There was a user who wanted a wav-file to be played once a second. This was not fully possible since files were too long. After looking at the file more closely it was clear, the TTS-service add’s a bit too much silence to the files so it needs a bit of fixing.
Issue and Solution
Let’s have a look on two audio files in visual form. This is Text To Speech -generated audio files from text “RC-Thoughts”, the unfixed version on top, fixed version below:
Looking at the unfixed version “Untrimmed” on top we can see extra silence in the beginning of file and especially in the end of the file. I examined about 50 different audio-files you users have generated and the amount of extra silence is really constant. So I added fixed trimming.
The “Trimmed”-file below has 0.12 seconds trimmed from the beginning and 0.75 seconds from the end. And the result is just about perfect! Also note that total length of the file is reduced from 1.8 seconds to about half, ~0.92 seconds. So yes, goal achieved!
Best part is that you do not have to do anything, fix is automatically done!
Enjoy the improved free Text-To-Speech here!
It took you 36 minutes to come up with an post a solution to the user request?
You're getting slow :)
As always, thanks!
Yeah, sorry it took so long! :D