What can I use to cut 5 frames?

[efc1978]

Hi,
I have 2 DVD's I've losslessly authored with DVD Shrink by cutting the end of one and cutting the start of the other so the same scene from the 2 different sources joins together, however when I demux the VOB's and join the two .m2v's in Muxman there's 0secs.20 / 5 frames repeated, as DVD Shrink is only useful for cutting every 12 frames I'm looking for a free programme to losslessly cut the end or start of one of the two m2v's by 0secs.20 / 5 frames.
Could somebody please recommend a free programme and easy way to losslessly do this?
Thank-You

[DB83]

Maybe you did not cut on an I-frame.

Try avidemux with copy for both audio and video output

[efc1978]

Thanks for your suggestion but I'm pretty certain Avidemux doesn't work with .m2v's.

They're 2 completely different video sources, one official DVD, the other a different sourced VHS, but they're both the same video, so trying to join both at the same moment / frame isn't going to be as easy as using DVD Shrink which only cuts every 12 frames, that's why I need something which will cut more precisely, I know it's 5 frames (0secs.20) I need to cut off the end.

[DB83]

a m2v is just mpeg2 surely. And avidemux works with mpeg2

[efc1978]

I have m2v's made from VOBs using PGCDemux, I don't know how to losslessly make a format that will work in Avidemux, then losslessly convert back to an m2v so I can mix it back with the audio and all the other m2v's using Muxman.

[DB83]

If you are cutting 5 frames from a m2v you are going to lose sync with the audio anyway.

You really need to mux the m2v and the audio together into mpeg2-ps i.e .mpg and then cut the frames as I suggest.

Suggestion2: Create single mpg2 video from the dvd's using vob2mpg and then cut losslessly with avidemux.

[efc1978]

No, the reason I need to cut 5 frames (0secs.20) is because it will put the video in sync with the audio, I already have 2 hours of audio synced to 19 m2v's, separately using each m2v's original audio to sync the new superior audio to by 0secs.01, it all perfect except for this 1 particular join where I have the two m2v's of the same footage but different DVD sources which currently have the 5 frames (0secs.20) repeated.
I simply need to cut 5 frames off the end of one m2v or cut 5 frames off the start of the other, the audio doesn't even have to be part of it, I have it all calculated and otherwise sync is perfect,
I just need to cut 5 frames, it can even be just the video without the audio as I've calculated all frames to match with the audio.

[pippas]

I don't think you will be able to be that exact with an interframe format like Mpeg2 (or m2v). At least not without some re-coding. You can only cut with frame accuracy with an intraframe format (like DV-AVI for example), where every frame is a keyframe.
With mpeg2 you will only get keyframes ('I' frames) at indeterminate and irregular intervals... as you've already discovered...

Alternatively, you can force 'fake' frame accurate cutting with m2v files form within an editor that lets you do 'smart renders'. But that will involve some re-coding, so won't be totally lossless...

[DB83]

Ok. So have you actually tried to use avidemux ?

You might have to save it as -ps and then demux again back to m2v

MpegstreamClip, VideoRedo etc. might be able to write back to m2v.

[efc1978]

Thanks I'll give it s try after the weekend.
But I think I've tried something similar without success using Avidemux some time ago....

[efc1978]

Are there any other free programmes I could use to losslessly make frame accurate cuts to MPEG/m2v not on I frames?, even if they re-encode just the cut end to the next I frame.
Avidemux doesn't seem like it's good to do the job.
Thanks

[smrpix]

There are no free smart renderers -- which is what you're asking for. Both VideoReDo and TMPGenc Smart Renderer have trial periods that may get you over the hump.

[efc1978]

OK thanks.

[jagabo]

You can always smart render manually. Cut the video into chunks saving the uncut GOPs sections without re-rendering, re-render the cut GOPs, then appending all the pieces back together. That's not too big a hassle to remove one small section of one video.

[pandy]

You can always smart render manually. Cut the video into chunks saving the uncut GOPs sections without re-rendering, re-render the cut GOPs, then appending all the pieces back together. That's not too big a hassle to remove one small section of one video.

Depend - relatively easy with closed GOP but it can be difficult for open GOP and also he need to create proper timestamps...

[jagabo]

You can always smart render manually. Cut the video into chunks saving the uncut GOPs sections without re-rendering, re-render the cut GOPs, then appending all the pieces back together. That's not too big a hassle to remove one small section of one video.

Depend - relatively easy with closed GOP but it can be difficult for open GOP and also he need to create proper timestamps...

Yes, of course.

[efc1978]

mmm
It seems even when opening in Avidemux before I even do anything like move the cursor it says the total length of the VOB or m2v file is about 3 frames smaller than it really is.
Why would that be?, would the few frames it's not including in the total frame count be missing from the start or end?
Does this have a reason / explanation?

I've also read conflicting information about which part from A selection to B Avidemux actually saves or deletes (not sure if that is because it works different ways in different release versions?), but I think when cutting a selection A frame will be the first frame deleted, and the frame before B will be the last frame deleted?, does anybody know for sure, and if this applies to all versions?
What about selecting a section to keep?, if keeping the selection worked the same way as I have read deleting a section works (as explained above), then it would mean you can not save the very last frame of a file, am I correct?
Thanks

[DB83]

The A and B are markers for footage kept. It works thus:

Say you want to lose 5 frames at the end, A is marked at the first frame and B is marked the 6th frame from the end. So you get a long coloured line.

Effectively, it is a TRIM

Transversly, to lose 5 frames at the beginning, A is marked at the 6th frame and B is marked at the last one.

[efc1978]

VideoRedo seems like it's working on any frame, I guess it's smart rendering, I haven't looked into much detail yet.
I have a question about the total length of the original files VideoRedo is giving.

I know the last number VideoRedo gives is the frame, and not a fraction of a second, but it still doesn't make sense,
for example the 10 frames on the end of the first example convert to 0secs.40, but that's still 0secs.12 shorter than the real file length.

So example 1, 3 and 4 VideoRedo says the files are 3 frames / 0secs.12 shorter than the original files, example 2
it says the file is 1 frame / 0secs.04 shorter than the original file.

This is very confusing as I know I need to cut 5 frames / 0secs.20 off the start of a file, so where are these 'missing' times cut from?, the start?, end?, what's happening here?

I have a VOB of 9638 frames / 6min25secs.52, but when loaded into VideoRedo the green bar says its 6min25.10 and program, selection and selection and output say 6min25secs.11.

Another example is a VOB 612 frames / 24secs.48 when loaded into VideoRedo the green bar says 24secs.11, and program, section and output says 24secs.12 (well 24secs.12 is actually correct as it works out to 24secs.48 the same time length as the original file).

Third VOB I tried is 8926 frames / 5min57.04, but when loaded into VideoRedo green bar says 5min56.23 and program, selection and output says 5min56.24.

Finally the fourth VOB I tried is 8353 frames / 5min34secs.12, but when loaded into VideoRedo the green bar says 5min34.00, and program, selection and output says 5min34secs.01.

Why does VideoRedo give these time lengths which are different to the 'real' lengths?
This is causing me confusion as I know I need to cut 5 frames / 0secs.20 off the start of my video file to match with my new audio, however as VideoRedo is saying the original file time length is different to what it really is I don't know if I can make my cuts accurate....?

Thanks

[DB83]

It has already been stated that frame-accurate/lossless/smart editing/rendering is only possible on i-frames.

Allow me to make a suggestion.

Upload one of those vobs - the fourth might be the best since that appears to give the greatest variances, as an attachment to this topic to let others use your actual file to see if accurate editing really is possible

[efc1978]

I thought the whole purpose of smart rendering (known as smart copy in Avidemux) was that you can cut any frame, and only the frames from the cut to the next I frame are lossy, and the rest lossless, so you're telling me cutting on a frame which isn't an I frame with VideoRedo will put the entire file through a lossy conversion?, and not just from the cut to the next I frame?

I guess the smart render manually idea should work OK then,
I presume there won't be any visible join when I merge the lossy and lossless part if the 2 files are
ended and started on consecutive frames from the original VOB source, for example if the rendered part ends on frame 9 and the non rendered (losslessly cut part) starts on an I frame which is frame 10 ?

[jagabo]

As long as you don't change codecs in VideoRedo it will smart render. It will only re-render cut GOPs. Eg, MPEG 2 in, MPEG 2 out, smart render. MPEG 2 in, h.264 out -> full render.

[efc1978]

As long as you don't change codecs in VideoRedo it will smart render. It will only re-render cut GOPs. Eg, MPEG 2 in, MPEG 2 out, smart render. MPEG 2 in, h.264 out -> full render.

What does one of the above posts mean I have to create proper time stamps?
I think as long as I use the boxes highlighted in green which show at 1 frame at a time and use them to select start and end to either keep or delete the section, it should smart render properly with the majority being losslessly?
I haven't changed any settings since downloading.
And when you say it will only work for MPEG2 in and out you do mean VOB import is OK don't you?, also experimenting today I think it was exporting VOB as m2v, is that OK for smart rendering?
Thanks

[jagabo]

What does one of the above posts mean I have to create proper time stamps?

That was in reference to open GOPs and manually cutting and pasting. You don't need to worry about that with VideoRedo.

I think as long as I use the boxes highlighted in green which show at 1 frame at a time and use them to select start and end to either keep or delete the section, it should smart render properly with the majority being losslessly?

Yes, as long as you output MPEG 2.

And when you say it will only work for MPEG2 in and out you do mean VOB import is OK don't you?, also experimenting today I think it was exporting VOB as m2v, is that OK for smart rendering?

VOB and m2v are both MPEG 2. And since Pandy is going to be pedantic, VOB can contain MPEG 1 video but that is very rare.

It will be very obvious whether the program is smart rendering or reencoding. With a short video like yours a smart render will take seconds. A rerender will take minutes.

[efc1978]

What does one of the above posts mean I have to create proper time stamps?

That was in reference to open GOPs and manually cutting and pasting. You don't need to worry about that with VideoRedo.

I think as long as I use the boxes highlighted in green which show at 1 frame at a time and use them to select start and end to either keep or delete the section, it should smart render properly with the majority being losslessly?

Yes, as long as you output MPEG 2.

And when you say it will only work for MPEG2 in and out you do mean VOB import is OK don't you?, also experimenting today I think it was exporting VOB as m2v, is that OK for smart rendering?

VOB and m2v are both MPEG 2. And since Pandy is going to be pedantic, VOB can contain MPEG 1 video but that is very rare.

It will be very obvious whether the program is smart rendering or reencoding. With a short video like yours a smart render will take seconds. A rerender will take minutes.

Thanks for the info.
Yes the few tests I did today it took just a few seconds, which is why I was pretty sure it was smart rendering, as I said in my earlier post this evening, I'm not sure why DB84 responded to my earlier post by saying "It has already been stated that frame-accurate/lossless/smart editing/rendering is only possible on i-frames".
I think DB84 has misunderstood/mistyped something there...

Oh I just remembered, I read something where somebody was suggesting that Avidemux would only smart render the start of a video (but not the end), will VideoRedo smart render both the start and end if both are needed?
Thanks

[DB83]

Who is DB84 ?

Ok. Cut on a b-frame or a p-frame at the start of a video and see if you can smart render. You might be able to and do not come back and state you are seeing rubbish in the screen.

If you want any more help from me in this topic you will honour my suggestion and post that sample vob.

But, to be frank, I do not see how this editing will ultimately help you even if full sync is now achieved (many people would not be able to detect a 1/5 of a second delay). When you come to link all those vobs back together (if that is your intent) to sync with the 2 hours of video, you end up re-encoding it. Or rather the authoring program will.

[efc1978]

Who is DB84 ?

Ok. Cut on a b-frame or a p-frame at the start of a video and see if you can smart render. You might be able to and do not come back and state you are seeing rubbish in the screen.

If you want any more help from me in this topic you will honour my suggestion and post that sample vob.

But, to be frank, I do not see how this editing will ultimately help you even if full sync is now achieved (many people would not be able to detect a 1/5 of a second delay). When you come to link all those vobs back together (if that is your intent) to sync with the 2 hours of video, you end up re-encoding it. Or rather the authoring program will.

No, nothing is re-encoded, the VOB's are all output as .m2v's which I then combine with my new audio in Muxman, it's all lossless apart from the start and end cuts to I frames of 3 very small sections totalling less than 1 minute!
Yeah I agree not many people will notice 5 frames / 0secs.20 out of synch, but the rest of my 2 hours each song is individually in synch to 0secs.01 with the old original audio, so I'm just being a perfectionist by making sure this final part is as perfect as I can get it, I probably wouldn't notice this tiny bit out of synch myself, but when comparing side by side with the correctly in synch original audio I can see the slight differences, so might as well fix it considering the rest of the 2 hours is perfect.

[poisondeathray]

5 frames at 25FPS will be obviously out of sync (200ms) . 2 frames is about the threshold . The threshold is about 40-80ms for most people

Oh I just remembered, I read something where somebody was suggesting that Avidemux would only smart render the start of a video (but not the end), will VideoRedo smart render both the start and end if both are needed?

avidemux doesn't smart render mpeg2 , only mpeg4-asp (xvid, divx) . videoredo can smart render mpeg2 anywhere, but it can't accept elementary video (not m2v, it has to be multiplexed)

Another free oldie is cuttermaran . It could smart render m2v , when you had the free tmpgenc version 2.5 installed .
https://www.videohelp.com/software/Cuttermaran

The other option is to add audio, but I guess that doesn't make sense if it's the same scene and you're doing a middle section

[efc1978]

5 frames at 25FPS will be obviously out of sync (200ms) . 2 frames is about the threshold . The threshold is about 40-80ms for most people

The other option is to add audio

Another free oldie is cuttermaran . It could smart render m2v , when you had the free tmpgenc version 2.5 installed .
https://www.videohelp.com/software/Cuttermaran

Yeah I noticed when doing a few test runs today it wouldn't accept m2v as input but would VOB,
although I'm pretty sure I did successfully output one of my tests as m2v.

5 frames / 0secs.20 being obviously out of sync depends what the video footage is showing I guess. In this case of the small section I'm editing there's not really any close ups of instruments, and just a pretty short couple of shots of a singer speaking.
However if the footage was out 5 frames / 0secs.20 for example when a drumstick was being shown hitting the drums I would find that quite obvious and annoyingly out of sync!

[poisondeathray]

5 frames / 0secs.20 being obviously out of sync depends what the video footage is showing I guess. In this case of the small section I'm editing there's not really any close ups of instrumental, and just a pretty short couple of shots of a singer speaking.
However if the footage was out 5 frames / 0secs.20 for example when a drumstick was being shown hitting the drums I would find that quite obvious and annoyingly out of sync!

Speech will be visibly out of sync as much as drumsticks . Or did you mean the speaking has no vocals (speech actually isn't heard), only background music ?