Audio Question...

[zerobyte01]

I have a clip that was encoded to 23.976 fps and am writing an avisynth script to convert these types of clips to DVD NTSC 29.97 with 3:2 pulldown done with HCenc; Does this affect the audio?(I assume it would) And if so what would I need to use in my script to sync audio with the video?

Also, using the cli mediainfo I saw there was an offset to the audio of -5ms in the original clip....so would this offset need to be applied after solving the above?


I woud appreciate any input thanks

[Bernix]

Hi,
AFAIK audio is fps independent, so there shouldn't be a problem. And 5ms is not noticable, you can ignore it.
Edit: even much bigger offset is acceptable. Try calculate what distance sound did, when you know it is roughly 300m/s

Bernix

[zerobyte01]

thank you Bernix for the info much appreciated

[jagabo]

When you use 3:2 pulldown (any pulldown, really) you care duplicating frames (as fields) to achieve 59.94 fields per second. So there is no change in running time.

5 ms audio skew is not noticeable. Keep in mind that 23.976 fps is about 42 ms per frame.

[JVRaines]

Perceptual studies show that lip sync errors of -20 to +40 milliseconds have no effect whatsoever.

[smrpix]

Perceptual studies show that lip sync errors of -20 to +40 milliseconds have no effect whatsoever.

While I'm not condoning it -- internet videos have given folks an even wider tolerance than that.

[pandy]

Studies shows that people are able to accept even over -100 ... +200ms lipsync delay (more sensitive for audio advancing video, distance from display is very important as audio is way slower than light so few meter distance from display may trig unacceptable lipsync).
However this is less important than main problem - if audio duration is the same as video duration there should be no problems. Duration is a key framerate and audio delay may be related to packet structure - various data packets are interleaved within container and frequently some offset is added to compensate packet layout (where most of packets are video data)...

Perceptual studies show that lip sync errors of -20 to +40 milliseconds have no effect whatsoever.

While I'm not condoning it -- internet videos have given folks an even wider tolerance than that.

This is old Dolby requirement - never are different for Dolby MS11 this is -20 ... +30ms but this is for HW/SW vendors applying for Dolby certification and Dolby i would say is not so very strict on this.

[manono]

I have a clip that was encoded to 23.976 fps and am writing an avisynth script to convert these types of clips to DVD NTSC 29.97 with 3:2 pulldown done with HCen

There's nothing more for AviSynth to do if the video is already 720x480. Let HCEnc take care of applying soft pulldown. Just check the "3:2 pulldown" box and hit "Make DVD compliant". You certainly don't want to create hard 3:2 pulldown in a script.

And if so what would I need to use in my script to sync audio with the video?

I don't believe HC-Enc even handles audio. It's purely an MPEG-2 video encoder. If the delay bothers you, extract the audio and remove the delay using DelayCut and mux it with the M2V that HC-Enc creates when muxing with Muxman or whatever you use. Or add in the delay when muxing. But as mentioned several times, 5ms is unnoticeable by even the sharpest ears.

[Bernix]

Sorry for this,
isn't it rather sharpest eyes? It is about lipsync, so eyes seems to me be more important. Of course deaf people and lipsync is bit problematic. But probably wrong again. And sorry for this, even if I'm right.


Bernix

[manono]

isn't it rather sharpest eyes? It is about lipsync, so eyes seems to me be more important. Of course deaf people and lipsync is bit problematic.

And blind people won't do well trying to fix lipsynch errors, either. But I'll stick with what I said. Someone else might disagree or say it differently. jagabo and you both phrased it as "not noticeable" because, I think, both the eyes and ears work together in this. When I try and fix it just by listening and watching, I believe the ears are more important. When I want to fix a delay error more accurately, I use an AviSynth filter to do it and the eyes are all that's needed.

[Bernix]

Hi Manono,
glad you take it easy. I just want to say, that even if you have 10% of hearing abilities, you hear the sound at exact time as anybody else with sharpest ears. But if you have 10% of seeing abilities, you can have problem to see things that makes sound. For example lips. So therefore sharper eyes seems to me be more important. Of course both senses are important, just from my view is seeing bit more important. Thats all.
But this is very little related to OP, but don't want to bother you with with P.M. Just clarifying my reason what is more important in A/V sync. Sharper seeing/hearing.


Bernix

[Hoser Rob]

If you have a video with a frame rate of 23.976, say, one frame lasts about 40 msec. So talking about audio delay of under 40 msec being significant doesn't make a lot of sense to me.

I've found the type of video makes a difference. With a lot of them it's hard to manually sync them, but with some it's easy. I have a concert video that's all basically funk with dancers on the stage. They're all moving ... you can't play fink without moving something ... and they all know exactly where the downbeat is. That one was a snap to sync.