So I captured some PAL VHS footage at 30 fps and...

[Soundguy]

I know it should have been 25 fps. Is this a huge deal? I've actually recorded several hours and would rather not have to re-do it all. I have a bunch of VHS gameplay tapes I'm currently archiving.

I've re-recorded some samples at 25 fps for comparison's sake and in practice can see little difference between the two, though I've analysed the footage frame by frame and noticed that the 30 fps captures include some unique frames not in the 25 fps captures and vice versa, though the 30 fps seem to have more unique frames overall.

I assume unique frames are missing with either framerate due to my VCR and capture device not being perfectly in sync, not much I can do about that. Capturing at 50 fps includes all unique frames, with a lot of double ups obviously, but as these captures are SD and destined for YouTube, they get cut down to 25 fps anyway, and the frames that are missed in the 25 fps get cut also. So whatever I do I'm going to lose frames.

The game itself is on the choppy side anyway (GTA:SA on PS2), making any jerkiness caused by framerate mismatches hard to spot.

All this considered, what is my best course of action?

1) Scrap everything and re-record at 25 FPS (would rather not, have already uploaded several hours of footage)

2) Keep recording at 30 fps, bearing in mind this seems to capture more unique frames and is the highest I can go for SD on YouTube

3) Not worry about the 30 fps I've already done, but switch to 25 fps for the remainder of the tapes (about 10 more to go)

Any advice appreciated.

Samples below, note timecoded parts SFW but other parts of the video might have NSFW language

25 fps sample - https://www.youtube.com/watch?v=rK9V0c_6LaU&t=33s

30 fps sample - https://www.youtube.com/watch?v=or9McuCHtQo&t=1406s

[dellsam34]

And what? I don't see the first post, It may need approval from the mods I think.

[Soundguy]

I did some edits and the post vanished, though it's back now.

[Alwyn]

Are ads really necessary on videos you are asking for technical comments on?

[Soundguy]

Are ads really necessary on videos you are asking for technical comments on?

I assume you're not a YouTube content creator. The ads aren't mine, it's GTA, it has licensed music, any ads are beyond my control.

[DB83]

And uploading vids to youtube for help is not going to help you that much since yt will have re-encoded your upload.


BTW You do not get 'unique' frames if your capture a 25 fps source at 30 fps. You get duplicate frames. If the frame you see and interpret as 'unique' is not a duplicate of the previous one your capture is corrupted. Whether that is noticeable at normal playback you are your own judge.


But it may also be useful if you need more assistance were you to state the hardware/software used for the capture.

[Soundguy]

And uploading vids to youtube for help is not going to help you that much since yt will have re-encoded your upload.

Yes, but the frames of motion still match the source despite that, plus that's the versions viewers will see. It was just to provide a comparison for the two different framerates, I can't see a huge difference personally, but I wanted some second opinions. YouTube can sometimes drop frames when it encodes, but that's rare, trust me I've analysed and compared source to YouTube many times.

BTW You do not get 'unique' frames if your capture a 25 fps source at 30 fps. You get duplicate frames. If the frame you see and interpret as 'unique' is not a duplicate of the previous one your capture is corrupted. Whether that is noticeable at normal playback you are your own judge.

I think you misunderstood me. Yes, there are duplicate frames in the 30 fps capture, that goes without saying. When I say "unique", I meant frames that contain differential images, unique frames of motion. In one second of 25 fps there might be 21 frames that show a unique point of motion, with four that show dupes (it's a choppy game, so will sometimes run at less than 25 fps). Whereas the equivalent second at 30 fps might have 23 unique points of motion, with 7 dupes. So even though the 30 fps capture contains more dupes, it also contains more of the motion data. Hope that makes sense.

And like I said, there will be frames of animation captured in 25 fps that don't show in 30 fps and vice versa. The only way I can capture every point of motion from the tapes is to use 50 fps, but 50 fps SD videos aren't a thing on YouTube.

Summed up, the 30 fps capture is grabbing frames from the tape that the 25 fps skips, and vice versa, though the 30 fps seems to be skipping less of them, by a marginal amount.

But it may also be useful if you need more assistance were you to state the hardware/software used for the capture.

OBS software, just a VCR hooked up to a USB capture, nothing fancy. I have a handle on the set up, that's not what I'm seeking advice for. Just want to know want other people would do in my situation, option 1, 2 or 3.

[DB83]

It is not a Q. of mis-understanding. It is a question of the use of the word 'unique'. Whether it is the device or the software that creates a frame dis-similar (maybe a better choice of wording) to the previous frame there is corruption.


OBS is the worst possible software to capture VHS. Don't take my word for it. There are tons of topics on here. It is, at best, a screen-capture program and a streaming AP.


Anyway for proper analysis do a sample capture at 25 fps and the same footage at 30 fps and upload these direct to the forum as attachments.


Mind you I am not a gamer and know zilch about how the consoles output their video and, equally important, if there is a constant progressive stream independent of broadcast standards. What I allude to is if the console outputs video at 30 fps which you have recorded to PAL VHS at 25 fps - not, in my eyes, a typical process - the recording has lost frames and maybe that reverse process is even more confused.

[jagabo]

Capture at 50 fps, Then TDecimate() to 25p or 30p.

[Soundguy]

It is not a Q. of mis-understanding. It is a question of the use of the word 'unique'. Whether it is the device or the software that creates a frame dis-similar (maybe a better choice of wording) to the previous frame there is corruption.

I still don't think you're understanding me, it's not corruption, it's certain frames on the tape that are being grabbed at different points in accordance with the frame rate used for capturing. 30 fps will grab some frames that 25 fps doesn't and vice versa. I'm assuming it's because the VCR and capture device aren't synced, so the capture device starts grabbing frames at an arbitrary point from when I hit record, that doesn't perfectly line up with the frames on tape. I've explained it as best I can, apologies if it's still not clear.

Recording at 50 fps fixes that, like I said

OBS is the worst possible software to capture VHS. Don't take my word for it. There are tons of topics on here. It is, at best, a screen-capture program and a streaming AP.

That may be so but that's not the issue here, it was my fault for inadvertently using a 30 fps template instead of 25 fps, my main concern is how to proceed from there, again, option 1, 2 or3.

Anyway for proper analysis do a sample capture at 25 fps and the same footage at 30 fps and upload these direct to the forum as attachments.

Can do, but trust me, the frames will match the YouTube vids linked, if not the fidelity of the image. I've already analysed them extensively though, I know what's going on in them frame by frame. I just want to know if anyone spots a night and day difference from a casual view.

Mind you I am not a gamer and know zilch about how the consoles output their video and, equally important, if there is a constant progressive stream independent of broadcast standards. What I allude to is if the console outputs video at 30 fps which you have recorded to PAL VHS at 25 fps - not, in my eyes, a typical process - the recording has lost frames and maybe that reverse process is even more confused.

They are old tapes, recorded years ago, typical for back then. The console is PAL, like the VCR equipment used, it may have lost frames from source but I can't see GTA on PS2 running higher than 30 fps tbh. That's not the issue though, the issue is how to proceed with capturing what I already have on tape.

[Soundguy]

Capture at 50 fps, Then TDecimate() to 25p or 30p.

So do that with what's remaining and leave what's been done? or scrap everything and do that starting from scratch?

I'm not encoding these files btw, I want to upload the source files "as is". I'm losslessly editing in Avidemux for that reason. YouTube butchers them enough without more encoding my end, if TDecimate works losslessly then I'll look into it, if not then I'd rather suffer the frame loss. In which case Just want to know which of the 3 options from my OP I should go for.

[krykmoon]

So do that with what's remaining and leave what's been done? or scrap everything and do that starting from scratch?

I'm not encoding these files btw, I want to upload the source files "as is". I'm losslessly editing in Avidemux for that reason. YouTube butchers them enough without more encoding my end, if TDecimate works losslessly then I'll look into it, if not then I'd rather suffer the frame loss. In which case Just want to know which of the 3 options from my OP I should go for..

There's a lot of confusion, here.

Soundguy.. First thing first : OBS isn't capturing your source footage at all... simply because it can't. OBS is a software to capture WHAT's on your screen (whatever it is a game, your desktop or a video playing). So it will be NEVER be in sync with your original footage whatever the frame rate you may choose. That's why is highly not advisable to use OBS for tasks like this. The videos you will get in the end with OBS will ALWAYS be radically different from your source footage, since capturing the source footage IS NOT what OBS does. It can't.. OBS captures in realtime the pixels that are on your screen (*), nothing more than this, regardless (and totally unaware) of the source footage.

Second: if you "capture" your video with OBS... IT IS ALREADY heavily encoded, so using a lossless codec at this stage after you edited the video to preserve image "fidelity" with the source and quality has no purpose at all, in particular considering that such fidelity (and quality) has never been in existence from the start (how I explained before).

To sum things up: the point is that with this method you'll simply generate a lot of HUGE videos that are totally different from your original source footage.

Now the choice on what to do is yours.

If you're happy with this and you don't care too much about preserving the original material.. you can go on this way. But at least use a lossy codec to encode the "captured" video after editing it, using a lossless one has not purpose at all with this kind of workflow.




(*) NOTE: to be 100% accurate and precise about how OBS works.. it captures the pixels from its "render buffer".

[Soundguy]

[QUOTE=krykmoon;2619900]

So do that with what's remaining and leave what's been done? or scrap everything and do that starting from scratch?

I'm not encoding these files btw, I want to upload the source files "as is". I'm losslessly editing in Avidemux for that reason. YouTube butchers them enough without more encoding my end, if TDecimate works losslessly then I'll look into it, if not then I'd rather suffer the frame loss. In which case Just want to know which of the 3 options from my OP I should go for..

There's a lot of confusion, here.

Soundguy.. First thing first : OBS isn't capturing your source footage at all... simply because it can't. OBS is a software to capture WHAT's on your screen (whatever it is a game, your desktop or a video playing). So it will be NEVER be in sync with your original footage whatever the frame rate you may choose. That's why is highly not advisable to use OBS for tasks like this. The videos you will get in the end with OBS will ALWAYS be radically different from your source footage, since capturing the source footage IS NOT what OBS does. It can't.. OBS captures in realtime the pixel that are on your screen, nothing more than this, regardless (and totally unaware) of the source footage.

This is just a more technical description of what I've already said, yes, it's not perfectly in sync, I established that in the first post, but "radically different"? Not really, there are minor differences at best between the 25 fps, 30 fps, and 50 fps captures. Capturing at 50 fps will include all the frames the tape has, all the motion, if not perfectly in sync. I can't do that for YouTube so will have to accept some frame loss. No-one has yet said whether they can see actually see a night and day difference between the two samples I posted, which was one of the main reasons I posted.

OBS is capturing what's on my screen, which is a signal from the capture device, which is coming direct from the VCR. Syncing issues aside, it's capturing what's on the tape, so I don't see the issue with describing it as such. That said, if anyone has suggestions for a better VHS capture app then fire away, I use OBS because it was the only thing I could find that worked with my device and was free. Edit - I dug out the disc that came with my capture device, it has software on it, might give it a try and see if that makes a difference.

I feel like I'm being a little nit-picked here tbh, no-one has said what they'd do out of the 3 options I described either, and people are pouncing on things I say, that have little to do with my issue, without understanding what I meant, like this...

Second: if you "capture" your video with OBS... IT IS ALREADY heavily encoded, so using a lossless codec at this stage after you edited the video to preserve image "fidelity" with the source and quality has no purpose at all, in particular considering that such fidelity (and quality) has never been in existence from the start (how I explained before).

To sum things up: the point is that with this method you'll simply generate a lot of HUGE videos that are totally different from your original source footage.

Now the choice on what to do is yours.

If you're happy with this and you don't care too much about preserving the original material.. you can go on this way. But at least use a lossy codec to encode the "captured" video after editing it, using a lossless one has not purpose at all with this kind of workflow.

When did I say anything about using a lossless codec? I'm using Avidemux to edit the source files losslessly, so I don't have to re-encode the file and lose quality, that's it. Even if the files aren't great quality to begin with (they aren't) why would I want to lessen it even more?

[krykmoon]

This is just a more technical description of what I've already said, yes, it's not perfectly in sync, I established that in the first post, but "radically different"?

The frame rate you specify in OBS is NOT related in anyway to the frame rate of your source material.. it's its "sampling" rate. To be clear: OBS has absolutely not knowledge at all about your original footage, it simply gets pixel from the video player engine that's rendering that source..
It does it in realtime, at the specified frame-rate .. Even not considering the possibility of having dropped frames... the frame rate at which OBS is "capturing" the video may be exactly the same (as a value) as the source footage, or a perfect multiple of it.. but the twos will never be truly in sync..

To make it easiest as possible to understand: OBS simply put every 1/fps seconds into the video it is capturing a new "frame".. getting the pixels from what it is stored in its video "buffer" at that instant .. and that doesn't reflect exactly and precisely where the source footage REALLY is at that instant.

So for every frame generated from OBS you can always have as an outcome one of those : (a) the same exact frame as in the source material at that instant; (b) a previous frame in the source (in this case it will be a duplicated frame) or (c) a following frame (that's when you got a dropped frame).

That's what will happen for every 1/fps inside OBS for all the duration of your footage: for every frame generated by OBS you may have either an (a), (b) or (c) (randomly, since your footage and OBS will never be perfectly in sync) .. that's why I wrote the "captured" video will be RADICALLY different: because.. IT WILL.

And that's only one aspect of the matter.. there's also the aspect of the rendering/filter video chain.

So this is the reason why OBS is absolutely the worst tool to do any kind of "footage" capture.

When did I say anything about using a lossless codec? I'm using Avidemux to edit the source files losslessly, so I don't have to re-encode the file and lose quality, that's it

I see.. When you said "edit" I thought about full editing, not simply doing some minor cuts.. Now I got the point.

[dellsam34]

Again, OBS is not for analog video capture, it captures your computer's video card canvas or screen not from the capture card's ADC based on what I've read about it.

[DB83]

Yes, you asked for advice of either 1,2 or 3. But such advice should also come with the 'science' behind it.


That is what is being attempted here. Whether you accept it or not.


So here is my 'advice' on your options


1. Preferred.
2. A non-starter.
3. A compromise.

[krykmoon]

Again, OBS is not for analog video capture, it captures your computer's video card canvas or screen not from the capture card's ADC based on what I've read about it.

OBS is not for any kind of "footage" capture, analog or digital, since it will ALWAYS get pixels from a place that is different than their true source and it can't assure any kind of sync with the source..

People trying to capture with it streaming content from the web (which is a digital source) have the same identical problems .. and nasty results.

It's a very good software for real time capturing in games and for real time streaming: but you can't use a hammer to slice bread.

[Soundguy]

This is just a more technical description of what I've already said, yes, it's not perfectly in sync, I established that in the first post, but "radically different"?

The frame rate you specify in OBS is NOT related in anyway to the frame rate of your source material.. it's its "sampling"rate. To be clear: OBS has absolutely not knowledge at all about your original footage, it simply gets pixel from the video player engine that's rendering that source..
It does it in realtime, at the specified frame-rate .. Even not considering the possibility of having dropped frames... the frame rate at which OBS is "capturing" the video may be exactly the same (as a value) as the source footage, or a perfect multiple of it.. but the twos will never be truly in sync..

To make it easiest as possible to understand: OBS simply put every 1/fps seconds into the video it is capturing a new "frame".. getting the pixels from what it is stored in its video "buffer" at that instant .. and that doesn't reflect exactly and precisely where the source footage REALLY is at that instant.

So for every frame generated from OBS you can always have as an outcome one of those : (a) the same exact frame as in the source material at that instant; (b) a previous frame in the source (in this case it will be a duplicated frame) or (c) a following frame (that's when you got a dropped frame).

That's what will happen for every 1/fps inside OBS for all the duration of your footage: for every frame generated by OBS you may have either an (a), (b) or (c) (randomly, since your footage and OBS will never be perfectly in sync) .. that's why I wrote the "captured" video will be RADICALLY different: because.. IT WILL.

And that's why OBS it's absolutely the worst tool to do any kind of "footage" capture.

OK, thanks, I understand more about why it's not syncing, I'm going to try the software that came with the device, not sure what it is, as the disc is pretty nondescript, but maybe it will do a better job at capturing all the frames at 25 fps, in which case I'll use that from now on.

I'd still like to know if anyone can see a noticeable difference between the two samples, and if they think it worth starting from scratch if I can improve the way I capture, or maybe it's not worth it.

You can analyse the frames till the cows come home, but if in practical terms there's no discernible difference to the casual viewer, in what is already quite a choppy game, would you go through all the hassle?

[Soundguy]

Again, OBS is not for analog video capture, it captures your computer's video card canvas or screen not from the capture card's ADC based on what I've read about it.

OBS is not for any kind of "footage" capture, analog or digital, since it will ALWAYS get pixels from a place that is different than their true source and it can't assure any kind of sync with the source..

People trying to capture with it streaming content from the web (which is a digital source) have the same identical problems .. and nasty results.

OK point taken, people seem to love OBS in other circles so I was swayed I guess, I'm literally using it because it recognised my device and was free. I'm checking out another app as we speak. Any suggestions I'd be pleased to hear them, free though, I only have a few tapes to record, so it's not worth me splashing out on anything fancy.

[Soundguy]

Yes, you asked for advice of either 1,2 or 3. But such advice should also come with the 'science' behind it.


That is what is being attempted here. Whether you accept it or not.


So here is my 'advice' on your options


1. Preferred.
2. A non-starter.
3. A compromise.

That's cool, I appreciate the input, did you A B the YouTube samples? Can you see a noticeable difference between the 30 fps and 25 fps? Enough to be concerned with? It's a choppy game, like I keep saying, so I'm just wondering if re-doing everything will make enough difference to be worth it.

At the moment I'm leaning towards 3, but if enough people say they can see a huge improvement with the 25 fps version (crappy OBS syncing aside) then I might go for 1, though it will be a massive pain. Best to do it now before the videos gain traction though.

[DB83]

I glanced at them but not being a gamer I do not hold any real opinion on them.


Now were you to do a sample record of normal, non game-play, video with spoken dialogue from actors with visible faces I might then appreciate the difference.


You asked for an alternative capture app. The one currently in vogue is AmarecTV. You do need the older version and thus avoid the obligatory codec that the newer version uses. But all is explained, or should be, on the download page.

[jagabo]

Some free video capture programs: VirtualDub, AmarecTV, ffmpeg.

[Soundguy]

I glanced at them but not being a gamer I do not hold any real opinion on them.


Now were you to do a sample record of normal, non game-play, video with spoken dialogue from actors with visible faces I might then appreciate the difference.


You asked for an alternative capture app. The one currently in vogue is AmarecTV. You do need the older version and thus avoid the obligatory codec that the newer version uses. But all is explained, or should be, on the download page.

OK fair enough, it sounds like you have a more discerning eye than the majority of YouTube viewers, so if you couldn't spot anything glaring that's a good sign. Obviously the footage is grainy and rough, those tapes are nearly two decades old, and the game is blurry and choppy at source anyway, so it's futile me getting bogged down about the image quality, it's not good and isn't expected to be. It's the framerate stuff I'm most concerned with. Thanks for the software recommendation, I'll look into it.

[Soundguy]

Some free video capture programs: VirtualDub, AmarecTV, ffmpeg.

Thanks, I think I have VirtualDub and ffmpeg, though I couldn't figure them out years back when I first downloaded them, maybe I should take another look, I think VD wouldn't recognise my device iirc, but I have a new one now.

[Soundguy]

OK, so I had a look at some of the other apps.

VirtualDub doesn't want to play nice, the image I get is very choppy in the preview screen, and it will only display the picture on one of my mirrored monitors, the other just shows a black screen. Not sure what's up with that, still a bit confused by the interface, took me a while to even to get the signal displayed as the option I needed was buried in a mass of menus.

Honestech VHS to DVD is the software that came bundled with my device, it's simple to use and seems to sync a lot better to the source, the framerate is definitely smoother in the captures I made, but it's not perfect, there are still frames present in my original OBS captures that are being skipped. It's also lacking in options for adjusting the quality, it has a few presets but little else. The image seems sharper than the OBS captures, but the colours and contrast are a lot darker for some reason. Not sure if I like it, the captured image seems a bit cheaper and blockier in places too, like a low quality encode, even though it's actually running at a higher quality than I had with OBS, could just be because I'm not used to it.

AmaRecTV, took a while to figure it out, it's a Japanese app and it's English doesn't always make sense. It gave me a garbled mess and delayed sound at first but seems to have sorted itself out. It could be the one I stick with as I like the image quality in the viewer, but it defaults to a 720x480 resolution and won't let me change it, that's not a resolution I've ever used before, it's not the same as the other PAL settings I've used at least. If that's the proper resolution I should be using for that device then fine, though it wasn't an option in the manufacturer's own software

My main issue is that it won't let me capture anything, I keep getting a "can not support colorspace by video codec" message every time I hit record, I have no idea what that means or how to fix it. It will say that whatever codec I choose, many of the ones it displays are "not install" anyway, and I have no idea how to install them or where to get them from for this particular app. I got the AMV4 codec, but that's lossless and gave me the same error when I tried to record.

Any ideas how I can fix this "colorspace" issue?

[jagabo]

The good thing about VirtualDub is that's it's very flexible and has many options -- it can be made to work with just about any capture device. The bad thing about VirtualDub is that it's very flexible and has many options -- it can be very hard to get it working properly with some devices. Be sure not to play the audio while capturing -- it will cause sync problems like those you are seeing. You can capture the audio, just don't play it (Audio -> Enable Audio Playback, disabled). Also don't compress the audio while capturing, and use a fast lossless compression codec (UT Video Codec or Lagarith). Oh, and be sure to use VirtualDub, not VirtualDub2 which is reputed to have problems with video capture (I can't confirm that myself).

For all capture programs, set the capture device to capture YUY2 or other YUV 4:2:2 format. Use a codec that keeps that format. Note that Amarectv doesn't show all installed codecs until you press the Update Codec List button.

[DB83]

Isn't that AMV codec the one I warned you about in my earlier post ?


Go back to the download page and read the comment about installing version 3.1 NOT version 4.xx (maybe that is no longer an install option but the codec is still pay-ware but may have a limited default ie 720*480 which is NOT what you want)


The Honestech will inevitably only capture as mpeg2 with the provided software

[Soundguy]

The good thing about VirtualDub is that's it's very flexible and has many options -- it can be made to work with just about any capture device. The bad thing about VirtualDub is that it's very flexible and has many options -- it can be very hard to get it working properly with some devices. Be sure not to play the audio while capturing -- it will cause sync problems like those you are seeing. You can capture the audio, just don't play it (Audio -> Enable Audio Playback, disabled). Also don't compress the audio while capturing, and use a fast lossless compression codec (UT Video Codec or Lagarith). Oh, and be sure to use VirtualDub, not VirtualDub2 which is reputed to have problems with video capture (I can't confirm that myself).

For all capture programs, set the capture device to capture YUY2 or other YUV 4:2:2 format. Use a codec that keeps that format. Note that Amarectv doesn't show all installed codecs until you press the Update Codec List button.

Thanks, yeah I have the original VirtualDub, not 2. I'll go back and see if the audio was causing that choppiness. Though I would prefer to be able to monitor the audio as I often leave the tapes recording while doing things around the house, if I can hear it playing I know the tape is still running. I always record to lossless audio if I have the option, audio is actually my background, the video stuff I just sort of fumble my way through, I have no formal background in it (as if that wasn't obvious). I just like recording and editing gameplay, was recording it to VHS long before sites like Twitch and YouTube became a thing.

I'm still having no luck trying to fix this "colourspace" issue though with Amarectv, yeah I pressed the update codec list, some of the ones that show up are "not install" though, I'm not sure where to get them from. The one I currently have active is IYUV as that seems to be the closest match to the ones you mentioned.

[Soundguy]

Isn't that AMV codec the one I warned you about in my earlier post ?


Go back to the download page and read the comment about installing version 3.1 NOT version 4.xx (maybe that is no longer an install option but the codec is still pay-ware but may have a limited default ie 720*480 which is NOT what you want)


The Honestech will inevitably only capture as mpeg2 with the provided software

No I don't think it's the same codec, probably an earlier version, as that's commercial and this was just grabbed by the software and installed for free. I have 3.1, but every time I hit record I get that "colourspace" error, any ideas what that is and how to fix it?

No matter what codec I choose or what settings I make, the video info in the bottom right says 720x480 50 fps, it did say 50i for a while, so I'm not quite sure what's happening, to make it even more confusing I have the viewer set to fixed 4:3 and it displays at 868x651.

Yeah it's mpeg2, I probably won't use the Honestech software, I'd like to get this AmaRecTV working and see what the captures are like, if they're good I'll probably stick with it.

[DB83]

Not something I have personally used but since I have, by the recc of others, promoted, it I will test it myself.


Not tonight since the red grape is talking but watch this space.