Audio import/export (wav) problem

[ggoblin]

... According to Youtube help their recommended audio sample rate is "96khz or 48khz" ...

There is also a support page for YouTube that states.... "Sample rate 44.1 kHz recommended. Higher sample rates are accepted but not required (for example, 48 kHz or 96 kHz)" - https://support.google.com/youtube/answer/6039860 But I think this is old/older.

I think that recommendation is for 'music videos' - due to prevalence of music CDs, I think they are kinda still stuck at 44.1kHz? But for all other videos its 48kHz as thats what most video recording devices today will record at.

[ggoblin]

Perhaps I should clarify, I'm not saying we shouldn't use other software, I certainly need to use other software. But personally what I find is it costs more to fix something if I have to cross software boundaries so my workflow minimizes this.

So for example once the voice over is cleaned and normalized in a sound editor, and sound effects trimmed and cleaned, they never returns back to the sound editor once they are imported into Moho. In Moho the voice-over gets chopped - pauses added as needed by animation - and animation is lipsynced to it, and sound effects synced to animation. Once video is exported from Moho to video editor it never returns back to Moho. In the video editor final touches are added like background music and subtitles, titles, credits, etc. For my small 2 minute social media projects which may have a deadline of only 8-10 days of which half is spent on the script, thats an efficient workflow. The idea of re-syncing the audio in the video editor once its already been synced in Moho (because audio in moho is only for 'reference') is time I don't have.

I appreciate many pros here will have other workflows to suit their needs.

But all of this is besides the point, whatever our workflow differences may be, Moho has no excuse to degrade the audio. And its an easy fix to implement - just let FFmpeg do its job.

[synthsin75]

I still can't see it here, with 48 kHz wav:


And even converting that to MP3:

[ggoblin]

I still can't see it here, with 48 kHz wav:

Thats why I would still like to see the spectrograms of the original audio before it was imported into Moho to compare with the spectrograms of the exported audio for both 44.1kHz and 48kHz.

Since I didn't personally notice the degradation I checked my original voice-over wav file only to discover it had already gone through a filter to remove high frequencies which are not used in speech as part of its clean up before it was imported to Moho, so I would not expect much degradation in any case. I will look again when I get a chance.

[synthsin75]

I still can't see it here, with 48 kHz wav:

Thats why I would still like to see the spectrograms of the original audio before it was imported into Moho to compare with the spectrograms of the exported audio for both 44.1kHz and 48kHz.

Since I didn't personally notice the degradation I checked my original voice-over wav file only to discover it had already gone through a filter to remove high frequencies which are not used in speech as part of its clean up before it was imported to Moho, so I would not expect much degradation in any case. I will look again when I get a chance.

Both of those (the wav and mp3) are the original stereo above the Moho exported stereo audio.

[ggoblin]

Okay I've done some tests.

Original wav used: 48kHz PCM wav
https://docs.google.com/uc?export=downl ... Fa9ICMKu1O
from:
Download Sample .WAV File For Testing
https://www.appsloveworld.com/download- ... or-testing
(only first 10 secs used)

48kHz TEST

1. Spectrogram of original 48KHz before import to Moho:



2. Spectrograph of output from Moho (original 48kHz): MP4 file with compressed audio at 128kbs 44.1kHz - serious degradation clearly visible especially in the higher frequencies



3. Spectrograph of output from Moho (original 48kHz): AVI file with PCM audio 44.1kHz - still serious degradation clearly visible especially in the higher frequencies




44.1kHz TEST

4. Spectrograph of original 48kHz converted to 44.1kHz wav file in sound editor - before import to Moho:



5. Spectrograph of output from Moho (original 41.1kHz): MP4 file with compressed audio at 128kbs 44.1kHz - degradation is a lot less than (2) & (3) when original is 48kHz



6. Spectrograph of output from Moho (original 44.1kHz): AVI file with PCM audio 44.1kHz - degradation is less than (5),(2)(3) - much closer to original




Conclusions:

There is serious degradation of 48kHz audio irrespective if the output from Moho is compressed 128kbs MP4 file or PCM audio in AVI / PNG wrapper. If anything the degradation is more visible in the PCM output suggesting mp4/aac compression smooths some of the degradation over.

For best output from Moho: choose 44.1kHZ input and export to codec which uses PCM audio (any of the codecs with [PNG alpha- PCM] in the export menu).

This totally concurs with Alexx's findings.

[ggoblin]

I still can't see it here, with 48 kHz wav:

Thats why I would still like to see the spectrograms of the original audio before it was imported into Moho to compare with the spectrograms of the exported audio for both 44.1kHz and 48kHz.

Since I didn't personally notice the degradation I checked my original voice-over wav file only to discover it had already gone through a filter to remove high frequencies which are not used in speech as part of its clean up before it was imported to Moho, so I would not expect much degradation in any case. I will look again when I get a chance.

Also checked again, noticed my original voice-over wav file was actually 44.1kHz.. so that explains why I didn't notice any degradation

[bebemustacio]

For best output from Moho: choose 44.1kHZ input and export to codec which uses PCM audio (any of the codecs with [PNG alpha- PCM] in the export menu).

This totally concurs with Alexx's findings.

Thank you ggoblin for tests. This is exactly how we start our next episode.

We take original sound files and downsample it from 48kHz to 44.1kHz, then do moho magic, export WAV and picture, import everything in AE to make composting with 3D BGs , export uncompresed WAV from AE, and finally send it to Alexx who will make all sound editing and mixing (in sound editor) and export it as 48 khz...

This is still the best workflow for us, no trying to make final sound in animatic, no double syncing ...

cheers!

[Greenlaw]

Sorry guys, I still don't ,understand where this idea of 're-syncing' or 'double-syncing' in editorial is coming from. In the video editor, when I overcut the animatic scenes with my final rendered scenes, the footage is already in sync. I literally drop the footage on a scene marker and it's in sync. I never have to 're-sync' anything because my animation was based the reference exported from the animatic to begin with. It's very simple and mostly foolproof.

I also don't understand why this workflow is considered only for a 'big studio'. This is the same workflow I've been using in my personal one-man projects since the early 2000's, back when I had a fairly modest computer and hardly any budget. For me, this is just a very efficient and low-cost way to work, which is generally why many 'big studios' have a similar workflow.

I'm not saying everyone must follow the same workflow--you should follow what works best for you--but so far the arguments being made against working the way I do as an 'indie animator' haven't made much sense to me.

Anyway...I do agree, Moho cans use some improvements to how it handles audio files, and I'm sure that's coming. Probably not for the 13.5.x cycle (they must be done with 13.5.x by now) but eventually.

[synthsin75]

Sorry guys, I still don't ,understand where this idea of 're-syncing' or 'double-syncing' in editorial is coming from. In the video editor, when I overcut the animatic scenes with my final rendered scenes, the footage is already in sync.

Doesn't sound like they're doing an animatic at all. Sounds like they're just figuring out the timing as they animate...which certainly must take longer.

[synthsin75]

I see the problem now. I had been doing export audio from right-clicking the layer, which apparently is just a copy of that layer's audio. Not sure why this exists.

[Daxel]

I agree that animatics are the way to go, but I don't think that every animation has to have an animatic and that it has to be set on stone, so I understand why one could have to re-sync.

Moho is actually very permissive with experimentation and modifications on the spot, because we are usually working with a rig that we can easily pose to see how something feels, while going back to change the animatic hand-drawing a couple of frames could take way more time for no good reason.

[synthsin75]

I agree that animatics are the way to go, but I don't think that every animation has to have an animatic and that it has to be set on stone, so I understand why one could have to re-sync.

Moho is actually very permissive with experimentation and modifications on the spot, because we are usually working with a rig that we can easily pose to see how something feels, while going back to change the animatic hand-drawing a couple of frames could take way more time for no good reason.

If you stray from your own animatic and have to re-sync, that sounds like a problem with you own workflow. And changing the animatic AFTER starting animation sounds even worse.

A better option would be to follow your animatic (finalized before animation starts), and then cut or do pick up shots for final edit. Now, you can do added stuff while animating, but you'd be much better off doing a camera cut to add them to the edit later. If there are not enough cuts to do so, it may be a sign that the animation could risk being boring, with overlong shots.

[ggoblin]

I see the problem now. I had been doing export audio from right-clicking the layer, which apparently is just a copy of that layer's audio. Not sure why this exists.

Yeah I tested it as well, I though maybe it would include timeline information like what frame it starts playing, when it is muted by not being visible, audio levels keyframes, etc. But no, it ignores all that and just makes a copy of the original sound file you just imported!

Pointless, and misleading. I suppose its one way, the only way, to get Moho to export a 48kHz audio file

[synthsin75]

Pointless, and misleading.

Agreed.