Is it worth it recapturing videos?

[rgr]

As the gee gee goes behind jump 2, you can see the bottom of the tape goes pearshaped. Damaged, I reckon.

Tested, confirmed.

You mean YOUR tests.

I wasn't the only person to confirm this.

[Riker0007]

How should I proceed and would anyone want to see a different capture? Perhaps the JVC with TBC off and just through passthrough? That being said, when I was testing with both virtualdub2 and Amrectv I notice that with the JVC and Pany passthrough the duplicate/drop frame was not racking up the numbers like it did with just the JVC (TBC was on). FYI, the Blackmagic was like filling up your gas tank with dup/dropped frames. Other tapes that where maybe 30 mins or 1 hr only had like 10 with just the JVC w/tbc on, but not sure then when I go to change it to prores will it spit out dups like it did before or was it just my bad script?

[Sharc]

Other tapes that where maybe 30 mins or 1 hr only had like 10 with just the JVC w/tbc on, but not sure then when I go to change it to prores will it spit out dups like it did before or was it just my bad script?

Conversion to ProRes should not multiply drops/inserts unless something is bad with your script or setup IMO. Start with a short interlaced capture (few minutes .avi) with no drops/inserts to practice and assess your conversion to ProRes. Also be clear about interlaced and deinterlaced sources before throwing the clip to some conversion script or commandline. And post your script and/or commandlines, not just the results so it is easier for other users to follow your issues.
Lastly, maybe the condition of (some of) your tapes (and perhaps the condition/wear of the player as well) make the addition of the ES-15 (and JVC with TBC OFF as this may have caused the vertical jumps) a prerequisite for a more or less decent capture?

[jagabo]

How should I proceed and would anyone want to see a different capture? Perhaps the JVC with TBC off and just through passthrough?

Yes.

[Riker0007]

How should I proceed and would anyone want to see a different capture? Perhaps the JVC with TBC off and just through passthrough?

Yes.

https://files.videohelp.com/u/139455/jvc%20w%20pass%20tbc%20off.avi
https://files.videohelp.com/u/139455/JVC%20w%20pass%20tbc%20off.txt

[Sharc]

This looks like a valid capture to me, the best until now. Interlaced TFF, 29.97fps. There is a glitch at field#944, and I am not sure about a drop around field# 820 (?). I don't see the need for a recapture though. It's probably as good as it can get.
So the next steps would be to either
- Deinterlace and then convert to ProRes progressive, or
- Convert to Prores interlaced (and then deinterlace etc.)

(we postpone denoising, de-rainbowing etc. for a later discussion if needed. Depends what you leave for DaVinci)

Let's see the findings from other users.

Edit:
I uploaded deinterlaced and interlaced ProRes variants of your capture here. They should open in DaVinci without problems.
https://mega.nz/file/DMFlBQbQ#X1GYsXGsAv6f-vIgsZKoJyUa-cN46uqr2AyMameRxjI
https://mega.nz/file/CYUj1SyB#btOmaJPPDEX__cD5Xw8Uv3VS4B5Z26hSFvqZsGTYl1k

[Riker0007]

This looks like a valid capture to me, the best until now. Interlaced TFF, 29.97fps. There is a glitch at field#944, and I am not sure about a drop around field# 820 (?). I don't see the need for a recapture though. It's probably as good as it can get.
So the next steps would be to either
- Deinterlace and then convert to ProRes progressive, or
- Convert to Prores interlaced (and then deinterlace etc.)

(we postpone denoising, de-rainbowing etc. for a later discussion if needed. Depends what you leave for DaVinci)

Let's see the findings from other users.

Edit:
I uploaded deinterlaced and interlaced ProRes variants of your capture here. They should open in DaVinci without problems.
https://mega.nz/file/DMFlBQbQ#X1GYsXGsAv6f-vIgsZKoJyUa-cN46uqr2AyMameRxjI
https://mega.nz/file/CYUj1SyB#btOmaJPPDEX__cD5Xw8Uv3VS4B5Z26hSFvqZsGTYl1k

Is this capture, Jvc with pass through AND TBC off better than the one with JVC TBC on and pass through on from above? In recapturing I’ve only have had one tape that had a glitching flag at the top with the tbc on and pass through. Turing it off flag went bye bye. It was the same thing before when I did not have the pany for pass through though.

[Sharc]

Is this capture, Jvc with pass through AND TBC off better than the one with JVC TBC on and pass through on from above?

Yes. With JVC TBC OFF and ES-15 in passthrough you didn't get these annoying vertical up/down picture jumps for certain scenes which you got before with the JVC TBC ON and ES-15 in passthrough.

[Riker0007]

Is this capture, Jvc with pass through AND TBC off better than the one with JVC TBC on and pass through on from above?

Yes. With JVC TBC OFF and ES-15 in passthrough you didn't get these annoying vertical up/down picture jumps for certain scenes which you got before with the JVC TBC ON and ES-15 in passthrough.

So should I just capture with the TBC off on the JVC and panny pass through on the rest of the tapes?

[jagabo]

This looks like a valid capture to me, the best until now. Interlaced TFF, 29.97fps.

A dropped frame was replaced with a duplicate. That missing frame can be motion interpolated from the frames before and after. See attache sample.

There is a glitch at field#944, and I am not sure about a drop around field# 820 (?)

.
I think it's just a camera jerk.

This is a much better cap than the those in post #109.

[Sharc]

So should I just capture with the TBC off on the JVC and panny pass through on the rest of the tapes?

There is no strong and strict rule to this. The prevalent opinion is to switch the VCR TBC off when a DVD recorder like ES-15 in passthrough is in place, because the ES-15 provides this similar TBC functionality itself, and having this function 2x in the loop can do more harm than good. But there are always exceptions to the rules. You have to gain your own experience. Capturing can be laborious at times.

Now, after 131 posts, did you find (for yourself) the the answer to your original question?

[Alwyn]

A dropped frame was replaced with a duplicate. That missing frame can be motion interpolated from the frames before and after. See attache sample.

That looks good but how did you actually achieve that? Did you RIFE the whole file or just that section?

[jagabo]

I used RIFE to generate those two frames. Search these forums for "ReplaceFramesRIFE", the function I wrote to do that.

[Riker0007]

So should I just capture with the TBC off on the JVC and panny pass through on the rest of the tapes?

There is no strong and strict rule to this. The prevalent opinion is to switch the VCR TBC off when a DVD recorder like ES-15 in passthrough is in place, because the ES-15 provides this similar TBC functionality itself, and having this function 2x in the loop can do more harm than good. But there are always exceptions to the rules. You have to gain your own experience. Capturing can be laborious at times.

Now, after 131 posts, did you find (for yourself) the the answer to your original question?

Yep. Recapturing. Just not exactly sure on jvc tbc on or off with the pass through. Doing entire tapes both ways but would like to know how to really scan them correctly for the better capture. If during capture amrectv show like 5 or less dupe frames, is it worth recapturing with tbc off on jvc? From memory a lot of both tbc and pass through same as just jvc and tbc.

[Sharc]

Now, after 131 posts, did you find (for yourself) the the answer to your original question?

Yep. Recapturing. Just not exactly sure on jvc tbc on or off with the pass through. Doing entire tapes both ways but would like to know how to really scan them correctly for the better capture. If during capture amrectv show like 5 or less dupe frames, is it worth recapturing with tbc off on jvc? From memory a lot of both tbc and pass through same as just jvc and tbc.

If you are in doubt you need not to re-capture the entire tape. You can re-capture the flawed section(s) only (maybe 10+ frames) and - if it has fixed the issue - eventually stitch it frame (field) accurate together with the original capture (overlay, replace). Easiest using the timeline of an NLE like your DaVinci for doing this.
Or you fix the flaws (if not too many faulted frames in a row) by motion interpolation of the adjacent healthy frames, as jagabo has suggested.

[Riker0007]

Now, after 131 posts, did you find (for yourself) the the answer to your original question?

Yep. Recapturing. Just not exactly sure on jvc tbc on or off with the pass through. Doing entire tapes both ways but would like to know how to really scan them correctly for the better capture. If during capture amrectv show like 5 or less dupe frames, is it worth recapturing with tbc off on jvc? From memory a lot of both tbc and pass through same as just jvc and tbc.

If you are in doubt you need not to re-capture the entire tape. You can re-capture the flawed section(s) only (maybe 10+ frames) and - if it has fixed the issue - eventually stitch it frame (field) accurate together with the original capture (overlay, replace). Easiest using the timeline of an NLE like your DaVinci for doing this.
Or you fix the flaws (if not too many faulted frames in a row) by motion interpolation of the adjacent healthy frames, as jagabo has suggested.

I'll have to figure out how to do that. A capture last night had 3 dups according to amrectv when capturing, but I need to learn how to run the scans and find the bad frames. Thanks!

[Sharc]

I'll have to figure out how to do that. A capture last night had 3 dups according to amrectv when capturing, but I need to learn how to run the scans and find the bad frames. Thanks!

Or if the flaws are not annoyingly eyecatching when playing the video you may want to live with it and leave as is. Depends how much learning and time you are willing to spend with it.

[Riker0007]

Yep. I think I will finish with capturing with the JVC TBC on and Passthrough ( I can see when it glitches with TBC on and then turn off and restart) and then proceed with the deinterlace and then convert to prores for editing. I'll have to re-read all the helpful posts on the steps again to get it right.
Thanks!

[Riker0007]

I've been trying to write a script that will batch all .avi files I've created to use qtgmc and convert to prores
It keeps failing: Any ideas?


@echo off
setlocal enabledelayedexpansion

REM Output folder name
set outdir=Converted
set "avsdir=avs_temp"

REM Make sure folders exist
if not exist "%outdir%" mkdir "%outdir%"
if not exist "%avsdir%" mkdir "%avsdir%"

for %%a in ("*.avi") do (
echo Converting: %%a
set "filename=%%~na"
set "avsfile=%avsdir%\!filename!.avs"

REM Create AviSynth script dynamically
(
echo AVISource("%%a")
echo AssumeTFF()
echo QTGMC(Preset="Slow", Preset=3)
echo LanczosResize(720,480)
) > "!avsfile!"

REM Run FFmpeg: take video from AVS, audio passthrough, ProRes output
ffmpeg -y -i "!avsfile!" -c:v prores_ks -profile:v 3 -bits_per_mb 8000 -pix_fmt yuv422p10le -c:a copy "%outdir%\!filename!.mov"
)

echo Done!
pause


Ideally, I'd like to then convert them to h.264 files


Thanks

[Alwyn]

I can't help you with your script failing but I don't see any code in there that will reshape your videos to 4:3. The 720x480 (3:2) has to be brought down to 4:3, or 640x480 equivalent, unless you're going to do that in Resolve.

Also, is the resize necessary? The video is already 720x480.

[Riker0007]

I can't help you with your script failing but I don't see any code in there that will reshape your videos to 4:3. The 720x480 (3:2) has to be brought down to 4:3, or 640x480 equivalent, unless you're going to do that in Resolve.

Also, is the resize necessary? The video is already 720x480.

How do you do your vids? One at a time? Honestly, I've been plugging it into chatgpt for help with the "all files in x folder". If I do one at time, I'm good. I think the resize was just extra stuff it included. As far as the 4:3, I thought it needed to be that since that was how the original was? 3:2 is fine with me if that's the way it should be.

[Alwyn]

I do a couple of tapes at a time but I don't use any command line code; I use batching in VDub2.

As far as the 4:3, I thought it needed to be that since that was how the original was?

That's correct: it should be 4:3. But if you don't explicitly tell it to set the DAR to 4:3 or set the correct SAR, it will just create a DAR of 3:2 from the frame ratio of 720x480 (see your files in post #109). As I say, Resolve might be able to do that for you but I'd do it at this initial stage so you don't have to worry about it later.

[Alwyn]

Re the SAR, I spotted a "setsar" code in Sharc's post #82.

[Sharc]

@Riker: Have you succeeded to produce a deinterlaced (qtgmc) ProRes version of your .avi capture, before jumping onto some chatgpt batch mass production? Did your QTGMC(....) deinterlacing work correctly? QTGMC(Preset="Slow", Preset=3) makes no sense and will fail. Delete the Preset=3. And as said include -vf setsar=10/11 or -vf setdar=4/3 in the ffmpeg commandline for 4:3 display (unless you do the sizing later in DaVinci).
(Footnote only: And if you want to have it in .h264 anyway you could skip ProRes and convert your .avi directly to lossless x264 which DaVinci should open as well).

Edit: For batch scripting you may find this useful:
https://www.tutorialspoint.com/batch_script/index.htm

[Riker0007]

@Riker: Have you succeeded to produce a deinterlaced (qtgmc) ProRes version of your .avi capture, before jumping onto some chatgpt batch mass production? Did your QTGMC(....) deinterlacing work correctly? QTGMC(Preset="Slow", Preset=3) makes no sense and will fail. Delete the Preset=3. And as said include -vf setsar=10/11 or -vf setdar=4/3 in the ffmpeg commandline for 4:3 display (unless you do the sizing later in DaVinci).
(Footnote only: And if you want to have it in .h264 anyway you could skip ProRes and convert your .avi directly to lossless x264 which DaVinci should open as well).

Edit: For batch scripting you may find this useful:
https://www.tutorialspoint.com/batch_script/index.htm

Preset 3 was a typo, should have been 2. Yes I got it correct doing one file. Was just doing prorest to edit, then have it as h.264 for playback.

[_Al_]

It keeps failing: Any ideas?

to get intermediate 4:2:2, in prores this should work, not sure how 10bit conversion can benefit, but anyway:

@echo off
setlocal
rem Output folder name
set "outdir=Converted"
set "avsdir=avs_temp"
set "ffmpeg=F:\tools\ffmpeg.exe"

rem Make sure folders exist
if not exist "%outdir%" mkdir "%outdir%"
if not exist "%avsdir%" mkdir "%avsdir%"

for %%a in ("*.avi")  do call :make_intermediate_422 "%%a" "%avsdir%" "%outdir%"

endlocal
echo DONE & pause
goto :eof


:make_intermediate_422 <videofile,  avisynth dir name,  output dir>
set "avsfile=%~2\%~n1.avs"
rem careful, avisynth needs absolute video filepath
call :make_avs_file "%CD%\%~1"  "%avsfile%"
"%ffmpeg%" -y -i "%avsfile%" -c:v prores_ks -profile:v 3 -bits_per_mb 8000 -pix_fmt yuv422p10 -c:a copy "%~3\%~n1.mov"
goto :eof

:make_avs_file <videofile, avs filepath>
rem for top field first video, loading 4:2:0 by avisource and then change format to 4:2:2
echo AVISource("%~1")               >  "%~2"
echo AssumeTFF()                    >> "%~2"
echo QTGMC(Preset="Slow")           >> "%~2
echo ConvertToYV16()                >> "%~2"
goto :eof

:make_avs_file2 <videofile, avs filepath>
rem for NTSC DV, to loading 4:1:1 by ffms2 and then change format to 4:2:2 directly (without intermediate 4:2:0)
echo FFVideoSource("%~1")           >  "%~2"
echo AssumeBFF()                    >> "%~2"
echo ConvertToYV16(interlaced=true) >> "%~2"
echo QTGMC(Preset="Slow")           >> "%~2"
goto :eof

get rid of setting ffmpeg path if using windows PATH,
also, use subroutines (that is how windows calls functions), it makes structure, makes it modular, also more clear, and no ai would generate it for you like that, people refuse to use it for some reason, so ai cannot copy it from examples ...
and also added a version to load video by ffms2, which loads NTSC DV as 4:1:1, because you change it to 4:2:2, you resize chroma directly into it, not using 4:2:0 which your VFW codec using AviSource might do. Watch field order though, in that script, using ffms2, it is set to BFF after loading avi.

[_Al_]

just realized, there is no audio using :make_avs_file2, so it can be done, using FFAudioSource, but that gave me some access violation (not sure, missing some codec or something), so used AviSource to load audio (bit wild, but anyway):

:make_avs_file2 <videofile, avs filepath>
rem for NTSC DV, to loading 4:1:1 by ffms2 and then change format to 4:2:2 directly (without intermediate 4:2:0)
echo FFVideoSource("%~1")             >  "%~2"
echo AssumeBFF()                      >> "%~2"
echo ConvertToYV16(interlaced=true)   >> "%~2"
echo QTGMC(Preset="Slow")             >> "%~2"
echo AudioDub(last, AviSource("%~1")) >> "%~2"
goto :eof

[Sharc]

Preset 3 was a typo, should have been 2.

No. Read what I wrote.

[Riker0007]

Preset 3 was a typo, should have been 2.

No. Read what I wrote.

Got it. Thanks!

[Riker0007]

It keeps failing: Any ideas?

to get intermediate 4:2:2, in prores this should work, not sure how 10bit conversion can benefit, but anyway:

@echo off
setlocal
rem Output folder name
set "outdir=Converted"
set "avsdir=avs_temp"
set "ffmpeg=F:\tools\ffmpeg.exe"

rem Make sure folders exist
if not exist "%outdir%" mkdir "%outdir%"
if not exist "%avsdir%" mkdir "%avsdir%"

for %%a in ("*.avi")  do call :make_intermediate_422 "%%a" "%avsdir%" "%outdir%"

endlocal
echo DONE & pause
goto :eof


:make_intermediate_422 <videofile,  avisynth dir name,  output dir>
set "avsfile=%~2\%~n1.avs"
rem careful, avisynth needs absolute video filepath
call :make_avs_file "%CD%\%~1"  "%avsfile%"
"%ffmpeg%" -y -i "%avsfile%" -c:v prores_ks -profile:v 3 -bits_per_mb 8000 -pix_fmt yuv422p10 -c:a copy "%~3\%~n1.mov"
goto :eof

:make_avs_file <videofile, avs filepath>
rem for top field first video, loading 4:2:0 by avisource and then change format to 4:2:2
echo AVISource("%~1")               >  "%~2"
echo AssumeTFF()                    >> "%~2"
echo QTGMC(Preset="Slow")           >> "%~2
echo ConvertToYV16()                >> "%~2"
goto :eof

:make_avs_file2 <videofile, avs filepath>
rem for NTSC DV, to loading 4:1:1 by ffms2 and then change format to 4:2:2 directly (without intermediate 4:2:0)
echo FFVideoSource("%~1")           >  "%~2"
echo AssumeBFF()                    >> "%~2"
echo ConvertToYV16(interlaced=true) >> "%~2"
echo QTGMC(Preset="Slow")           >> "%~2"
goto :eof

get rid of setting ffmpeg path if using windows PATH,
also, use subroutines (that is how windows calls functions), it makes structure, makes it modular, also more clear, and no ai would generate it for you like that, people refuse to use it for some reason, so ai cannot copy it from examples ...
and also added a version to load video by ffms2, which loads NTSC DV as 4:1:1, because you change it to 4:2:2, you resize chroma directly into it, not using 4:2:0 which your VFW codec using AviSource might do. Watch field order though, in that script, using ffms2, it is set to BFF after loading avi.

THANKS!!!
This is working. I'm not getting the errors! Thanks for your help.