Rockbox mail archive
Subject: There *is* hope for a cleaner spoken voicefont.
From: Christi Alice Scarborough (christi_at_chiark.greenend.org.uk)
Two pieces of good news wrt the spoken voicefont.
I've been mailed offlist by someone who has tried Nero 6's Noise
Reduction on the samples, and the results are, if not quite perfect,
pretty good. There doesn't seem to be any way to automate this process
Also, I've faked up a "poor woman's pre-amp setup" using a Realistic
33-9031 500 ohm dynamic microphone connected up to my ageing (12 year
old!) stereo's mic input and headphones output connected to the sound
card's line in. This does, as others have suggested, produce a much
Given the clipping present in the first voicefont, my instinct tells me
that it would be worth re-recording. It might even be possible not to
have to hand clip the samples this time, which is more time consuming
than the recording and, it appears, not done very well by me. Also, if
I'm going to do this properly, I should probably invest in a mic
pre-amp. (They're a not too unreasonable 20 pounds or so from Richer
I'm pretty certain the samples weren't as badly clipped in the initial
WAVs as they sound on the Jukebox. Could the MP3 encoding process have
caused greater clipping? Does the JB definitely play the whole sample?
Finally, are there enough potential users of the spoken voicefont that
it's worth me putting in another day or so's effort on re-doing it?
I'll not be doing anything to it for a week or so anyway, since I'm
currently working on the manual for 2.3.
Page was last modified "Jan 10 2012" The Rockbox Crew