Elastic Audio - How does it do what it does?

[ihate100bees]

I just started to dabble into using Elastic Audio, experimenting to change a tempo of a complete song. Seems like a good feature to have in your tools box.
But... what does it really do to the files and what else can you do with it that is useful?

Someone on a YouTube video said that it works by "Quantizing"? Should I be aware of artifacts or another type of deterioration to the files, by pushing it up about 8 - 10 bpm's? Does it thin out the sound of Audio (guitar/bass/voc) tracks?

Tell me what you know about using Elastic Audio and what does it do to the files?

Thanks

[massivekerry]

https://youtu.be/PTko9Q4pf0w

https://youtu.be/yYoQSRAHhIc

https://youtu.be/sKSncdygAas

[nednednerb]

It recalculates the same levels across a different time.

Imagine taking 44.1 kHz audio and playing it twice as fast but keeping it at 44.1kHz.

It takes the same levels (bit depth), but takes half the minutes:seconds to output those levels.

If it's the same levels and the calculation is good (ie. see Elastic Audio quality and make your own opinions!), you ought to hear the same audio at double the frequency.

Some algorithm of pitch and time compression try to offset the final frequency so in the above example, the original and output pitch would be the same, and theoretically the timing only *should* change.

That might be by some kind of frequency shift math. So the algorithms are doing this, in practice for little drum hits and whole passages to comform hits, quantize notes or beat match sections. If you group all your drum mic audio tracks, then EA them together, they should keep in phase. Can use EA to match a drum loop's groove with a different rhythm. Can use EA for vocal lines, to synch in the length (timing) of a held vowel without chopping with another edit tool. That can be more natural maybe..

So many things to do with EA that are sound experimental and weird too, like very exaggerated slow downs, etc.

Some of the "artifacts" sound like "robotic style" but imagine changing the pitch of your human voice by 2x. Since typically that refers to our vowel vocalization, and NOT the sibilant and consonant frequencies, 2x the frequency of those parts is not going to be natural sounding, hence performing the trick is more than running just one omni algorithm.

[ihate100bees]

Thanks so much for this explanation in detail.

[ihate100bees]

Thank you for sharing these videos about Elastic Audio in Pro Tools!