* Audio PDA with Speech Synth Pkans
@ Doug Sutherland
` Jonathan Duddington
0 siblings, 1 reply; 8+ messages in thread
From: Doug Sutherland @ UTC (permalink / raw)
To: speakup
Folks,
I am working towards producing a small pocket
sized linux system based on ARM920T processor
with onboard hardware speech synth. It will have
both USB host and device and will be able to
operate like a PDA when not connected to PC,
will support both normal keyboard plus some
simpler button/switch type interfaces, to allow
audio PDA functions, read your notes, contacts,
schedule, directions, etc.
It will also operate as a speech synth when
plugged into PC, and I have an ambitious goal
of making it driverless. By correctly implementing
the USB CDC (communications data class) serial
interface, no drivers will be needed to access
what will appear as a virtual serial port on both
windows and linux. Further, I believe there is a
way software-wise to make a USB storage class
entity appear to the system as a CDROM and
automatically do the AUTORUN.INF process,
which could then launch a screen reader and
other specialized audio based apps. This would
allow plugging into any PC, windows or linux,
and having a working speech synth, without
installing anything.
It will be some time before some of the software
is worked out, but I am deep into the hardware
now and should have prototypes soon. I am very
interested in feedback, especially things like what
kind of features would you like to have in both a
speech synth and a personal audio based PDA
type organizer. How can the human interface be
more ergonomic if we ditch the keyboard
paradigm for some controls, perhaps like a
small remote control for controlling the speech
playback and volume etc while roaming or even
when connected to PC in some instances, also
supporting full keyboard when necessary.
My strategic goal is to produce these in low
volume and sell them at very low price, much
lower than anything similar. The goal is not to
make money but to slowly build infrastructure
and skills to continue working on other similar
products, mostly geared towards portable
mobile systems and with special emphasis on
assistive interfaces of many kinds, not just
speech but others also.
I already have facility for CAD design of the
hardware, production soldering, assembly
and testing, and I'm starting on software now.
When I get further along I may ask if anyone
wants to participate in some beta testing by
trying out the device and giving feedback.
Also as I said I am very interested in feedback
especially out of the box thinking: you do not
need a PC to do these functions, old rules
go away and new possibilities emerge like
bluetooth, wifi, gsm/gprs, a specialized
small keypad/remote to control the speech.
There could be a mode where you plug into
PC USB port and walk away with only a
bluetooth headset and small keyboard,
but still can read email and web within a
certain amount of fixed restriction, or move
to full keyboard for serious work.
If this all sounds interesting to you and you
have some ideas or think you'd like to try
out early prototypes, please contact me
offline. I will also surely be trying out the
speakup, compiled into ARM kernel, with
playback of that kernel console plus ability
to act as a speech synth for pc running
speakup. That work should be in process
very soon. But I would love to hear any
suggestions or ideas, especially in the area
of usability, simplicity, ergonomics, how
can a speech synth be more usable and
more portable, how can it become much
more assistive, if you have a processor
running linux in your pocket?
Please contact me off list doug@proficio.ca
if you'd like to discuss this with me. I will
likely be giving away a certain number of
free prototypes eventually, but it may be
some time before that happens.
-- Doug
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: Audio PDA with Speech Synth Pkans
Audio PDA with Speech Synth Pkans Doug Sutherland
@ ` Jonathan Duddington
` Doug Sutherland
0 siblings, 1 reply; 8+ messages in thread
From: Jonathan Duddington @ UTC (permalink / raw)
To: speakup
In article <45C99217.8070209@proficio.ca>,
Doug Sutherland <doug@proficio.ca> wrote:
> I am working towards producing a small pocket
> sized linux system based on ARM920T processor
> with onboard hardware speech synth.
This sounds interesting.
What is the onboard hardware speech synth?
Can you give a link to information about it?
I wonder whether a small software synth such
as eSpeak might be better, or at least useful
as an alternative.
Does the hardware synth add to the cost of the
product?
Does it perform both the spelling-to-phoneme
translation as well as the phoneme-to-sound
production, or is the former done in software
on the main processor?
Can it speak different languages?
How do you compile software for the PDA? Do
you use a cross-compiler on another system?
Are you restricted to C, or is C++ available?
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: Audio PDA with Speech Synth Pkans
` Jonathan Duddington
@ ` Doug Sutherland
` Jonathan Duddington
` Chris Norman
0 siblings, 2 replies; 8+ messages in thread
From: Doug Sutherland @ UTC (permalink / raw)
To: Speakup is a screen review system for Linux.
The hardware is doubletalk rc8660
There is info on the rcsys site
http://www.rcsys.com/
It could be done in software, but I have worked
with this hardware and am pleased with the sound.
Also, I want to keep the processor free of using
many cycles for speech synth, this is not a PC and
it will operate at 180Mhz top speed. It will add to
the cost but I think it is worth the trade-off.
Software is linux cross development toolchain,
the entire toolchain is already done and avail
on processor manufacturers web site, along
with buildroot for flash file system, u-boot
bootloader etc. C, C++, assembly, all of the
traditional tools will be avail.
Processor for first version is AT91RM9200.
Later version will move towards EP9315 to
take advantage of PCMCIA (will map that to
CF format for wifi modem) and IDE (will be
set up to support toshiba 1.8" hard drives,
same as in iPod.
-- Doug
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: Audio PDA with Speech Synth Pkans
` Doug Sutherland
@ ` Jonathan Duddington
` Doug Sutherland
` Chris Norman
1 sibling, 1 reply; 8+ messages in thread
From: Jonathan Duddington @ UTC (permalink / raw)
To: Speakup is a screen review system for Linux.
In article <45C9A862.2050305@proficio.ca>,
Doug Sutherland <doug@proficio.ca> wrote:
> I want to keep the processor free of using
> many cycles for speech synth, this is not a PC and
> it will operate at 180Mhz top speed.
That may not be a problem.
I run eSpeak here on a 200MHz ARM processor and
I measured that it took 51 sec to process some
text to generate 6m 3s of speech output. That's
about 14% of the processor.
Anyway, it's something consider.
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: Audio PDA with Speech Synth Pkans
` Jonathan Duddington
@ ` Doug Sutherland
` Jonathan Duddington
0 siblings, 1 reply; 8+ messages in thread
From: Doug Sutherland @ UTC (permalink / raw)
To: Speakup is a screen review system for Linux.
That is very interesting about espeak on ARM. The only software
speech synth that was bearable to me was IBM TTS which was
really eloquence branded IBM. Is it possible to generate a wav
or mp3 of what it sounds like? Could you possibly send me an
audio sample? I am deeply into hardware at the moment but I'm
interested in knowing how that software synth sounds.
-- Doug
Jonathan Duddington wrote:
I run eSpeak here on a 200MHz ARM processor and
I measured that it took 51 sec to process some text to
generate 6m 3s of speech output.
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: Audio PDA with Speech Synth Pkans
` Doug Sutherland
@ ` Jonathan Duddington
` Doug Sutherland
0 siblings, 1 reply; 8+ messages in thread
From: Jonathan Duddington @ UTC (permalink / raw)
To: Speakup is a screen review system for Linux.
In article <45CA06F9.6090104@proficio.ca>,
Doug Sutherland <doug@proficio.ca> wrote:
> Is it possible to generate a wav or mp3 of what it sounds like? Could
> you possibly send me an audio sample?
Yes, it can generate WAV file output.
It doesn't sound as natural as a large synthesizer, but that doesn't
seem to be the most important attribute for regular users.
There's a large audio file sample on the web site:
http://espeak.sourceforge.net/
and you can download Windows and Linux versions of the software.
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: Audio PDA with Speech Synth Pkans
` Jonathan Duddington
@ ` Doug Sutherland
0 siblings, 0 replies; 8+ messages in thread
From: Doug Sutherland @ UTC (permalink / raw)
To: Speakup is a screen review system for Linux.
Thanks, here are samples of the rcsys rc8660,
description of each demo first followed by the
wav file. I have found that with tweaks using
commands I get this sounding very good.
"Print and forget" text-to-speech conversion.
"In the beginning God created the heaven and
the earth. And the earth was without form, and void;
and darkness was upon the face of the deep.
And the Spirit of God moved upon the face of the
waters."
http://www.rcsys.com/Downloads/demo1.wav
Using an exception dictionary, you can trigger the
playback of a recorded message or sound, generate
tones, or synthesize any message based on the rules
you specify.
"What the **** do you think you're doing?" (profanity filter)
"$GPGGA,123456,3095.2007,N,09527.0923,W" (GPS receiver output)
http://www.rcsys.com/Downloads/demo2.wav
Common abbreviations and acronyms handled
automatically.
Mr. & Mrs. Forrest Gump
1122 St. John St.
Ft. Knox, KY 40121
"The recipe calls for 12 oz of tomato paste, 3 g
of salt, and 2 lbs of sugar."
"The co. had net income of $247M and earnings
per share of $1.07 in fiscal year 2002. Revenues
rose 9.1% to $1.2B."
http://www.rcsys.com/Downloads/demo3.wav
Demonstration of the sinusoidal tone generators
and the wide dynamic range of the TTS synthesizer.
http://www.rcsys.com/Downloads/demo4.wav
Demonstration of the touch-tone dialer, sinusoidal
tone generators, and recording memory. The example
shows how words and phrases that have been recorded
on-chip can be played back to form complete sentences.
http://www.rcsys.com/Downloads/demo5.wav
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: Audio PDA with Speech Synth Pkans
` Doug Sutherland
` Jonathan Duddington
@ ` Chris Norman
1 sibling, 0 replies; 8+ messages in thread
From: Chris Norman @ UTC (permalink / raw)
To: Speakup is a screen review system for Linux.
Do you have any idea of how much this will cost when it's built?
Cheers,
On Wed, 2007-02-07 at 05:22 -0500, Doug Sutherland wrote:
> The hardware is doubletalk rc8660
> There is info on the rcsys site
> http://www.rcsys.com/
>
> It could be done in software, but I have worked
> with this hardware and am pleased with the sound.
> Also, I want to keep the processor free of using
> many cycles for speech synth, this is not a PC and
> it will operate at 180Mhz top speed. It will add to
> the cost but I think it is worth the trade-off.
>
> Software is linux cross development toolchain,
> the entire toolchain is already done and avail
> on processor manufacturers web site, along
> with buildroot for flash file system, u-boot
> bootloader etc. C, C++, assembly, all of the
> traditional tools will be avail.
>
> Processor for first version is AT91RM9200.
> Later version will move towards EP9315 to
> take advantage of PCMCIA (will map that to
> CF format for wifi modem) and IDE (will be
> set up to support toshiba 1.8" hard drives,
> same as in iPod.
>
> -- Doug
>
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
--
Chris Norman.
<!-- cnorman@rnibncw.ac.uk -->
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~ UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
Audio PDA with Speech Synth Pkans Doug Sutherland
` Jonathan Duddington
` Doug Sutherland
` Jonathan Duddington
` Doug Sutherland
` Jonathan Duddington
` Doug Sutherland
` Chris Norman
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).