[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Speech-reco] RE: [VINUX-DEVELOPMENT] Test speeding up LPC synthesized v
[Speech-reco] RE: [VINUX-DEVELOPMENT] Test speeding up LPC synthesized voice
Mon, 4 Oct 2010 22:37:52 -0400
Interesting that it breaks down by orders of magnitude disproportional to the
In other words ... The 4x one seems to be far more than 33% less comprehensible
than the 3x one. It almost seems to have a
logarithmic decay of comprehensibility.
Also, all tonal or pitch elements seem to be completely removed.
Only the rhythm is maintained from the original, in my opinion.
One additional note. Your example is at jaws level distinction. Can we hear
this with someone talking naturally?
My guess is that it will sound far worse, because the process strips out way
too much. It sounds close enough to the original
because you spoke in a military clipped staccato style.
Very cool stuff!
From: address@hidden [mailto:address@hidden On Behalf Of Bill Cox
Sent: Monday, October 04, 2010 9:30 PM
To: Vinux Development; address@hidden; Jonathan Duddington
Subject: [VINUX-DEVELOPMENT] Test speeding up LPC synthesized voice
I found some very simple LPC encode/decode software on the net which performs
better than what I had so far, so I'm switching over
Here's a sample of my voice, and then the result after encoding and decoding.
I also sped up my voice by 2X, 3X, and 4X. Output
samples attached. Don't worry too much that the encoded voice sounds fairly
poor. The encoder and decoder are only 300 lines of C
code! I've also attached the the C source code, for any hackers out there.
You received this message because you are subscribed to the Google Groups "The
Vinux Development Forum" group.
To post to this group, send email to address@hidden
To unsubscribe from this group, send email to
For more options, visit this group at
Vinux Home Page: http://vinux.org.uk/