Does anyone have any “best practices” type recommendations on audio file sizes? E.g., keep it to 1MB or less per file?
A skill of mine that was experiencing absolutely no crashing when there were short audio files and Alexa responses now crashes at erratic moments after longer audio clips have been loaded. They are below Amazon’s 240-second and 10MB per file limits, but my theory is that the larger files still are problematic.
Where are you uploading/storing your audio files? Also, is this short audio (i.e. under 240 seconds, and part of the speak block feature)? Or long audio (i.e. using stream blocks)?
But I still notice more errors ( “this skill produced an unexpected response”) than I do when using the same Voiceflow structure and only text (Alexa-spoken) responses. If you have any other tips I’m all ears.
In my case, I found that the problem wasn’t the length of the individual audio files. Instead, it was that skills are not allowed to have more than five consecutive audio files lined up to play in a row without a chance for user input.
If that possibility of 5 in a row exists in any of the possible paths forward, you may intermittently trigger that problem at a point earlier in the chain.
Your explanation is very clear, and this was my exact issue. Now it works. Thanks. I do see what they are getting at, ensuring a baseline of interactivity.