Capturing Tapes to ProRes vs DV

[celsoac]

Thank you all of you who help me with the installation of Hybrid in Mac and with the use of the QTGMC deinterlacer. I did solve these issues (so far? ) thanks, particulary, to poisondeathray. I've also learned about the intrinsic, relative quality of DV versus other formats, about archiving and (not) deinterlacing, etc. For example, I often converted DV files from VHS into a cooler-looking ProRes just for archiving. But I now realize it's counterproductive and unnecessary, as DV can be used for archiving and used for output files as needed.

Now, having said all that, just one thing. I did not try to take over the thread with questions not directly related to the original post by Christina. Sorry if it seemed that way (sorry, Christina). You guys hang around these forums a lot. I am a newbie and an amateur -- I digitize and edit videos for my work and other practical purposes, not as a specialist, and I only come to forum.videohelp.com (and to videohelp itself, for many years) for specific info and help. Perhaps some of you, being helpful as you are, assume the interlocutor knows or should know more than they do, in terms of how video capturing, processing and editing works computationally. So your expertise needs to be complemented with a little more didactic approach, instead of decapitating right away whomever says something off-the-track. (We) learners need a lot of show-and-tell, not just the theory of why things are the way they are.

Cheers,
-celso

[lordsmurf]

little more didactic approach, instead of decapitating right away whomever says something off-the-track. (We) learners need a lot of show-and-tell, not just the theory of why things are the way they are.

No. False/bad/wrong info needs to be halted right away. Only then can we have a conversation about doing things the proper way.

Requesting sample images is nice and all, but it takes more effort that you realize. I often don't have time for that.

Thank you all of you who help me

In the end, that's what matters.

[Christina]

Not a problem at all for me! No worries. If I decide to install Hybrid I know who to ask for help now.

Given the amount of time, research, and troubleshooting you did to get it working, is the result significantly better than what you get with yadif/bob in handbrake to make it worth your while? I would love to see 2 side by side comparisons of some of your deinterlaced captures using both methods. Apologies if that’s already been done - I have only been half paying attention to this thread for the past week or so and will go back and read through again.

[celsoac]

Not a problem at all for me! No worries. If I decide to install Hybrid I know who to ask for help now.

Given the amount of time, research, and troubleshooting you did to get it working, is the result significantly better than what you get with yadif/bob in handbrake to make it worth your while? I would love to see 2 side by side comparisons of some of your deinterlaced captures using both methods. Apologies if that’s already been done - I have only been half paying attention to this thread for the past week or so and will go back and read through again.

Christina,

I haven't compared the two methods. I'll try to do that. Hybrid also has yadif.

Did you fix the display aspect problem with your mp4 ?

-celso

[Christina]

In my particular playback devices, aspect ratio flags are being ignored, so to get the correct display aspect ratio I just have to encode my 720x480 4:3 to 640x480 1:1 (square pixels). I haven’t played around with it too much yet but I suppose you can say my problem is fixed. Really wish I didn’t have to change the aspect ratio though.

[jagabo]

is the result (WTGMC) significantly better than what you get with yadif/bob in handbrake to make it worth your while?

Once you've used QTGMC you will not want to use anything else (except for some special purpose quick/dirty deinterlacing).

[Christina]

How does the decomb filter in handbrake compare to QTGMC? From what I understand and correct me if I am wrong, it uses yadif but only where it detects it’s needed- therefore I assume it leaves more of the video “untouched”. Is QTGMC always a better option or does it depend on the footage/source? I would love to see the 3 options (decomb, deinterlace with yadif/bob, deinterlace with QTGMC) side by side with actual samples (video samples, not frame jpegs). I’m sure it’s been done somewhere but.. where, I don’t know. I can do my own tests with the 2 methods in handbrake but haven’t attempted to install Hybrid yet. Also I feel like when I first started researching this, all I kept seeing was handbrake’s decomb, and lately all I keep seeing is QTGMC.

[Christina]

Also, with interlaced DV as my source file, I tried handbrake's yadif/bob deinterlacing plus changing frame rate from 29.97 to 59.94. I am not sure how I feel about this - It definitely looks super smooth on playback, but almost in an unnatural or unusual way.

[celsoac]

In my particular playback devices, aspect ratio flags are being ignored, so to get the correct display aspect ratio I just have to encode my 720x480 4:3 to 640x480 1:1 (square pixels). I haven’t played around with it too much yet but I suppose you can say my problem is fixed. Really wish I didn’t have to change the aspect ratio though.

Would you mind posting a short cut (a couple of seconds would be enough) of a "distorted" MP4? I am curious about one thing, though I don't want to say it aloud for fear .

[celsoac]

Also, with interlaced DV as my source file, I tried handbrake's yadif/bob deinterlacing plus changing frame rate from 29.97 to 59.94. I am not sure how I feel about this - It definitely looks super smooth on playback, but almost in an unnatural or unusual way.

Are you converting telecined home recordings (original film at 18fps or 16fps)?

[Christina]

Would you mind posting a short cut (a couple of seconds would be enough) of a "distorted" MP4? I am curious about one thing, though I don't want to say it aloud for fear .

Here you go. This plays fine on my computer but on tv it's playing it 720x480 1:1.

[Christina]

Also, with interlaced DV as my source file, I tried handbrake's yadif/bob deinterlacing plus changing frame rate from 29.97 to 59.94. I am not sure how I feel about this - It definitely looks super smooth on playback, but almost in an unnatural or unusual way.

Are you converting telecined home recordings (original film at 18fps or 16fps)?

It's 29.97 interlaced home video recordings, as far as I know.

[Seed]

If you need to fix the problem with sound, perhaps you can help the program to convert video files from mp4 to mov https://www.videoconverter.com/how-to-convert-mp4-to-mov.html. It is very convenient, seriously, just very easy to convert files, without loss of quality and also with a convenient interface. It's just great!

[Christina]

Any audio problem can be fixed with simple video editing software like free VE from Movavi. It is easy to learn and the output quality is quite good.

Thanks but I wasn't having any audio problems (at the moment, anyway). We were talking about my playback devices ignoring aspect ratio flags.

[Seed]

Any audio problem can be fixed with simple video editing software like free VE from Movavi. It is easy to learn and the output quality is quite good.

Thanks but I wasn't having any audio problems (at the moment, anyway). We were talking about my playback devices ignoring aspect ratio flags.

Oh, then I do not quite understand you, now I'll try to look for something on this topic. Excuse me!

[jagabo]

is the (QTGMC) result significantly better than what you get with yadif/bob in handbrake to make it worth your while? I would love to see 2 side by side comparisons

I don't have any good interlaced SD material here for demonstration. So I created an interlaced DV AVI video from a progressive high definition source. The stockholm video in the 720p folder from here:

https://media.xiph.org/ldv/pub/test_sequences/

I deinterlaced with qtgmc and Handbrake's yadif+bob at both 29.97 fps and 59.94 fps. MKV files attached. All encoded with x264, Slow preset, at (C)RF 18.

The clip starts out with no motion so you can verify that yadif uses both fields when there is no motion. Once the camera starts panning you can see bad moire artifacts on the stairs in the 29.97 fps yadif video (and less visible aliasing artifacts in other parts of the frame). The artifacts are still there in the 59.94 fps yadif video but they are less visible because of the high frame rate. The qtgmc video is very clean throughout. Even if you step through the video frame by frame there are hardly any moire/aliasing artifacts. Also note how the bitrate of the 59.94 fps yadif+bob video is much higher because of all the flickering edges.

The source video (dv.avi) is included so you can experiment on your own.

[celsoac]

[[Deleted. Post repeated]]

[celsoac]

Would you mind posting a short cut (a couple of seconds would be enough) of a "distorted" MP4? I am curious about one thing, though I don't want to say it aloud for fear .

Here you go. This plays fine on my computer but on tv it's playing it 720x480 1:1.

Thanks. OK, I'm going to say something that may make people jump at my throat: I think the culprit, weird as it sounds, is the "TV standard: NTSC" flag. I don't know how to edit that (with Subler I can only play with AR). What tool do you use to edit MP4 flags? Try, if you can, setting TV standard: undefined, and Interlacing: undefined. Below is a report for the file from Videospec (if you don't have it, it's free, very useful).

Another test: I've changed the display size to 720x540, with Subler, no reencoding. Try that.

Also try to just make whateve capture from any source at 720x480 encoded pixels, with AR 8:9 = 640, but logically without the TV standard flag.

(Ready to receive criticism)

-celso

====

shortclipARtest

***** First Analyzed File Results *****

*** General Parameters ***
- Name: shortclipARtest.mp4
- Container: MP4 - QuickTime
- Creation Date: 2019-04-25 18:36:29
- Size: 5.9 MiB
- Duration: 0:00:12
- Bitrate: 4 089 Kbps
- Encoding Library: Undefined
- Encoding Application: Undefined

*** Video Track Parameters ***
- Format: H.264/MPEG-4 AVC
- Size: 5.73 MiB (97%)
- FourCC: H264
- Track number(s): 1
- Bitrate: Max.: Undefined
Average: 3 949 Kbps
Min.: Undefined
- Frame rate (fps): Max.: ---
Average: 29.970
Min.: ---
- Bitrate mode: Variable
- Encoding profile: Main@L4.0
- Resolution: 24 bits
- Width (Pixel number): 720
- Height (Pixel number): 480
- Pixel Aspect Ratio: 8:9
- Display aspect ratio: 4:3
- Chroma subsampling format: YUV420p
- TV standard: NTSC
- Interlacing: Progressive
- Encoding library: x264 core 155 r2901 7d0ff22
- Additional Parameters: CABAC: Yes
Reference Frames: 4
Bits/(Pixel*Frame) ratio: 0.381

*** Audio Track(s) Parameters ***
- Format: AAC - MPEG-4 audio
- Size: 191 KiB (3%)
- FourCC: MP4A
- Number(s) and language(s): 1: Undefined
- Details: -
- Profile: Low Complexity
- Bitrate: 251 Kbps
- Bitrate mode: Variable
- Resolution: 16 bits
- Rate: 48.0 KHz
- Channel(s): 2 (stereo)
- Position: Left Right
- Encoding Library: Undefined
- Additional Parameters: Spectral Band Replication: No

*** Miscellaneous ***
- Subtitle(s): No Subtitle
- Metadata: Album: Undefined
Part number: Undefined
Track name: Undefined
Track name number: Undefined
Performer: Undefined
Screenplayer/Writer: Undefined
Genre: Undefined
Encoded date: UTC 2019-04-25 14:10:29
Comment: Undefined
Album artist: Undefined
Grouping: Undefined
Copyright: Undefined

[Christina]

I deinterlaced with qtgmc and Handbrake's yadif+bob at both 29.97 fps and 59.94 fps. MKV files attached. All encoded with x264, Slow preset, at (C)RF 18.

Thank you for doing this! Definitely looks good with QTGMC. I may have to try it out on my own videos.

A few things going through my head:

-Your source footage looks way better than my source footage.
-This looks like a very challenging scene (lot of movement, horizontal panning, lots of details in the windows, sign on building, etc.).

[Christina]

Any audio problem can be fixed with simple video editing software like free VE from Movavi. It is easy to learn and the output quality is quite good.

Thanks but I wasn't having any audio problems (at the moment, anyway). We were talking about my playback devices ignoring aspect ratio flags.

Oh, then I do not quite understand you, now I'll try to look for something on this topic. Excuse me!

Something funky might have happened to the test video I uploaded when I tried to extract a short clip from it, which might explain why you assumed I was having an audio problem. It seemed to play fine when I previewed it prior to uploading it from home, but now downloading it from a different computer to view, the audio does seem off. However, that is probably an effect of how I extracted the segment, which I did by setting In and Out points in MPEGStreamClip and using Save-As command. Perhaps not the best approach?

[Christina]

Thanks. OK, I'm going to say something that may make people jump at my throat: I think the culprit, weird as it sounds, is the "TV standard: NTSC" flag. I don't know how to edit that (with Subler I can only play with AR). What tool do you use to edit MP4 flags? Try, if you can, setting TV standard: undefined, and Interlacing: undefined.

I don't know enough about this to comment, nor would I know how to remove the "NTSC flag". To generate this file I used Handbrake. But as I noted in a previous post I tried to extract a clip using MPEGStreamClip and I might have messed something up when doing that.

Also try to just make whateve capture from any source at 720x480 encoded pixels, with AR 8:9 = 640, but logically without the TV standard flag.

I have no idea what this even means to be honest ("but logically without the TV standard flag").

Thank you for testing!

[celsoac]

Thanks. OK, I'm going to say something that may make people jump at my throat: I think the culprit, weird as it sounds, is the "TV standard: NTSC" flag. I don't know how to edit that (with Subler I can only play with AR). What tool do you use to edit MP4 flags? Try, if you can, setting TV standard: undefined, and Interlacing: undefined.

I don't know enough about this to comment, nor would I know how to remove the "NTSC flag". To generate this file I used Handbrake. But as I noted in a previous post I tried to extract a clip using MPEGStreamClip and I might have messed something up when doing that.

Also try to just make whateve capture from any source at 720x480 encoded pixels, with AR 8:9 = 640, but logically without the TV standard flag.

I have no idea what this even means to be honest ("but logically without the TV standard flag").

Thank you for testing!

I don't think there is a problem with splicing files like that with MPEGStreamclip. I do it all the time and all the video properties are preserved. If you paste together two MP4s coded differently, weird things may happen with audio or video (most often, they get lost), but if you cut a piece, no problem, in my experience.

I call it "TV standard flag" but perhaps it's metadata, I don't know. It's just information about how to display that file as a "TV" file. I've created a screen capture video at 720x480 real square pixels. It does not come from a TV source. When I add the flag about PAR 8:9 so that the display is 640x480, then Videospec tells me that that file is interpreted as "TV Standard NTSC".

So, if your computer displays an MP4, 720x480 => 640x480 file correctly, but your TV doesn't, perhaps (my naive reasoning) it is precisely because the TV thinks it is an NTSC file, but something else (I don't know what) doesn't match. So if you are able to remove that "TV Standard" information from the MP4, perhaps the device will treat the file as any other video file where an PAR is applied, and the number of encoded and displayed pixels don't match.

For example, try to make an MP4 file with MPEGStreamclip with other dimensions and also the 8:9 PAR, or with the 720x480 dimensions but another PAR.

[celsoac]

Christina, try this file. If it displays alright both in your computers and in your TV, then all you have to do is:
- you clip 8 pixels both on left and right from the DV file (which is 720x480 SAR and has PAR 8:9), so that the resulting encoded pixels will be 704 x 480; so, in fact, you'll reduce the width of those black bands on both sides. In fact, those 704 encoded pixels are the standard sampling both for NTSC and PAL systems; the added pixels up to 720 are for overscan or errors.
- you convert this with PAR 10:11 so that => 640x480 display. If the output is MP4 H.264 and by default you get 1:1, you can correct PAR with Subler.

What I did to create this file was to convert your sample to DV, then convert it back to MP4; so disregard the quality.

[celsoac]

is the (QTGMC) result significantly better than what you get with yadif/bob in handbrake to make it worth your while? I would love to see 2 side by side comparisons

I don't have any good interlaced SD material here for demonstration. So I created an interlaced DV AVI video from a progressive high definition source. The stockholm video in the 720p folder from here:

https://media.xiph.org/ldv/pub/test_sequences/

I deinterlaced with qtgmc and Handbrake's yadif+bob at both 29.97 fps and 59.94 fps. MKV files attached. All encoded with x264, Slow preset, at (C)RF 18.

The clip starts out with no motion so you can verify that yadif uses both fields when there is no motion. Once the camera starts panning you can see bad moire artifacts on the stairs in the 29.97 fps yadif video (and less visible aliasing artifacts in other parts of the frame). The artifacts are still there in the 59.94 fps yadif video but they are less visible because of the high frame rate. The qtgmc video is very clean throughout. Even if you step through the video frame by frame there are hardly any moire/aliasing artifacts. Also note how the bitrate of the 59.94 fps yadif+bob video is much higher because of all the flickering edges.

The source video (dv.avi) is included so you can experiment on your own.

My two cents: blending fields with JES Deinterlacer for Mac, 29.97p; and deinterlacing by both fields (bob-like), 59.94p, also with JES Deinterlacer.

My humble opinion: JES Deinterlacer is bettter than yadif, worse than QTGMC 59.94 (we don't know about QTGMC 29.97 because we don't have it here).

[jagabo]

QTGMC at 29.97 fps looks the same as the 59.94 fps version, just with half as many frames.

[celsoac]

QTGMC at 29.97 fps looks the same as the 59.94 fps version, just with half as many frames.

Well, I am sure the field blending is equally great, but movement and flow must look different.

Many thanks for the samples, by the way.

[jagabo]

Yes, QTGMC at the original frame rate is jerky/flickery like any 30p video.

[celsoac]

Yes, QTGMC at the original frame rate is jerky/flickery like any 30p video.

Thank you. Yes it is.

[jagabo]

QTGMC has the ability to add motion blur to reduce the flicker. Here's it's maxed out:

[lordsmurf]

My humble opinion: JES Deinterlacer is bettter than yadif

There's no way that blended deinterlace is better than Yadif. Also realize that Yadif has many options, and has mod methods as well. It can be almost as good as QTGMC, or almost as bad as drop-field, depending on what was done.