DECTalk question, was speakup using different synths with software speech?

[al Sten-Clanton]

Tony:

I recently tried using a DEBTalk.  I tried installing Fedora with it.  It
started out fine, but very quickly during the installation there was a long
lag between keystroke and speech.  Indeed, I had a lot more trouble than
usual figuring what choice I was on at various points of the installation.
I went through with the install the last time; I found that the boot-up was
fine, at least to a point, but then the lag kicked in bit-time.

I think this is a problem somebody mentioned on this list in December.  (He
was using Slackware.) My very crude understanding is that it resulted from
changes in the kernel's interaction with serial ports.  Has Debian worked
around or avoided this problem?  I ask because my work with GRML last
weekend suggests that it may have.  (I had another problem with GRML, but
that's off the point.)

The recent speakup fixes seemm almost to have eradicated the problems I was
having with the TripleTalk LT.  Still, I like the speed of the DECTalk
Express, and also one of its voices.

Thanks for any info.

Al 

-----Original Message-----
From: speakup-bounces at braille.uwo.ca [mailto:speakup-bounces at braille.uwo.ca]
On Behalf Of Tony Baechler
Sent: Friday, June 05, 2009 7:23 AM
To: Speakup is a screen review system for Linux.
Subject: Re: speakup using different synths with software speech?

All,

There still seems to be misunderstandings on what I'm trying to say.  
Again, I really don't mind paying a reasonable amount for software speech,
assuming I can afford it.  The problem is that I've never yet heard software
speech that I liked.  The other problem is that I don't like using non-free
software but I don't think I could get used to ESpeak.  My favorite voice is
the hardware DEC-talk Express.  I can use it to read books, email, work on
my other Linux boxes, etc.  I can listen to it for hours without growing
tired of it.  It has a very fast speech rate while still being
understandable.  It doesn't have a muffle like most software synths,
specifically the software DEC-talk.  It can easily be customized to have the
exact pitch and inflection I want.  
When I bought it, it was around $1200 US and is still worth it, even though
the price dropped since then.  I understand that the USB version isn't as
flexible.  I've tried many different sets of speakrs for software speech,
but all either have too much bass, a muffle or static.  
Probably the best for reading was Realspeak but it was very, very slow.  
I've even tried software speech on a high-end stereo system.  That did help,
but it still wasn't as good as my old DEC-talk Express.  I guess my next
favorite would be the Trippletalk, but it mispronounces things and has other
problems, like stuttering and a fairly small text buffer.  
One really great thing about the DEC Express is that if I'm reading and the
power goes out, it stores at least two screens of text in its buffer, so it
will keep reading for a few minutes.

As I said, I'll look at Voxin.  Perhaps it's not as bad as I think.  I know
speakers do have a lot to do with it and it's partially what I'm used to,
but what I really want is to just use my hardware synthesizers.  Software
speech is enough of an issue for me that I have put off really exploring
Orca.

Also, one thing not mentioned here is my other reason for not leaving
Windows.  That is a lack of GUI audio software.  I'm surprised that no one
else has pointed this out, but a really big problem with software speech and
audio production is that you don't want speech in your recordings.  Even if
you have multiple sound cards or a good multichannel card, a good microphone
will still pick up software speech, I know from experience.  Even if you use
headsets, it's still very inconvenient when you're doing restoration and
editing to hear speech in your ears at the same time as the sound you're
concentrating on.  I work with old time radio shows which have many pops and
clicks.  It's hard to hear the disc noise with speech chattering away in my
ears.  On the other hand, with hardware speech I can't hear it at all with a
headset on, but at least I don't have to repeat the same half second of
sound multiple times because the speech wouldn't shut up.  That's the
problem I'm having now on a Windows machine with only software speech.  I'm
sure there's an easy solution, but the volume of all software synths seems
to dominate all other sound, even if the volume is lowered.

Michael Whapples wrote:
> As for voxin, if you weren't pleased by eloquence, you are unlikely to 
> be pleased with voxin as it uses IBM viavoice which sounds the same as 
> eloquence (or at least very close, I think the voices are very 
> slightly different).
>
> As for contributing to espeak, you could suggest how it could sound 
> better, it may not require altering the actual code it may be a matter 
> of altering the voice files. I don't mean it in a bad way, but you 
> seem to be hard to please with speech output, it may help if we knew 
> what makes a voice good to you.

_______________________________________________
Speakup mailing list
Speakup at braille.uwo.ca
http://speech.braille.uwo.ca/mailman/listinfo/speakup