Increasing the playback speed

[gizmo_1]

I'm working on some interviews. The most goal is to shorten the playback times, so I did all the possible cuts. Final media will be DVD-PAL.

Now I need to increase a little bit (25%) the overall clips speeds without audio distortion but I don't know where to start.

I did a search but I found only topics about changing fps but not the final duration.

Could someone point me the way?

THX

[Boulder]

25% isn't a little bit IMHO;)

AssumeFPS() + SoundStretch are your friends.

[gizmo_1]

Thank you for your advise.

After a test we decide to not use this approach.
The impression is too much forced for this particular job.

Anyway I was wondering how could I encode in MPEG2
such a file (not 25fps)?

THX

[Mug Funky]

you'll have to put:

leakKernelBob(order=1, threshold=4)
converttoyuy2()
convertfps(50)
assumebff() # probably unnecessary, but depends on input maybe
separatefields()
selectevery(4,1,2)
weave()
assumetff()
#bob() # sanity check for field-order - uncomment to check playback

at the end of your script.

depending on your encoder, you may have to convert to yv12 again too.

PAL DVD obviously only accepts 25fps (at 720, 704, or 352 by 576 lines)

encode interlaced, top-field-first.

[edit]
btw, if you can't get the play time down any more, then you could just encode at a lower bitrate to fit it on the disc. interviews = talking heads = compress very well on their own, and artefacts like GOP blinking are forgiven (use open-GOP to help there, if possible). so go crush it like you've never crushed before :). with a good mpeg-2 encoder you can get away with 3.5mbps for it at full D1 size. at half-D1 you can go lower (but don't use this with subtitles).

[Boulder]

Actually if bitrate is the issue, deinterlacing would be a good option to try. You don't need to have a video-look with interviews. Also some denoising would reduce the bitrate demand quite a bit.

[Mug Funky]

my experience with talking heads and a decent mpeg-2 encoder is that deinterlacing doesn't really help at all. the big part of the screen is static anyway, and deinterlacing harms motion-detection. there's probably sod-all difference, but my all-time pet hate with video is unnecessary deinterlacing. 3.5 mbps with a talking head is enough if the encoder is any good.

but i use an encoder that's designed specifically for messy stuff like interlace and even field-blending (though it's pretty old).

denoising may help, but it depends on the source (it might just make it look bad - moving faces are the enemy of temporal denoisers).

if your source is a 3-chip camera with good lighting and a simple background, denoising is not necessary. if it's an outdoor interview with trees in the background, shot on Hi-8 with a shaky, grainy, 1-chip camera, then denoising would be a good idea :)

[Boulder]

Originally posted by Mug Funky

if your source is a 3-chip camera with good lighting and a simple background, denoising is not necessary. if it's an outdoor interview with trees in the background, shot on Hi-8 with a shaky, grainy, 1-chip camera, then denoising would be a good idea :)
Not to mention the Deshaker and/or DePan method;)