Oops... It was pointed out that I got the credits wrong in the header.
 I've regenerated the second patch to correct it.

Bill

On Fri, Jan 8, 2010 at 1:08 PM, Bill Cox <[email protected]> wrote:
> Hi, Luke.  Adding the ability to set PulseAudio parameters in
> speechd.conf was harder than I thought it would be, but it's done.
> Attached are two patches.  The first is the patch file to get the
> basic pulseaudio driver that you've tested.  The second is the work I
> had to do to get the PulseAudio buffering parameters exposed to the
> user through speechd.conf.  Just add these to your debian/patches
> directory and the 00list file.
>
> I wasn't sure where to e-mail these... sorry for the spam!
>
> Bill
>
> On Thu, Jan 7, 2010 at 7:56 AM, Bill Cox <[email protected]> wrote:
>> Hi, Luke.  That's great!  For some reason, I keep expecting the Magic
>> Code Fairies to clean up code before it gets included anywhere.
>> Earlier in this thread, Rui Batista said he was going to do some
>> cleanup - making PulseAudio buffer parameters configurable again.  So,
>> naturally I didn't bother.  If you can wait a few hours, I'll do that
>> change, and post an improved patch here.
>>
>> Bill
>>
>> On Thu, Jan 7, 2010 at 12:31 AM, Luke Yelavich
>> <[email protected]> wrote:
>>> On Thu, Jan 07, 2010 at 03:25:47PM EST, Luke Yelavich wrote:
>>>> On Thu, Jan 07, 2010 at 03:19:32PM EST, Bill Cox wrote:
>>>> > Hi, Luke.  The new pulse drivers work well for me on both Karmic and
>>>> > Lucid, and I haven't heard of users with poor performance so far, but
>>>> > it could be machine specific.  The strange thing is you're reporting
>>>> > libao works well, and all I did was cut and paste the pulseaudio calls
>>>> > libao makes into the libao driver, so I would expect similar
>>>> > performance, other than for the buffering settings I make in the
>>>> > pulse-simple interface.
>>>>
>>>> *slaps head. I didn't apply the patch I created with the new pulse file. 
>>>> I'll test again and get back to you,
>>>
>>> Ok, now that I actually have it aplying and built, the difference is *VERY* 
>>> noticable. Currently using espeak with pulse, but will try with portaudio 
>>> again since bits of short text, especially if arrowing very fast, are a 
>>> little bit clicky.
>>>
>>> Long term, there is the power consumption issue that Lennart raised, since 
>>> we are using the simple API, but we need to work that out as best we can, 
>>> and go from there, but I think this will be going into Ubuntu very shortly, 
>>> once I have done a few more tests.
>>>
>>> Great work!
>>>
>>> Luke
>>>
>>
>

Attachment: patches.tar.gz
Description: GNU Zip compressed data

-- 
Ubuntu-accessibility mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-accessibility

Reply via email to