Ask HN: Is Adobe Podcast Audio Enhancer Currently Broken?
The AI powered audio enhancer [0] seems broken for me -- tried multiple times and every time it takes a long time to process and reports an error at the end:
> The audio file is in a format we can't understand. Ensure the file is not corrupted or convert it to a standard wav or mp3 format to try again.
I used this service without issue a few days ago. Not sure what's changed now.
0. https://podcast.adobe.com/enhance
32 comments
[ 1.5 ms ] story [ 69.6 ms ] threadThanks for your patience! We’re really sorry about the inconvenience.
Thanks for acknowledging and working on it. As a unpaid customer I can surely wait :)
Can I sneak in a couple of questions while I have your attention?
1. The upload says it only accepts MP3 and WAV -- but it seems to accept AAC as well in the past when we force it to? Also, why not support all other commonly used audio formats as well - should be (relatively) simple to add a conversion step?
2. Can we get some controls on how aggressive the noise filtering applied is? The result in some cases sounds like abrupt cuts, flubbed words, speaker suppressing a hiccup or burp etc :))
Oh and a suggestion --
If the issue we saw yesterday was due to a service being down or broken, ideally the error message displayed should say so -- so users don't assume the issue is on their side with their input file and try repeatedly and waste their own time.
EDIT:
The service now seems to be up but the AI processing is very poor. The resulting audio has so much of warbling/trembling artifacts on every syllable that it cannot be used at all. The original recording even with the noise and reverb from my room is much more listenable and understandable.
I hope you have some standard tests for your algorithm to measure teh quality of audio output for a given input clip
Thanks!