I have a PAL source that is blended pretty badly.
srestore(frate=23.976) gets rid of it or at least reduces it pretty well.
I slowed the audio down to match the NTSC audio pitch and this increased the length of the audio track. It now does not sync up with the video after doing this.
What I would like to do, is add in duplicate frames (after the srestore removed the blended frames) in order to pretty much make it 25fps video again so I can then use assumefps and make it playback at 23.976fps. This will get the video and audio back in sync.
It syncs perfectly with the slowed down audio if I do not use srestore at all or use srestore(25) and use assumefps to make it playback at 23.976fps, but then it is still blended to where its annoying due to srestore not being used.
My goal is to remove the blends with srestore(frate=23.976) then put back in some duplicate frames instead of blended frames to make it as it was to begin with and have less blending. Then I wish to slow the video framerate so that it matches the slowed down audio.
How would I go about adding in duplicate frames after removing blended ones to make it 25fps again if this can be done?
+ Reply to Thread
Results 1 to 18 of 18
All this assumes the 'base' framerate before blending was really 23.976fps. If it's not, then you used SRestore incorrectly and that's the source of your problem.
It syncs perfectly with the slowed down audio if I do not use srestore at all
Srestore(frate=23.976) didn't change the video length but slowing the audio down to 23.976 speed/pitch increased the length of the audio. The audio is now longer than the video. I need to increase the length of the video in order to syn them up.
The audio is now 23.976fps, which is slower than 25fps like it was with the pal video. The video after srestore is also 23.976fps, problem is, the length was not changed at all like with the audio. Thats why I want to make it 25fps again, but this time not use blended frames so the blending is reduced.
I slowed down the audio to match the NTSC voice pitches, and the 25fps video or srestored 23.976 fps video will not sync with it. It syncs fine with it when assumefps(24000,1001) is used without srestore but if I try using it on the srestored version, it wont sync because its still the same length of the VOB video.
[QUOTE=manono;2414393]Meaning the audio included with your source wasn't in synch? How can the source video synch with slowed/lengthened audio?[QUOTE]
No it syncs with the source just fine, I was talking with the source having assumefps(24000,1001) used on it. The slowed down audio syncs with the video having assumefps used and no srestore used, but if srestore is used, it no longer syncs with it because the video is shorter length than the slowed audio and is the same length as the source video before assumefps was used.
Last edited by killerteengohan; 15th Oct 2015 at 02:23.
Source video = 22:04 length
Original PAL audio = 22:04 length
Slowed audio = 23:01 length
assumefps video = 23:01 length
srestore video = 22:04 length
assumefps+srestore = 22:04 length
Only the assumefps without srestore syncs with the slowed audio and it plays back at 23.976fps like the NTSC version both in video and audio, but all the blending is still in the video. Thats why I wanted to reduce the blending, then make it the same as before with duplicate frames added in to replace the blended ones, so when assumefps is used, it will sync up and not have the blending in it.
Ive never done this before and if stutter is bad or annoying then I will have to try something else or deal with the higher pitch audio. I want to see how it turns out, but did not know how to do it.
I will decide if I hate the outcome or not if told how to do it.
Last edited by killerteengohan; 15th Oct 2015 at 02:44.
You didn't say you used an 'AssumeFPS(23.976)' on the original PAL video before going through the unblending process. That's a really dumb thing to do because you'll screw up the unblending. You'll get a 23.976fps video anyway as a result of using SRestore(Frate=23.976) so what's the point of the AssumeFPS(23.976)? And what's the point of slowing the audio if using SRestore? The audio pitch should already be correct.
I would do it because the audio pitch for PAL is sped up and higher pitch and I am used to USA voice pitches
1. 23.976fps is the wrong unblended framerate
2. It's both blended and truly victimized by PAL speedup.
In both cases 23.976fps isn't correct. How do you know the audio is higher pitched? Because you compared it to NTSC versions with the same voices? Do you have an untouched video sample?
Thanks, that seems to have made it the correct length.
Thats why I want to make it 25fps again, but this time not use blended frames so the blending is reduced.
Have you tried unblending it to 25fps followed by an AssumeFPS(23.976)? Then you should be able to use your slowed-down audio.
I agree with getting the audio back to the correct pitch and do it all the time myself. But what you're doing isn't the way to go.
Yes, I tried doing it to 25fps then assumefps afterwards, and it did not get rid of the blending. It took srestore(23.976) in order to reduce the blending. But I was told in another thread that doing that was stupid to do and that I should use the srestore(23.976) because it would make it 23.976fps the same exact way as srestore(25) with assumefps right after, only it would do a slightly better job at removing the blends. It was claimed that using srestore(25) wouldn't do anything and that it would keep the same framerate as the source.
I dont see the point in using srestore(25) if the source is already 25fps. Its not even going to do anything is it??
Apparently manono Theres a slight misunderstanding from your reading if you are under the impression that I did that and used assumefps first then unblended. Let me try to clear it up some more here. I already said I did the unblending then the assumefps and it didnt sync up. I even gave the video length of it in a previous post. I would never do it in reverse like that and use assumefps first and then unblend. I never said I used assumefps first then unblended afterwards. I did all of these separate into separate files using the same source and checked them all after they were done.
The NTSC source is blended as well, The PAL DVD's have PAL pitch/speed audio and video speedup.
Just trust me on this, its been sped up in both video and audio for the PAL release. Normally PAL seems to speed up the audio pitch, and then make video 25fps by adding in extra frames. In this case it seems they sped up the audio and sped the video up to 25fps without adding in any extra frames. When I use srestore to lessen blending, its removing frames and making it less frames than NTSC version because it was just sped up without extra frames added in from the looks of things. If all I do is use assumefps and nothing else, its the same length as the NTSC DVD episode at 23:01 and it syncs up perfectly with the slowed down audio.
I dont know how to dumb it down any more than this. Here is what Ive done and my goal with each.
The PAL version is sped up audio with higher pitch. It plays back faster than the NTSC audio if you play them side by side. So I slowed it down. There, now I have same pitch and speed of audio as the NTSC copy by slowing down the PAL audio. I played them side by side and they sound exactly the same pitch and are the same length.
Now the video is the issue. I used just srestore(frate=23.976) in one file and it syncs perfectly with the original PAL audio but not the slowed down audio.
I then used only assumefps(24000,1001) on the same source into another file and the video synced perfectly with the slowed down audio. problem is, it still has the blending in it.
I then used srestore(frate=23.976) and then assumfps(24000,1001) right after in the same script and its the same output as with just srestore alone, so the assumefps is pointless in this file.
Okay then at this point, just assumefps without srestore is the only one syncing with the slowed audio.
So I want to reduce blending with srestore(frate23.976) and since its shorter length than the slowed down audio, I need to make it longer by adding frames back in afterwards so its the same length as the slowed down audio when I use assumefps to reduce it from 25fps (This is now with dupe frames instead of blended) down to 23.976 again.
This PAL source doesnt seem to be NTSC 23.976fps with added in blended frames to make it 25fps, it seems to be the original NTSC 23.976fps just sped up to 25 fps, so the video and audio is playing faster than the NTSC DVD. The NTSC has the same exact blending as the PAL version only the PAL is playing back faster in both audio and video when compared side by side.
So basically, the original NTSC got sped up to 25fps with no added in extra frames. When I use srestore on it, the reduces the blending by removing frames and then makes the playback rate 23.976fps, only now its 23.976fps with less frames than before. Thats why I want to put the frames srestore removed back in, only instead of the blended ones, I want them to be duplicates. That way I get 25fps again with less blending in it. (This may cause stuttering as you mentioned and as I originally thought, but I wish to see this and see if Im willing to accept it or not) Then when that is done, I wish to use assumefps to make it 23.976 and this time it will sync up with the audio. It would just like using assumefps by itself before on a 25fps source, only the blending would be gone.
Make any more sense to you now what it is I have done and am wanting to do??
I didnt wanna argue about this with anyone, I wanted a simple answer as to how I can add back in a few duplicate frames to get what "I" want and see if I like the output or not. Im looking for how to do this specific thing, not for any and every reason on why I shouldn't do it or why you wouldnt do it.
Last edited by killerteengohan; 15th Oct 2015 at 07:42.
You keep talking about changing the video to match the audio. That is absolutely the wrong way to fix sync issues. Instead, you do whatever you have to do to eliminate duplicates and blends in the video, and to get it to play at the correct speed (where a video of a digital clock will keep perfect time when played back), and then you use an audio stretch plugin to make the audio match, both in speed and in pitch. In audio, you can adjust both the speed and the pitch independently of each other.
There is no right or wrong way to do a lot of things, but in this case there is a wrong way, and changing the video speed to match the audio is the wrong thing to do.
audacity and goldwave.
I changed speed to a slower, proper speed so its the same as the NTSC audio and the voice pitches went down from the speed slowing down and the length of the audio got longer because it plays back at a slower rate. The programs even mention that changing speed will affect both tempo and pitch. This fixed the voice pitches, but made the entire track longer as a result of slowing it down. (How are you supposed to lower the pitch, and yet not increase the length of the overall track from the slow down to change the speed of playback??)
I saw a change pitch option in audacity and lowered it -4.096%, but when I tried it and exported it, the audio had some stuttering/prolonged spoken words in it in a few parts. It didn't sound right anymore in some parts for about a second or less, even though the pitch seemed perfect. It did keep the same exact length though and not get longer from the change, which makes it perfect. Do you know how to do a pitch change in Goldwave?
EDIT: I tried the pitch change again and this time the output was fine. it changed the pitch to sound as I wanted it and the length of the track was not changed. I'm not sure how it did that because I was under the impression if you slow audio to restore an audio pitch that was sped up to begin with it would become longer, but things seemed to work out fine here.
Last edited by killerteengohan; 15th Oct 2015 at 16:54.
Apparently manono Theres a slight misunderstanding from your reading if you are under the impression that I did that and used assumefps first then unblended.
It's not a continuation of another thread in my opinion. That thread I was asking how to get rid of blending, this thread all I simply wanted was to know how to put duplicate frames into the video to make my video 25fps again. (My reasons are my own) Why you couldn't just tell me how without all the fuss, and trying to point out whats wrong with doing it is beyond me. I wasn't asking for the correct way here or for help making something better, I just wanted to know how to do what I wanted to do, so I could try it out and see if it's output is acceptable to me or not. Not read smartass or immature remarks that waste both our times.
The sample should still be up to my knowledge at the top of that other thread where I wanted to know how to reduce the blending. You should be able to get it there. If not, I will put one up again.
It still looked kind of blended though or at least the srestore(frate=23.976) looked better before I added that changefps and assumefps in so I decided not to use it.
I went ahead and used the change pitch tool in audacity instead of the change speed tool and it seems to have fixed the voice pitches without altering the length of the audio track. How it did that I'm still not sure but I'm not gonna complain since it did what I needed.
So I used that changed pitch audio with the srestore(frate=23.976) without changefps or assumefps and it syncs perfectly now from the looks of things.
So I got what I wanted and problem is solved mostly thanks to pitch change instead of speed change in the audio track.
Usually converting PAL audio and video similar to NTSC is a cinch and it syncs up fine, but this particular source was a pain in the ass.
Thanks for telling me how to do what I wanted to try!
Last edited by killerteengohan; 15th Oct 2015 at 17:04. Reason: Spelling Correction
Audacity is supposed to be pretty good) also then give you the option of keeping the pitch the same, which is what you just figured out how to do.
One note: if you do this for a living, some programs don't do a very good job at creating clean audio when keeping the pitch the same. If you make a really big change, like plus or minus 25%, you can determine how well your software can perform this operation. The really, really good programs (usually costing at least $500) do not create hollow flanging artifacts.
To give you an audible example of the difference between good software and average software, here is a zip file containing two short audio clips that were slowed down by 50% (half speed). One was done using cheap software, and the other was done using professional software.
Speed Change Example