
    Quote Quote  
  5. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Originally Posted by poisondeathray View Post
    Yes, it's still technically 1920x1088. 8bit or 10bit AVC, "1920x1080" are really 1920x1088. Both have the crop frame flag
    I just want to confirm with you that I didn't lose 23,936 pixels on the (.mp4) video file? Someone had asked the same question about the frame height being 1072 here: https://www.reddit.com/r/computers/comments/7m23up/what_is_the_difference_between_1912...2_and/drqyntb/
    I just want to make sure the (.mp4) video is good to keep. Another fact is, the (.mp4) video is unplayable across many Hd devices such as HDTV, bluray players through usb ports,..etc. I sent the (.mp4) video to my friends and asked them to file to play the (.mp4) video on their HD devices. The (.mp4) file could not be played on their HD devices either. But the (.ts) video can be perfectly played back on most HD devices of mine and my friends'. That was why I assumed that I lost 23,936 pixels on this (mp4) file.
    Quote Quote  
  6. Originally Posted by rgia View Post
    Originally Posted by poisondeathray View Post
    Yes, it's still technically 1920x1088. 8bit or 10bit AVC, "1920x1080" are really 1920x1088. Both have the crop frame flag
    I just want to confirm with you that I didn't lose 23,936 pixels on the (.mp4) video file? Someone had asked the same question about the frame height being 1072 here: https://www.reddit.com/r/computers/comments/7m23up/what_is_the_difference_between_1912...2_and/drqyntb/
    I just want to make sure the (.mp4) video is good to keep. Another fact is, the (.mp4) video is unplayable across many Hd devices such as HDTV, bluray players through usb ports,..etc. I sent the (.mp4) video to my friends and asked them to file to play the (.mp4) video on their HD devices. The (.mp4) file could not be played on their HD devices either. But the (.ts) video can be perfectly played back on most HD devices of mine and my friends'. That was why I assumed that I lost 23,936 pixels on this (mp4) file.
    You did not lose any pixels. Eitherway, a proper decoder will display as 1920x1080 and crop the 8 pixels

    The reason for playback issues and windows explorer issues - 10bit422 AVC video is usually not supported by many consumer applications and hardware

    It would be more ideal for 8bit 4:2:0 in MP4 container, because .ts container is not universally supported by devices either


    But there is a discerpancy in the frame rate. One is 59.94, one is 29.97 . The 1st looks badly deinterlaced
    Quote Quote  
  7. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Originally Posted by poisondeathray View Post
    Originally Posted by rgia View Post
    Originally Posted by poisondeathray View Post
    Yes, it's still technically 1920x1088. 8bit or 10bit AVC, "1920x1080" are really 1920x1088. Both have the crop frame flag
    I just want to confirm with you that I didn't lose 23,936 pixels on the (.mp4) video file? Someone had asked the same question about the frame height being 1072 here: https://www.reddit.com/r/computers/comments/7m23up/what_is_the_difference_between_1912...2_and/drqyntb/
    I just want to make sure the (.mp4) video is good to keep. Another fact is, the (.mp4) video is unplayable across many Hd devices such as HDTV, bluray players through usb ports,..etc. I sent the (.mp4) video to my friends and asked them to file to play the (.mp4) video on their HD devices. The (.mp4) file could not be played on their HD devices either. But the (.ts) video can be perfectly played back on most HD devices of mine and my friends'. That was why I assumed that I lost 23,936 pixels on this (mp4) file.
    You did not lose any pixels. Eitherway, a proper decoder will display as 1920x1080 and crop the 8 pixels

    The reason for playback issues and windows explorer issues - 10bit422 AVC video is usually not supported by many consumer applications and hardware

    It would be more ideal for 8bit 4:2:0 in MP4 container, because .ts container is not universally supported by devices either


    But there is a discerpancy in the frame rate. One is 59.94, one is 29.97 . The 1st looks badly deinterlaced
    I'm sorry if I was not clear. The (.ts) video files I sent to my friends across the US are perfectly playable on their HD devices. In contrast, the (.mp4) video is NOT.

    At first, I had the same visual perception as you do regarding the (.mp4) video. In fact, both (.mp4) and (.ts) files are DE-INTERLACED using the exact same algorithm. The (.ts) video has low frame rate, which you can obviously see the motion transition is hence NOT smooth compared to that of the (.mp4) video.
    Quote Quote  
  8. Originally Posted by rgia View Post
    I'm sorry if I was not clear. The (.ts) video files I sent to my friends across the US are perfectly playable on their HD devices. In contrast, the (.mp4) video is NOT.


    The MP4 is 10bit422, so you would not expect it to play - that is consistent with what was posted above. You would expect problems with playback and windows explorer

    The TS is playable, because it's 8bit420 . 8bit420 is more compatible with consumer devices - so you would expect it to play

    But I'm saying it would be even more playable than the TS, on devices such as portable devices , phones etc.. - if it was the same thing in MP4 container (ie. as 8bit 420 , not 10bit 422)

    At first, I had the same visual perception as you do regarding the (.mp4) video. In fact, both (.mp4) and (.ts) files are DE-INTERLACED using the exact same algorithm. The (.ts) video has low frame rate, which you can obviously see the motion transition is hence NOT smooth compared to that of the (.mp4) video.

    What is the capture from? Ideally you would cap it SD interlaced, then deinterlace it properly, then upscale if you wanted to . Right now it's done poorly

    It's not the same either in terms of image quality (not talking about frame rate now) - maybe you set it up the same, but the result is different. The first you can see deinterlacing artifacts, the 2nd you can see scan lines (but was encoded progressive)
    Last edited by poisondeathray; 19th Jan 2021 at 17:39.
    Quote Quote  
  9. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Originally Posted by poisondeathray View Post
    Originally Posted by rgia View Post
    I'm sorry if I was not clear. The (.ts) video files I sent to my friends across the US are perfectly playable on their HD devices. In contrast, the (.mp4) video is NOT.


    The MP4 is 10bit422, so you would not expect it to play - that is consistent with what was posted above. You would expect problems with playback and windows explorer

    The TS is playable, because it's 8bit420 . 8bit420 is more compatible with consumer devices - so you would expect it to play

    But I'm saying it would be even more playable than the TS, on devices such as portable devices , phones etc.. - if it was the same thing in MP4 container (ie. as 8bit 420 , not 10bit 422)

    At first, I had the same visual perception as you do regarding the (.mp4) video. In fact, both (.mp4) and (.ts) files are DE-INTERLACED using the exact same algorithm. The (.ts) video has low frame rate, which you can obviously see the motion transition is hence NOT smooth compared to that of the (.mp4) video.

    What is the capture from? Ideally you would cap it SD interlaced, then de-interlace it properly, then upscale if you wanted to . Right now it's done poorly
    May i ask which programs did you use to inspect the quality of every frame to come to a conclusion that it was done poorly? Which program are you using that tells you the de-interlacing process was mistakenly performed, seriously asking if you don't mind? I'd like to know about those programs so that I can self-offer a professional and constructive criticisms to my own in the future. Because one may say thing, others may think that, and some even see a completely different thing. So, it is the best bet to self-equip myself with your programs. Thank you in advance.
    Quote Quote  
  10. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Originally Posted by poisondeathray View Post
    Originally Posted by rgia View Post
    I'm sorry if I was not clear. The (.ts) video files I sent to my friends across the US are perfectly playable on their HD devices. In contrast, the (.mp4) video is NOT.


    The MP4 is 10bit422, so you would not expect it to play - that is consistent with what was posted above. You would expect problems with playback and windows explorer

    The TS is playable, because it's 8bit420 . 8bit420 is more compatible with consumer devices - so you would expect it to play

    But I'm saying it would be even more playable than the TS, on devices such as portable devices , phones etc.. - if it was the same thing in MP4 container (ie. as 8bit 420 , not 10bit 422)

    At first, I had the same visual perception as you do regarding the (.mp4) video. In fact, both (.mp4) and (.ts) files are DE-INTERLACED using the exact same algorithm. The (.ts) video has low frame rate, which you can obviously see the motion transition is hence NOT smooth compared to that of the (.mp4) video.

    What is the capture from? Ideally you would cap it SD interlaced, then deinterlace it properly, then upscale if you wanted to . Right now it's done poorly

    It's not the same either in terms of image quality (not talking about frame rate now) - maybe you set it up the same, but the result is different. The first you can see deinterlacing artifacts, the 2nd you can see scan lines (but was encoded progressive)
    I apologized for not having let you know that the original VHS tapes I used for digitizing already contained the artifacts you're talking about. I just don't notice them in the (.ts) video file as much as I see in the (.mp4) video file.
    The Hd video files were captured from a regular VHS tape. There is NO software used in my entire workflows. All hardware. I have the same VHS tapes captured as uncompressed .MOV and Prores HQ422. These files are meant to keep for distribution purposes.
    Quote Quote  
  11. Originally Posted by rgia View Post

    May i ask which programs did you use to inspect the quality of every frame to come to a conclusion that it was done poorly? Which program are you using that tells you the de-interlacing process was mistakenly performed, seriously asking if you don't mind? I'd like to know about those programs so that I can self-offer a professional and constructive criticisms to my own in the future. Because one may say thing, others may think that, and some even see a completely different thing. So, it is the best bet to self-equip myself with your programs. Thank you in advance.
    I'm using avisynth

    Beware there might be post processing applied if you're using a media player (software or hardware or TV), or video editor/NLE. ie. other filters between you and your friends. You have to be aware of the processing chain, and what is being done to the data/signal in between

    You can tell it was poorly deinterlaced because there are the "tell tale" signs such as buzzing lines, aliasing, stair stepping. Some of that is from poor upscaling too, there is some overlap in the types of artifacts, but those are classic signs


    Can you see the difference between these cropped screenshots between file 1 and 2 ? You said the same deinterlacing was applied to both, but clearly they are different. ( I' m not referring to the levels or color differences)
    1
    Image
    [Attachment 56851 - Click to enlarge]


    2
    Image
    [Attachment 56852 - Click to enlarge]
    Quote Quote  
  12. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Originally Posted by poisondeathray View Post
    Originally Posted by rgia View Post
    I'm sorry if I was not clear. The (.ts) video files I sent to my friends across the US are perfectly playable on their HD devices. In contrast, the (.mp4) video is NOT.


    The MP4 is 10bit422, so you would not expect it to play - that is consistent with what was posted above. You would expect problems with playback and windows explorer

    The TS is playable, because it's 8bit420 . 8bit420 is more compatible with consumer devices - so you would expect it to play

    But I'm saying it would be even more playable than the TS, on devices such as portable devices , phones etc.. - if it was the same thing in MP4 container (ie. as 8bit 420 , not 10bit 422)

    At first, I had the same visual perception as you do regarding the (.mp4) video. In fact, both (.mp4) and (.ts) files are DE-INTERLACED using the exact same algorithm. The (.ts) video has low frame rate, which you can obviously see the motion transition is hence NOT smooth compared to that of the (.mp4) video.

    What is the capture from? Ideally you would cap it SD interlaced, then deinterlace it properly, then upscale if you wanted to . Right now it's done poorly

    It's not the same either in terms of image quality (not talking about frame rate now) - maybe you set it up the same, but the result is different. The first you can see deinterlacing artifacts, the 2nd you can see scan lines (but was encoded progressive)
    Interestingly spoken. The two video clips were digitized from the same VHS tape through Composite jack, using all hardware. I'm not sure the artifacts you're talking about are the artifacts produced in the de-interlacing process or they are the native artifacts that are always there when using the composite jack as the output source and converted the analog composite signal to SDi signal.
    Quote Quote  
  13. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Did you receive the last two messages I sent? For some technical reasons, I cannot reply to you anymore.
    Quote Quote  
  14. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    I tried to reply, but my replies keep vanishing after I hit the submit response button.
    Quote Quote  
  15. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Originally Posted by poisondeathray View Post
    Originally Posted by rgia View Post
    I'm sorry if I was not clear. The (.ts) video files I sent to my friends across the US are perfectly playable on their HD devices. In contrast, the (.mp4) video is NOT.


    The MP4 is 10bit422, so you would not expect it to play - that is consistent with what was posted above. You would expect problems with playback and windows explorer

    The TS is playable, because it's 8bit420 . 8bit420 is more compatible with consumer devices - so you would expect it to play

    But I'm saying it would be even more playable than the TS, on devices such as portable devices , phones etc.. - if it was the same thing in MP4 container (ie. as 8bit 420 , not 10bit 422)

    At first, I had the same visual perception as you do regarding the (.mp4) video. In fact, both (.mp4) and (.ts) files are DE-INTERLACED using the exact same algorithm. The (.ts) video has low frame rate, which you can obviously see the motion transition is hence NOT smooth compared to that of the (.mp4) video.

    What is the capture from? Ideally you would cap it SD interlaced, then deinterlace it properly, then upscale if you wanted to . Right now it's done poorly

    It's not the same either in terms of image quality (not talking about frame rate now) - maybe you set it up the same, but the result is different. The first you can see deinterlacing artifacts, the 2nd you can see scan lines (but was encoded progressive)
    Interesting spoken!!! I used no software in the entire process of digitizing my VHS tapes. The two video clips were recorded from the same VHS tapes through composite jack, then composite signal is converted to sdi and upscale to 1080 using all hardware. There was no post processing as to de-interlacing, frame rate conversions..etc at all to be honest.
    Quote Quote  
  16. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    @poisondeathray
    Interestingly spoken!! The two video clips were digitized from the same VHS tape through composite signal converted to SDI, then this SDi signal is upconverted to HD1080p using all hardware. There is no software involved in any processed in my workflows. That being said, there is no post-processing on my end to be honest.
    I have the same files captured as uncompressed mov 4:2:2 and also proress hq422. These HD files I posted here are meant for distribution purposes.
    Quote Quote  
  17. Can you see this animated png ? If you use firefox or chrome it should animate if you click on it or open in a new window
    Image
    [Attachment 56853 - Click to enlarge]


    This is slowed down, but do you see those buzzing lines and jaggy artifacts ? This is from clip1


    Clip2 is not correct either, but it clearly has not the same treatment as clip1 . It's encoded progressively, but the 2nd field is slightly offset but really a duplicate, this results in the scan lines. But the fields that are encoded are potentially much cleaner than in clip1, because clip1 has be degraded by poor deinterlacing

    If you recorded it properly, then deinterlaced it properly, perhaps upscaled it properly, you wouldn't see those types of artifacts (or at least not as bad)


    The two video clips were recorded from the same VHS tapes through composite jack, then composite signal is converted to sdi and upscale to 1080 using all hardware. There was no post processing as to de-interlacing, frame rate conversions..etc at all to be honest.
    Ok, but clearly there are differences in the deinterlacing filter applied between the 2 clips. There are problems in both, probably avoidable problems
    Quote Quote  
  18. Member Cornucopia's Avatar
    Join Date
    Oct 2001
    Location
    Deep in the Heart of Texas
    Search PM
    In modern times, "All in Hardware" = "Used ASIC chips running custom software". That software (hardware) could still have poor algorithms or be setup incorrectly.

    To better analyse, we need DETAILS. Model #s, settings used...


    Scott
    Quote Quote  
  19. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Originally Posted by poisondeathray View Post
    Can you see this animated png ? If you use firefox or chrome it should animate if you click on it or open in a new window
    Image
    [Attachment 56853 - Click to enlarge]


    This is slowed down, but do you see those buzzing lines and jaggy artifacts ? This is from clip1


    Clip2 is not correct either, but it clearly has not the same treatment as clip1 . It's encoded progressively, but the 2nd field is slightly offset but really a duplicate, this results in the scan lines. But the fields that are encoded are potentially much cleaner than in clip1, because clip1 has be degraded by poor deinterlacing

    If you recorded it properly, then deinterlaced it properly, perhaps upscaled it properly, you wouldn't see those types of artifacts (or at least not as bad)


    The two video clips were recorded from the same VHS tapes through composite jack, then composite signal is converted to sdi and upscale to 1080 using all hardware. There was no post processing as to de-interlacing, frame rate conversions..etc at all to be honest.
    Ok, but clearly there are differences in the deinterlacing filter applied between the 2 clips. There are problems in both, probably avoidable problems
    I did not zoom the image to that smaller scale since the HD files are meant to be used for distribution purposes online. Yes, I can definitely see what you're saying now with the image being extensively zoomed in. I'm wondering if those buzzing lines are a result of me stretching the sides, top, and bottom to remove the overscan lines around four cornors of the video? How would you think?
    Quote Quote  
  20. Originally Posted by rgia View Post

    I did not zoom the image to that smaller scale since the HD files are meant to be used for distribution purposes online. Yes, I can definitely see what you're saying now with the image being extensively zoomed in. I'm wondering if those buzzing lines are a result of me stretching the sides, top, and bottom to remove the overscan lines around four cornors of the video? How would you think?

    The animated png is not a "zoom in"; it's a crop at 100%

    Those are deinterlacing artifacts as a result of deinterlacing with a low quality algorithm. It's a classic "textbook" example of bad deinterlacing.

    Resizing/resampling the image (what you can stretching) if done after deinterlacing is ok procedurally. But if you did it before deinterlacing , it would cause another sort of artifact. That is not the cause of those aliasing artifacts. What your hardware is essentially doing is doubling the scan lines - it's a "cheap" form of deinterlacing that causes those artifacts
    Last edited by poisondeathray; 19th Jan 2021 at 18:50.
    Quote Quote  
  21. Member
    Join Date
    Apr 2019
    Location
    USA
    Search Comp PM
    Originally Posted by poisondeathray View Post
    Originally Posted by rgia View Post

    I did not zoom the image to that smaller scale since the HD files are meant to be used for distribution purposes online. Yes, I can definitely see what you're saying now with the image being extensively zoomed in. I'm wondering if those buzzing lines are a result of me stretching the sides, top, and bottom to remove the overscan lines around four cornors of the video? How would you think?

    It's not a "zoom in"; it's a crop at 100%

    Those are deinterlacing artifacts as a result of deinterlacing with a low quality algorithm. It's classic "textbook" example of bad deinterlacing
    Interesting. Do you have any great de-interlacing hardware in mind to recommend?
    Quote Quote  
  22. Originally Posted by rgia View Post
    Interesting. Do you have any great de-interlacing hardware in mind to recommend?
    Not really.

    In hardware, the only decent ones I've seen are the motion adaptive ones on more expensive TV sets . I have not seen any devices that have anything close to this. Capture devices are designed for real time capture, not for quality. The deinterlacing quality on the expensive TV sets is almost as good as high quality software deinterlacing in motion (software is still better) , but better with no motion (they essentially do a perfect weave without degrading image, where a software deinterlacer might slightly degrade no motion segments)
    Quote Quote  
  23. Member Cornucopia's Avatar
    Join Date
    Oct 2001
    Location
    Deep in the Heart of Texas
    Search PM
    Best practice: Do not deinterlace. Or if you must, do it as late stage in the workflow as possible (assuming preceding tools are interlace-aware, especially resizing!), and then do it just once and use best algorithm possible (depending on content, but often this is QTGMC in software).

    Scott
    Quote Quote  



Similar Threads

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