Hello guys
Thanks for the thanks.
I have had a good few days working on changing the OMX audio code
library and the fork of espeakup known as piespeakup from static
Makefiles into Makefile and distribution .tar.gz files with autoconf and
automake.
For a long time I have wanted to work on a sizeable autotools project
and this has been a very enjoyable learning experience.
The use case for autotools is usually to provide a way of building
software on subtly different flavours of Unix/Linux. So, why use
autotools when this OMX audio code can only run on the Pi, with it's
particular GPU?
Well, firstly because it's a learning opportunity, but most importantly
because autotools gives a more robust way of checking stuff as it is
built, for example that the userland code libraries are there.
Next on the list now that I have the audio code running on Arch as well
as Raspbian are these things, not necessarily in the order in which I
will do them:
1. Put Talking Arch .img files on the site for both armv6 (pre Pi2) and
armv7 (Pi2 and later).
2. Put ilctts-1.0.0.tar.gz and piespeakup-1.0.0.tar.gz on the Web server
for download.
3. Write an Emacspeak speech server which uses this OMX code so that
Rill, and others of course, can use Emacspeak again stutter-free.
4. Update the Web site tutorials with the new build process for the OMX
audio code.
5. Probably a long way down the road yet, build a speech-dispatcher
audio module to use the OMX code so we can get Orca to work properly
with no stuttering and with NO USB sound dongle.
I realise now that the name 'ilctts' which I have given the audio
library is a bit misleading, the name standing for 'integration layer
client text-to-speech'. There is no OMX in the name and it doesn't say
it will only work with espeak, which is true, and it also doesn't say it
COULD be used to render ANY PCM, wherever it comes from. For example I
have a script which converts a wav file into raw PCM and one of the
library test programs will read the PCM and render it with the library.
But changing the name will lead to a whole raft of bug fixes unless I
can do a global search and replace. So I will leave it for now.
I could not believe when I looked at the code again that I first
completed it in November 2014, and that this list is now four years old.
As a last aside, I am going to look for a switch that could be operated
by either blowing down a tube or with the tongue, since using this
library on a Pi a paralysed person could 'speak' to people.
Mike
On 28/12/2016 19:06, Mobeen Iqbal wrote:
I couldn't have put it better my self, many thanks for everything mike!
On 28/12/2016 18:14, Tim Chase wrote:
In an effort to be more thankful in the year ahead, I wanted to just
put a little note on the list to say thanks, Mike, for all your work
on cajoling the Pi into successfully speaking as well as shepherding
this mailing list and the website.
So thanks!
-tim
===========================================================
The raspberry-vi mailing list
Archives: //www.freelists.org/archives/raspberry-vi
Administrative contact: <mike.ray@xxxxxxxxxxxxxx>
-----------------------------------------------------------
Raspberry Pi and the Raspberry Pi logo are trademarks of the Raspberry Pi
Foundation.
This list is not affiliated to the Raspberry Pi Foundation and the views and
attitudes expressed by the subscribers to this list do not reflect those of
the Foundation.
Mike Ray, list creator, January 2013
===========================================================
The raspberry-vi mailing list
Archives: //www.freelists.org/archives/raspberry-vi
Administrative contact: <mike.ray@xxxxxxxxxxxxxx>
-----------------------------------------------------------
Raspberry Pi and the Raspberry Pi logo are trademarks of the Raspberry Pi
Foundation.
This list is not affiliated to the Raspberry Pi Foundation and the views and
attitudes expressed by the subscribers to this list do not reflect those of
the Foundation.
Mike Ray, list creator, January 2013
