Audio import/export (wav) problem

[slowtiger]

I f someone with a recent Moho version could please import some WAV (bit of music, voice), export again, and open both in Audacity for comparison?

[synthsin75]

I f someone with a recent Moho version could please import some WAV (bit of music, voice), export again, and open both in Audacity for comparison?

So the original wav is 2.52MB and the exported wav is 10.1MB, but the only obvious difference is that the exported audio is now stereo.

Setting the export dialog to RAW PCM also doesn’t fix the problem. It is not the matter of the export - rather Import and playback engine of the Moho itself.

Apparently the OP knows it's only playback, which should not be expected to be fully uncompressed quality.


And here's the comparison for RAW AVI:

[ggoblin]

Wes has made two very good rational points - audio playback in Moho tells us nothing about the internal sound quality Moho is processing, and secondly, the way to resolve this should have been to simply produce an example of an uncompressed WAV imported and exported as uncompressed WAV showing degradadation. Instead we got side tracked into discussions of proper workflow and whether Moho is a sound editor.

I have very little knowledge about audio, but when Bebemustacio mentioned that in particular the high frequencies were degraded.. sounding metallic, alarm bells rang as I know the way mp3 and other audio compression works is by butchering the higher frequencies. And that kinda explained why I had not noticed any audio distortion in Moho as in my workflow only the voiceover and sometimes very small special effects are done in Moho with the music (which doesn't need exact syncing to the animation) added later in the video editor. Talking human voice is not a high frequency and audio compression is designed specifically to leave human voice alone so is seldom effected by audio compression unless the compression is very extreme.

However the screen shots Wes has produced show the volume, not the frequency, of the sound so are not very helpful in determining if compression is taking place. A frequency spectrum view, where the high frequencies are visible, would be more useful. An interesting discussion on the subject is available on stackoverflow:

https://stackoverflow.com/questions/252 ... compressed

Perhaps Bebemustacio can share audio files which demonstrate the problem, and which can then be viewed under a frequency spectrum to determine the level of degradation? The article above even links to a python program which put a number value on the level of degradation.

[Maestral]

Audacity has a spectrogram available

Click on the triangle opens a menu / Multi-view/Waveform/Spectrogram / where Multi-view shows both


Hope you'd not mind me asking again - is adding additional codecs to Moho possible? Either audio or video?

[synthsin75]

Aside from one being stereo, they look the same (wav vs raw AVI):

[ggoblin]

Thank you Wes.

They can be difficult to tell apart. The example given in the article shows the compressed stereo spectrum first with the uncompressed below:



The only difference in their 320kbs mp3 example is in the background noise above 16kHz. where there is a clear cut off in the compressed file.

Of course it also makes a difference where the 'uncompressed' source comes from. If its not been recorded by yourself then it might already be compressed, in which case we might not see any difference. In such cases there can still be a tell-tale sign as shown in this second example:



This again is a 320kbs mp3, but they dont have the original uncompressed version. "You can see the very hard limit at 20kHz as well as a milder cutoff at 19kHz. Were it lossless, that red blob in the middle would extend all the way up to 22kHz since the sample rate is 44.1kHz."

[bebemustacio]

Thank you Wes, and all other guys who join this discussion.

Our sound designer preparing examples and explanation and we will post it here soon.

[Greenlaw]

Sorry, I still don't understand why the production isn't mixing in the video editor. This way the same source audio is used before and after Moho, and it makes Moho's potentially altered audio quality irrelevant.

IMO, audio imported to Moho should be used for reference only, and that audio should come uncompressed directly from the video editor. When you use uncompressed audio, it will be exactly frame accurate in Moho to what's in the video editor. When the animation is completed, there's no reason to save the audio track along with the footage because the original audio already exists in the editor.

Just a suggestion.

[bebemustacio]

Sorry, I still don't understand why the production isn't mixing in the video editor. This way the same source audio is used before and after Moho, and it makes Moho's potentially altered audio quality irrelevant.

IMO, audio imported to Moho should be used for reference only, and that audio should come uncompressed directly from the video editor. When you use uncompressed audio, it will be exactly frame accurate in Moho to what's in the video editor. When the animation is completed, there's no reason to save the audio track along with the footage because the original audio already exists in the editor.

Just a suggestion.

To avoid more work than it is necessary. Why to sync again sound/voice that is already perfectly synced to animation. We are talking about bigger projects like 10 min x 26 episodes.

[Maestral]

Perhaps making an animatic in some other sw, which could preserve audio quality - then use that for reference in Moho. Animation and animatic should match - and it would be an easy final edit.
This animatic could be really simple (stills, key poses...), just to help with timing for audio clips. Not necessarily fully developed. If, by any chance, that "other sw" is the one you use for final edit - even if animatic and animation don't match - you'd have the audios separated on the timeline and fine-tuning is not much of a trouble.

I'd also check if there's a file type or some sort of specific settings for those raw/individual audios, which would be more "welcomed" by Moho.

[ggoblin]

Sorry, I still don't understand why the production isn't mixing in the video editor. This way the same source audio is used before and after Moho, and it makes Moho's potentially altered audio quality irrelevant.

IMO, audio imported to Moho should be used for reference only, and that audio should come uncompressed directly from the video editor. When you use uncompressed audio, it will be exactly frame accurate in Moho to what's in the video editor. When the animation is completed, there's no reason to save the audio track along with the footage because the original audio already exists in the editor.

Just a suggestion.

I guess that depends on your workflow. If your animatic is set in stone and you simply plug your animation in then yes you can discard the audio in Moho and everything will be in sync perfectly when you go back to the video editor. But if your workflow is more flexible allowing your animators a little leeway to move segments of audio by a few frames on the timeline for a better animation result, then you are relying on Moho not to screw up the audio, as that is the audio you will use in the video editor.

Also please have a little regard for those of us one-man teams working to very tight deadlines who dont have the luxury of an animatic.

For example my workflow for a typical 2 minute explainer talking head video on a tight deadline (6-10 days) is as follows:

1. Produce script
2. Produce simple storyboard which breaks script into scenes and shows how each scene will build up
2. Get voice over done
3. Open Moho, load voice over audio. Split voice over into sentences and scenes, leaving gaps for scene transitions, and the sound effects you dont wont playing over the voice-over.
4. Lipsynch characters with voice over, one sentence at a time to avoid 'drag' effect. Adjust as needed.
4. Create animation, scene by scene, guided by voice-over segments positioning. At end of each scene try tightening the scene if the animation allows by move audio closer and making animation snappier (for social media the shorter the animation, the better - microsoft study showed human attention span dropped from 12 secs to just 8 secs from 2000 to 2015 due to social media uptake, facebook is reporting 2 sec attention span for video content! ). Add any sound effects for that scene before moving onto next scene in case things need to be moved to accommodate the sound effect. The transitions between scenes are all done in Moho (often for continuity the transition will not effect all elements on the screen).
5. Export animation with synced audio to video editor. Place subtitles, watermark, etc, Add background music. Adjust audio levels as required. Export final video.

At each step we may discover issues which mean we have to go back to the previous step to fix, this can get expensive if we dont catch something and have to go back several steps (for example when we test video on a person not related to the project and they miss an important point from the video message we may need to go back and emphasis that point by changing motion graphic, or by adding a pause for example to give a person time to digest it).

I am always very grateful for suggestions to improve my workflow.

[ggoblin]

Hope you'd not mind me asking again - is adding additional codecs to Moho possible? Either audio or video?

Was there a specific codec you were looking for?

I was pleased with the recent versions of Moho adding a wide range of video export options. But I dont understand why we can't configure these options, for example set audio and video bitrates for lossy compressed formats like H.264 ? Or have I missed something?

Audio export (of whole timeline) would be a welcome addition.

[Maestral]

Guess my remark about animatic was misread.
I do believe animation works better with a 70-30 ratio where 70 is preparation and 30 realization. I've heard about this ratio but I've also tried the other ones so you don't waste your time on experimentation

Animatic does not really have to be fully developed but more like an animated storyboard. Though, it would be handy to have a developed storyboard - meaning: having (almost) all shots in it. Usually, that kind of animatic can help in defining potential issues. Of course, some issues would arise along the way but not really that many to set you back a step or two.

[ggoblin]

I do believe animation works better with a 70-30 ratio where 70 is preparation and 30 realization. I've heard about this ratio but I've also tried the other ones so you don't waste your time on experimentation

Animatic does not really have to be fully developed but more like an animated storyboard. Though, it would be handy to have a developed storyboard - meaning: having (almost) all shots in it. Usually, that kind of animatic can help in defining potential issues. Of course, some issues would arise along the way but not really that many to set you back a step or two.

I think currently I'm working with a 50-50 ratio, so for a 8 day project 4 days are spent on nailing the script and doing the story board.

I agree that for a sophisticated animation an animatic is essential, and will save time further down the work flow. But I am not convinced that in my typical type of animation I would see any dividend for the investment I make in an animatic. If I gave myself a day to create an animatic it would only leave say 2.5 days for animation and 0.5 day video editing. Thats still only 62 - 38 ratio . To achive 70-30 I would have less than 2 days to do the animation

Also I feel some of the benefit of an animatic is lost when doing motion graphic type of animation as until you see the final effect you cant really appreciate it (as everything is on timing) so it would be difficult to judge if the animatic works until the animation effect it describes is actually animated, and you would want the freedom to change it during animation if it doesn't work. Like you said "Animation and animatic should match", that would be difficult for motion graphics unless the animatic is very sophisticated and includes the motion graphics as part of it, in which case it would take much longer to do the animatic.

[slowtiger]

Most of my jobs usually skip the animatic anyway: there's the script, there's the storyboard (very sketchy), that's it. I speak the layout (temp) sound myself, mostly because client can't decide wether to pull out the money for a professional voice artist (in the end I often speak the VO myself, which is not too bad). Final timing of everything is done in video editor.