So, where'd the volume controls get to?

So, where'd the volume controls get to?

[Kirk Reiser]

Hello Samuel et al: The latest version of speakup appears to have
replaced the volume controls with inflection controls. I don't really
have a problem with this if the volume controls have just been moved
to another key combination but I haven't been able to find where they
are now. I may also be out to lunch but I don't remember any
discussion about moving them or the reason for doing so here on the
list. Can someone clarify the situation for me please?

BTW, a lot of people use the volume controls so I hope they haven't
been arbitrarily removed.

IMO espeakup is still entirely brain dead so a lot of us still have to
use very old versions from back before espeak-en to get decent
performence. It's just an observation.

   Kirk

Re: So, where'd the volume controls get to?

[Chime Hart]

Well Kirk-and-All, I've been asking Samuel for an inflection setting for 
DecTalk in Speakup. Certainly I completely agree, we still need access to 
volume controlls, as probably because of settings in these DecTalk
   DecTalk drivers, volume drops suddenly-and-I need to jossle it 
to restore to where I like it. 
I would love to try these new settings, but I am only on kernel 5.5.0 
although my daily upgrades are in 5.8 but after 174 days I am not ready to 
reboot. Please inform how I can run this newer Speakup? 
Thanks so much in advance-and-I am slightly surprised Samuel didn't make 
an announcement about whatever changes he made, but I would be eternally 
greatful for an inflection setting.
Chime

Re: So, where'd the volume controls get to?

[Samuel Thibault]

Hello,

Kirk Reiser, le ven. 09 oct. 2020 11:19:01 -0400, a ecrit:
> Hello Samuel et al: The latest version of speakup appears to have
> replaced the volume controls with inflection controls.

?? I'm not aware of this. Inflection was added in the /sys parameters,
but only to be configured manual by echoing there. What hardware are you
using?

> IMO espeakup is still entirely brain dead so a lot of us still have to
> use very old versions from back before espeak-en to get decent
> performence. It's just an observation.

"Performance"? What do you mean?

Again, I'm not an actual user of speakup, so people have to explain what
they mean in details otherwise I won't be able to grasp what problem
there actually is.

Samuel

Re: So, where'd the volume controls get to?

[Alexander Epaneshnikov]

hello Samuel. now when pressing speakup + 1/2 your changing enflection 
instead of volume.

10.10.2020 18:13, Samuel Thibault пишет:
> Hello,
>
> Kirk Reiser, le ven. 09 oct. 2020 11:19:01 -0400, a ecrit:
>> Hello Samuel et al: The latest version of speakup appears to have
>> replaced the volume controls with inflection controls.
> ?? I'm not aware of this. Inflection was added in the /sys parameters,
> but only to be configured manual by echoing there. What hardware are you
> using?
>
>> IMO espeakup is still entirely brain dead so a lot of us still have to
>> use very old versions from back before espeak-en to get decent
>> performence. It's just an observation.
> "Performance"? What do you mean?
>
> Again, I'm not an actual user of speakup, so people have to explain what
> they mean in details otherwise I won't be able to grasp what problem
> there actually is.
>
> Samuel
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

-- 
Sincerely, Alexander.

Re: So, where'd the volume controls get to?

[Chime Hart]

So Alexander, is their an easy way in Debian to run a later version, 
considering Speakup doesn't upgrade with an apt-get? Thanks so much in 
advance.
Chime

Re: So, where'd the volume controls get to?

[Kirk Reiser]

Well, I doubt if it matters exactly what hardware I'm running speakup
on but the two machines I've upgraded are both amd64 machines.

The volume controls have always been on speakup- one and two but now
they report inflection instead. Like I said I've not found the volume
controls moved any place else.

The espeakup problems are as I've pointed out many times before. The
voice doesn't flush quickly enough to be able to move up and down with
either the review keys or arrow keys so it generates a jumbling of
voices which is completely unusable when moving quickly. I'm not
exactly sure how to describe it but the problem shows up on all
machines I've tested with espeakup since espeak-ng. It is very easy to
produce just review up and down quickly with espeakup. I don't know
how speech-dispatcher works when using speechd-up and the espeak
voice. Someone with that would have to speak to that.

So, if you didn't move the controls who did? Curious minds and all
that.

   Kirk


On Sat, 10 Oct 2020, Samuel Thibault wrote:

> Hello,
>
> Kirk Reiser, le ven. 09 oct. 2020 11:19:01 -0400, a ecrit:
>> Hello Samuel et al: The latest version of speakup appears to have
>> replaced the volume controls with inflection controls.
>
> ?? I'm not aware of this. Inflection was added in the /sys parameters,
> but only to be configured manual by echoing there. What hardware are you
> using?
>
>> IMO espeakup is still entirely brain dead so a lot of us still have to
>> use very old versions from back before espeak-en to get decent
>> performence. It's just an observation.
>
> "Performance"? What do you mean?
>
> Again, I'm not an actual user of speakup, so people have to explain what
> they mean in details otherwise I won't be able to grasp what problem
> there actually is.
>
> Samuel
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>

Re: So, where'd the volume controls get to?

[Samuel Thibault]

[-- Attachment #1: Type: text/plain, Size: 395 bytes --]

Hello,

Kirk Reiser, le sam. 10 oct. 2020 14:56:38 -0400, a ecrit:
> So, if you didn't move the controls who did? Curious minds and all
> that.

It seems that it's the introduction of the inflection variable that
changed the values of the var_id_t enum, and these happen to be used by
the default speakupmap.h, thus shifting all meanings...

Would you be able to try the attached patch?

Samuel

[-- Attachment #2: patch --]
[-- Type: text/plain, Size: 1752 bytes --]

commit d9ef20b2061908f968d21579efd5b28d668f528a
Author: Samuel Thibault <samuel.thibault@ens-lyon.org>
Date:   Sat Oct 10 21:30:45 2020 +0200

    speakup: Fix keymap
    
    d97a9d7aea04 ("staging/speakup: Add inflection synth parameter") introduced
    a new "inflection" speakup parameter next to "pitch", but the values of the
    var_id_t enum are actually used by the keymap tables so we must not renumber
    them. The effect was that notably the volume control shortcut (speakup-1 or
    2) was then changing the inflection.
    
    This moves the INFLECTION value at the end of the var_id_t enum to fix back
    the enum values. This also adds a warning about it.
    
    Fixes: d97a9d7aea04 ("staging/speakup: Add inflection synth parameter")
    Reported-by: Kirk Reiser <kirk@reisers.ca>
    Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>

diff --git a/drivers/accessibility/speakup/spk_types.h b/drivers/accessibility/speakup/spk_types.h
index 7398f1196e10..37163118485a 100644
--- a/drivers/accessibility/speakup/spk_types.h
+++ b/drivers/accessibility/speakup/spk_types.h
@@ -32,6 +32,8 @@ enum {
 	E_NEW_DEFAULT,
 };
 
+/* XXX: add new members after, speakupmap.h depends on the values of the enum
+ * starting from SPELL_DELAY (see inc_dec_var) */
 enum var_id_t {
 	VERSION = 0, SYNTH, SILENT, SYNTH_DIRECT,
 	KEYMAP, CHARS,
@@ -42,9 +44,9 @@ enum var_id_t {
 	SAY_CONTROL, SAY_WORD_CTL, NO_INTERRUPT, KEY_ECHO,
 	SPELL_DELAY, PUNC_LEVEL, READING_PUNC,
 	ATTRIB_BLEEP, BLEEPS,
-	RATE, PITCH, INFLECTION, VOL, TONE, PUNCT, VOICE, FREQUENCY, LANG,
+	RATE, PITCH, VOL, TONE, PUNCT, VOICE, FREQUENCY, LANG,
 	DIRECT, PAUSE,
-	CAPS_START, CAPS_STOP, CHARTAB,
+	CAPS_START, CAPS_STOP, CHARTAB, INFLECTION,
 	MAXVARS
 };
 

Flushing issue (was: So, where'd the volume controls get to?)

[Samuel Thibault]

Kirk Reiser, le sam. 10 oct. 2020 14:56:38 -0400, a ecrit:
> The voice doesn't flush quickly enough to be able to move up and down
> with either the review keys or arrow keys so it generates a jumbling
> of voices which is completely unusable when moving quickly.

Ok.  Which versions of espeakup and espeak-ng do you use exactly?  I
do remember fixing some things in that regard, and it was then working
fine.  When I'm testing with the versions of Debian 10 for instance,
when I move around extremely quickly, the espeak-ng speech synthesis
doesn't even have the time to speak anything (since I interrupt it
extremely quickly)

Samuel

Re: So, where'd the volume controls get to?

[Chime Hart]

Hi  All, once again: So in looking in that patch, did I understand, once 
you fixed the volume, there are no push-buttons for the new inflection 
variable? Unfortunately insert+7 is an only unused in that roe. I suppose 
you could use insert+i for inflection or insert+v for volume. With an 
inflection option, instead of 2 keystrokes for up-and-down, we could type 
in a value. And lastly, can some1 please inform, how can I install a 
Speackup patch? I need to learn how. Thanks in advance
Chime

Re: So, where'd the volume controls get to?

[Kirk Reiser]

Hi Samuel: I am embarrassed to admit that it's been so many years
since I compiled a kernel and speakup I don't even remember how to do
it. If someone would tell me the commands for building the modules
I'll git pull a kernel and try it. It might be faster for one of the
folks that still do it regularly to try it also. They can probably do
it in no time.

   Kirk

On Sat, 10 Oct 2020, Samuel Thibault wrote:

> Hello,
>
> Kirk Reiser, le sam. 10 oct. 2020 14:56:38 -0400, a ecrit:
>> So, if you didn't move the controls who did? Curious minds and all
>> that.
>
> It seems that it's the introduction of the inflection variable that
> changed the values of the var_id_t enum, and these happen to be used by
> the default speakupmap.h, thus shifting all meanings...
>
> Would you be able to try the attached patch?
>
> Samuel
>

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Kirk Reiser]

Hahahahah, now that's fast! Unfortunately I can't type that fast. In
fact, if I just hold down the previous-lineor next-line keys it still
reads bits of the screen as it's going up or down. I tried to type
fast enough to not get any speech but couldn't do it. It almost speaks
immediately when I press the key or fast enough I can't tell the
difference.

The version of espeakup I use is 0.8 from 2008 and the libespeak
library as reported by ldd is:

libespeak.so.1 => /usr/lib/x86_64-linux-gnu/libespeak.so.1 (0x00007f6298

I'd be happy to use a newer version if I could find one that works
well. I give the new versions a chance with every new version but
nothing yet. If someone else can point me to a version that works I'd
be happy to try that also. I haven't heard complaints from anyone else
though so maybe it's me being silly. It's hard to believe that I
review faster than anyone else so maybe I am silly.




On Sat, 10 Oct 2020, Samuel Thibault wrote:

> Kirk Reiser, le sam. 10 oct. 2020 14:56:38 -0400, a ecrit:
>> The voice doesn't flush quickly enough to be able to move up and down
>> with either the review keys or arrow keys so it generates a jumbling
>> of voices which is completely unusable when moving quickly.
>
> Ok.  Which versions of espeakup and espeak-ng do you use exactly?  I
> do remember fixing some things in that regard, and it was then working
> fine.  When I'm testing with the versions of Debian 10 for instance,
> when I move around extremely quickly, the espeak-ng speech synthesis
> doesn't even have the time to speak anything (since I interrupt it
> extremely quickly)
>
> Samuel
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Samuel Thibault]

Kirk Reiser, le sam. 10 oct. 2020 19:17:42 -0400, a ecrit:
> In fact, if I just hold down the previous-lineor next-line keys it
> still reads bits of the screen as it's going up or down. I tried to
> type fast enough to not get any speech but couldn't do it. It almost
> speaks immediately when I press the key or fast enough I can't tell
> the difference.

Ok but isn't it getting interrupted by the next key press also very
fast?

Making the rate to 9 and running less on a long text, then keeping the
down key pressed, I do get gibberish speech indeed, since it basically
tries to speak the first words of each line very fast before the next
key press switches to the next line. But what else would be expected?

> the libespeak library as reported by ldd is:
> 
> libespeak.so.1 => /usr/lib/x86_64-linux-gnu/libespeak.so.1 (0x00007f6298

That is not precise enough, that has never changed since the espeak
times. The last version of espeak-ng I'm aware of is 1.50.

Samuel

Re: Flushing issue

[Tyler Spivey]

It might be getting interrupted quickly, but at the wrong time.
IIRC, the core issue with espeak-ng (without speech dispatcher) is that 
there's no way to flush the audio without at least one buffer of audio 
playing.
Speech-dispatcher, NVDA, etc don't care because they all use their own 
audio library. But espeakup and emacspeak use espeak-ng's built-in audio 
library and the performance isn't very good.
While I haven't tried with espeakup, emacspeak was completely unusable 
on my Ubuntu 20.04 system with PulseAudio and whatever espeak-ng it had.

I'll try with espeakup and report back, since I assume this is the 
configuration most people would actually want to use. Yes you can get by 
on a mostly text-only system without pulse, but it has a lot of 
disadvantages.

If I interrupt speech, I want it interrupted as soon as I hit the key, 
not after whatever buffer plays because espeak can't tell the audio 
driver to stop. I think the old espeak did that.


On 10/10/2020 4:53 PM, Samuel Thibault wrote:
> Kirk Reiser, le sam. 10 oct. 2020 19:17:42 -0400, a ecrit:
>> In fact, if I just hold down the previous-lineor next-line keys it
>> still reads bits of the screen as it's going up or down. I tried to
>> type fast enough to not get any speech but couldn't do it. It almost
>> speaks immediately when I press the key or fast enough I can't tell
>> the difference.
> 
> Ok but isn't it getting interrupted by the next key press also very
> fast?
> 
> Making the rate to 9 and running less on a long text, then keeping the
> down key pressed, I do get gibberish speech indeed, since it basically
> tries to speak the first words of each line very fast before the next
> key press switches to the next line. But what else would be expected?
> 
>> the libespeak library as reported by ldd is:
>>
>> libespeak.so.1 => /usr/lib/x86_64-linux-gnu/libespeak.so.1 (0x00007f6298
> 
> That is not precise enough, that has never changed since the espeak
> times. The last version of espeak-ng I'm aware of is 1.50.
> 
> Samuel
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
> 

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Kirk Reiser]

On Sun, 11 Oct 2020, Samuel Thibault wrote:

> Ok but isn't it getting interrupted by the next key press also very
> fast?

It does interrupt quickly but there is previously spoken speech that
gets jammed against the new speech.

> Making the rate to 9 and running less on a long text, then keeping the
> down key pressed, I do get gibberish speech indeed, since it basically
> tries to speak the first words of each line very fast before the next
> key press switches to the next line. But what else would be expected?

You do want to hear it speak fast but new speech not old and new
scrunched together. I don't feel I'm doing a good job of describing
it.

>> the libespeak library as reported by ldd is:
>>
>> libespeak.so.1 => /usr/lib/x86_64-linux-gnu/libespeak.so.1 (0x00007f6298
>
> That is not precise enough, that has never changed since the espeak
> times. The last version of espeak-ng I'm aware of is 1.50.

Okay, the espeakup I'm using isn't using espeak-ng at all. It's using:

libespeak1:amd64                       1.48.04+dfsg-9

   Kirk
   _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>

Re: So, where'd the volume controls get to?

[Gregory Nowak]

I'll try it either Sunday, or Monday evening MST. I'm running kernel
5.8.7, and am not seeing this issue. What kernel does this happen on,
so I can be sure I'm testing with a kernel version where this is an
issue?

Greg


On Sat, Oct 10, 2020 at 07:00:52PM -0400, Kirk Reiser wrote:
> Hi Samuel: I am embarrassed to admit that it's been so many years
> since I compiled a kernel and speakup I don't even remember how to do
> it. If someone would tell me the commands for building the modules
> I'll git pull a kernel and try it. It might be faster for one of the
> folks that still do it regularly to try it also. They can probably do
> it in no time.
> 
>   Kirk


-- 
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.

--
Free domains: http://www.eu.org/ or mail dns-manager@EU.org

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Gregory Nowak]

On Sun, Oct 11, 2020 at 01:53:55AM +0200, Samuel Thibault wrote:
> Ok but isn't it getting interrupted by the next key press also very
> fast?
> 
> Making the rate to 9 and running less on a long text, then keeping the
> down key pressed, I do get gibberish speech indeed, since it basically
> tries to speak the first words of each line very fast before the next
> key press switches to the next line. But what else would be expected?

Ok. I've just put up with this for the last two or three years, and
assumed that others either weren't experiencing this, or were just
putting up with it like me. If this can actually be gotten rid of
though, that would make software speech nice to use again. Here's how
to reproduce. I'm testing on a devuan Beowulf system, which is based
on debian buster. I have no pulseaudio installed, though as far as I
can tell, this is still the same with or without pulseaudio
running. This is specific to espeakup for me, it doesn't happen with
speechd-up, or with a hardware synthesizer. The espeakup version is
1:0.80-15, the espeak-ng version is 1.49.2+dfsg-8. 

1. modprobe speakup_soft

2. Start espeakup. I do this with

service espeakup start

3. Set speakup rate to 7 or 8, I have it set to 7.

4. Open the sound card with another application. This can be orca,
   aplay, mplayer, and probably something else too. The other sound
   source doesn't need to actually be playing something. It could be
   simply orca loaded in the GUI and ready to speak, or a file paused
   in mplayer. The sound device just needs to be opened and held ready
   for use by something else.

5. Load a screen full of content, man espeakup should be as good as
   anything.

6. Move to top of screen with speakup+9. Move down the screen with
   numpad 9, the speakup say next line command. Let the line read for
   the first few words, and press numpad 9 again to move to next
   line. What you should hear is that espeak doesn't finish reading
   the current line before the next line starts to speak. This causes
   the two speech chunks to overlap, making it difficult to hear the start
   of the next line of text.

I hope this demonstrates the issue. If there are other questions,
please ask.

Greg


-- 
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.

--
Free domains: http://www.eu.org/ or mail dns-manager@EU.org

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Chime Hart]

You know, Greg, running Voxin with Speakup on a laptop, I've noticed 
something similar, which I mentioned to Oralux. Voxin wasn't always like 
that, but yes, there was s smijin of a sound from a line above playing 
along with the current line. Voxin is essentially IBM TTS.

Chime

Re: So, where'd the volume controls get to?

[Gregory Nowak]

On Sat, Oct 10, 2020 at 08:24:44PM -0700, Gregory Nowak wrote:
> I'm running kernel
> 5.8.7, and am not seeing this issue. What kernel does this happen on,
> so I can be sure I'm testing with a kernel version where this is an
> issue?

Goes to show how much I use speakup's volume controls. Yes, I see the issue on 5.8.7, and will report if Samuel's
patch fixes it soon.

Greg


-- 
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.

--
Free domains: http://www.eu.org/ or mail dns-manager@EU.org

Re: Flushing issue

[Samuel Thibault]

Hello,

Tyler Spivey, le sam. 10 oct. 2020 17:46:58 -0700, a ecrit:
> It might be getting interrupted quickly, but at the wrong time.
> IIRC, the core issue with espeak-ng (without speech dispatcher) is that
> there's no way to flush the audio without at least one buffer of audio
> playing.

Which version of espeak-ng are you using exactly? In Debian we have
a patch to cancel the audio as early as possible, it was discussed
upstream on https://github.com/espeak-ng/espeak-ng/issues/172

The Debian patch is available on 

https://salsa.debian.org/a11y-team/espeak-ng/-/raw/master/debian/patches/fix-cancel

Samuel

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Samuel Thibault]

Kirk Reiser, le sam. 10 oct. 2020 20:58:56 -0400, a ecrit:
> On Sun, 11 Oct 2020, Samuel Thibault wrote:
> > Ok but isn't it getting interrupted by the next key press also very
> > fast?
> 
> It does interrupt quickly but there is previously spoken speech that
> gets jammed against the new speech.

So they get mixed? This really looks like what I fixed in the Debian
patch that upstream still hasn't integrated.

> > > the libespeak library as reported by ldd is:
> > > 
> > > libespeak.so.1 => /usr/lib/x86_64-linux-gnu/libespeak.so.1 (0x00007f6298
> > 
> > That is not precise enough, that has never changed since the espeak
> > times. The last version of espeak-ng I'm aware of is 1.50.
> 
> Okay, the espeakup I'm using isn't using espeak-ng at all. It's using:
> 
> libespeak1:amd64                       1.48.04+dfsg-9

That's really old them. You'd want to use a more recent version, at
least libespeak-ng1 1.49.0+dfsg-7

Samuel

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Samuel Thibault]

Hello,

Gregory Nowak, le sam. 10 oct. 2020 20:55:32 -0700, a ecrit:
> 4. Open the sound card with another application.

Oh, in that case I do get the mixture on my system indeed.

Samuel

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Samuel Thibault]

[-- Attachment #1: Type: text/plain, Size: 410 bytes --]

Samuel Thibault, le dim. 11 oct. 2020 16:35:16 +0200, a ecrit:
> Gregory Nowak, le sam. 10 oct. 2020 20:55:32 -0700, a ecrit:
> > 4. Open the sound card with another application.
> 
> Oh, in that case I do get the mixture on my system indeed.

It seems the buffering is happening inside alsa's dmix then. Could you
try the attached patch on pcaudiolib? (which is the one that configures
the buffering)

Samuel

[-- Attachment #2: patch --]
[-- Type: text/plain, Size: 871 bytes --]

Index: pcaudiolib-1.1/src/alsa.c
===================================================================
--- pcaudiolib-1.1.orig/src/alsa.c
+++ pcaudiolib-1.1/src/alsa.c
@@ -99,6 +99,7 @@ alsa_object_open(struct audio_object *ob
 
 	snd_pcm_hw_params_t *params = NULL;
 	snd_pcm_hw_params_malloc(&params);
+	snd_pcm_uframes_t bufsize = 128;
 
 	int err = 0;
 	if ((err = snd_pcm_open(&self->handle, self->device ? self->device : "default", SND_PCM_STREAM_PLAYBACK, 0)) < 0)
@@ -113,6 +114,8 @@ alsa_object_open(struct audio_object *ob
 		goto error;
 	if ((err = snd_pcm_hw_params_set_channels(self->handle, params, channels)) < 0)
 		goto error;
+	if ((err = snd_pcm_hw_params_set_buffer_size_near(self->handle, params, &bufsize)) < 0)
+		goto error;
 	if ((err = snd_pcm_hw_params(self->handle, params)) < 0)
 		goto error;
 	if ((err = snd_pcm_prepare(self->handle)) < 0)

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Samuel Thibault]

[-- Attachment #1: Type: text/plain, Size: 550 bytes --]

Samuel Thibault, le dim. 11 oct. 2020 17:16:40 +0200, a ecrit:
> Samuel Thibault, le dim. 11 oct. 2020 16:35:16 +0200, a ecrit:
> > Gregory Nowak, le sam. 10 oct. 2020 20:55:32 -0700, a ecrit:
> > > 4. Open the sound card with another application.
> > 
> > Oh, in that case I do get the mixture on my system indeed.
> 
> It seems the buffering is happening inside alsa's dmix then. Could you
> try the attached patch on pcaudiolib? (which is the one that configures
> the buffering)

And this version should be also fixing it for pulseaudio.

Samuel

[-- Attachment #2: patch --]
[-- Type: text/plain, Size: 1701 bytes --]

---
 src/alsa.c       |    3 +++
 src/audio_priv.h |    4 ++++
 src/pulseaudio.c |    7 +++++++
 3 files changed, 14 insertions(+)

--- a/src/alsa.c
+++ b/src/alsa.c
@@ -99,6 +99,7 @@ alsa_object_open(struct audio_object *ob
 
 	snd_pcm_hw_params_t *params = NULL;
 	snd_pcm_hw_params_malloc(&params);
+	snd_pcm_uframes_t bufsize = (rate * channels * LATENCY) / 1000;
 
 	int err = 0;
 	if ((err = snd_pcm_open(&self->handle, self->device ? self->device : "default", SND_PCM_STREAM_PLAYBACK, 0)) < 0)
@@ -113,6 +114,8 @@ alsa_object_open(struct audio_object *ob
 		goto error;
 	if ((err = snd_pcm_hw_params_set_channels(self->handle, params, channels)) < 0)
 		goto error;
+	if ((err = snd_pcm_hw_params_set_buffer_size_near(self->handle, params, &bufsize)) < 0)
+		goto error;
 	if ((err = snd_pcm_hw_params(self->handle, params)) < 0)
 		goto error;
 	if ((err = snd_pcm_prepare(self->handle)) < 0)
--- a/src/pulseaudio.c
+++ b/src/pulseaudio.c
@@ -74,6 +74,13 @@ pulseaudio_object_open(struct audio_obje
 	}
 
 	int error = 0;
+	pa_buffer_attr battr;
+
+	battr.fragsize = (uint32_t) -1;
+	battr.maxlength = (uint32_t) -1;
+	battr.minreq = (uint32_t) -1;
+	battr.prebuf = (uint32_t) -1;
+	battr.tlength = pa_bytes_per_second(&self->ss) * LATENCY / 1000;
 	self->s = pa_simple_new(NULL,
 	                        self->application_name,
 	                        PA_STREAM_PLAYBACK,
--- a/src/audio_priv.h
+++ b/src/audio_priv.h
@@ -52,6 +52,10 @@ struct audio_object
 	                         int error);
 };
 
+/* We try to aim for 10ms cancelation latency, which will be perceived as
+ * "snappy" by users */
+#define LATENCY 10
+
 #if defined(_WIN32) || defined(_WIN64)
 
 #include <windows.h>

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Gregory Nowak]

On Sun, Oct 11, 2020 at 05:16:40PM +0200, Samuel Thibault wrote:
> Could you
> try the attached patch on pcaudiolib? (which is the one that configures
> the buffering)

Yes, I'll grab the source for libpcaudio0, will patch, compile, and
test. I'll report back when I do assuming someone else doesn't beat me
to it first. That won't be until I test the kernel patch first from
the previous discussion. Stay tuned to both threads.

Greg


-- 
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.

--
Free domains: http://www.eu.org/ or mail dns-manager@EU.org

Re: So, where'd the volume controls get to?

[Gregory Nowak]

On Sat, Oct 10, 2020 at 09:36:36PM +0200, Samuel Thibault wrote:
> Would you be able to try the attached patch?

I can confirm that Samuel's patch fixes the issue so that speakup+1
and speakup+2 controls volume again.

Since I feel that volume control on the fly is fairly critical, I've
decided to release my kernel build. The debian packages were built for
a devuan Beowulf system, and should work on debian buster. This kernel
should boot on most if not all x86_64 systems. There are modules for a
variety of SATA controllers, and file systems. The debian packages
aren't signed, and the kernel isn't signed either which your UEFI
and/or boot loader might not like. If any of this makes you
uncomfortable, wait for the fix to make it to where ever you get your
kernel updates from, or build your own kernel with the patch (the www
is your friend).

For those who don't want to or can't install debian packages, the
amd64_binaries directory has binaries for the kernel, modules, and
headers. If installing this kernel breaks your system in any way, you
get to keep all the pieces.

You can find all of this at:

<https://www.gregn.net/linux-5.8.14/>

Greg



-- 
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.

--
Free domains: http://www.eu.org/ or mail dns-manager@EU.org

Re: So, where'd the volume controls get to?

[Janina Sajka]

+1 to Kirk. I'm also wondering about volume control and I'm also still
stuck on the old espeak.

Of the two the volume control is actually more important for me, because
I have Orca using the same audio device and need to level the volume
with Speakup. This was always so much easier to do on the Speakup side.

Best,

Janina

Kirk Reiser writes:
> Hello Samuel et al: The latest version of speakup appears to have
> replaced the volume controls with inflection controls. I don't really
> have a problem with this if the volume controls have just been moved
> to another key combination but I haven't been able to find where they
> are now. I may also be out to lunch but I don't remember any
> discussion about moving them or the reason for doing so here on the
> list. Can someone clarify the situation for me please?
> 
> BTW, a lot of people use the volume controls so I hope they haven't
> been arbitrarily removed.
> 
> IMO espeakup is still entirely brain dead so a lot of us still have to
> use very old versions from back before espeak-en to get decent
> performence. It's just an observation.
> 
>   Kirk
> 
> 
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

-- 

Janina Sajka
https://linkedin.com/in/jsajka

Linux Foundation Fellow
Executive Chair, Accessibility Workgroup:	http://a11y.org

The World Wide Web Consortium (W3C), Web Accessibility Initiative (WAI)
Co-Chair, Accessible Platform Architectures	http://www.w3.org/wai/apa

Re: So, where'd the volume controls get to?

[Kirk Reiser]

Thanks Greg for taking on testing the patches and releasing your
results. Also thanks again to Samuel for getting a fix out so quickly.

   Kirk

On Sun, 11 Oct 2020, Gregory Nowak wrote:

> On Sat, Oct 10, 2020 at 09:36:36PM +0200, Samuel Thibault wrote:
>> Would you be able to try the attached patch?
>
> I can confirm that Samuel's patch fixes the issue so that speakup+1
> and speakup+2 controls volume again.
>
> Since I feel that volume control on the fly is fairly critical, I've
> decided to release my kernel build. The debian packages were built for
> a devuan Beowulf system, and should work on debian buster. This kernel
> should boot on most if not all x86_64 systems. There are modules for a
> variety of SATA controllers, and file systems. The debian packages
> aren't signed, and the kernel isn't signed either which your UEFI
> and/or boot loader might not like. If any of this makes you
> uncomfortable, wait for the fix to make it to where ever you get your
> kernel updates from, or build your own kernel with the patch (the www
> is your friend).
>
> For those who don't want to or can't install debian packages, the
> amd64_binaries directory has binaries for the kernel, modules, and
> headers. If installing this kernel breaks your system in any way, you
> get to keep all the pieces.
>
> You can find all of this at:
>
> <https://www.gregn.net/linux-5.8.14/>
>
> Greg
>
>
>
>

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Kirk Reiser]

On Sun, 11 Oct 2020, Samuel Thibault wrote:

> Kirk Reiser, le sam. 10 oct. 2020 20:58:56 -0400, a ecrit:
>> Okay, the espeakup I'm using isn't using espeak-ng at all. It's using:
>>
>> libespeak1:amd64                       1.48.04+dfsg-9
>
> That's really old them. You'd want to use a more recent version, at
> least libespeak-ng1 1.49.0+dfsg-7

I won't upgrade from my old crusty but faithful version of espeakup
and espeak until what's available now is as usable. I do appreciate
you working on fixing it however and I hope this new set of patches
you've put together does in fact fix the problem.

Re: So, where'd the volume controls get to?

[Karen Lewellen]

If volume is restored, where is inflection now?
granted I am not a speakup user, but one should have both, no?
Karen



On Mon, 12 Oct 2020, Kirk Reiser wrote:

> Thanks Greg for taking on testing the patches and releasing your
> results. Also thanks again to Samuel for getting a fix out so quickly.
>
>  Kirk
>
> On Sun, 11 Oct 2020, Gregory Nowak wrote:
>
>>  On Sat, Oct 10, 2020 at 09:36:36PM +0200, Samuel Thibault wrote:
>> >  Would you be able to try the attached patch?
>>
>>  I can confirm that Samuel's patch fixes the issue so that speakup+1
>>  and speakup+2 controls volume again.
>>
>>  Since I feel that volume control on the fly is fairly critical, I've
>>  decided to release my kernel build. The debian packages were built for
>>  a devuan Beowulf system, and should work on debian buster. This kernel
>>  should boot on most if not all x86_64 systems. There are modules for a
>>  variety of SATA controllers, and file systems. The debian packages
>>  aren't signed, and the kernel isn't signed either which your UEFI
>>  and/or boot loader might not like. If any of this makes you
>>  uncomfortable, wait for the fix to make it to where ever you get your
>>  kernel updates from, or build your own kernel with the patch (the www
>>  is your friend).
>>
>>  For those who don't want to or can't install debian packages, the
>>  amd64_binaries directory has binaries for the kernel, modules, and
>>  headers. If installing this kernel breaks your system in any way, you
>>  get to keep all the pieces.
>>
>>  You can find all of this at:
>>
>>  <https://www.gregn.net/linux-5.8.14/>
>>
>>  Greg
>> 
>> 
>> 
>> 
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>
>

Re: So, where'd the volume controls get to?

[Samuel Thibault]

Karen Lewellen, le lun. 12 oct. 2020 13:39:22 -0400, a ecrit:
> If volume is restored, where is inflection now?

There is no default keybinding for it, but it could be added. I'm
however not sure whether it would be so useful as a keybinding? It can
be configured through the file in
/sys/accessibility/speakup/soft/inflection

Samuel

Re: So, where'd the volume controls get to?

[Karen Lewellen]

because speaking personally, there are times in different settings, when 
one   desires the flexibility of adjusting how much or how little 
inflection is provided..without having to be a computer programmer and edit 
that way.
Its about choices allowing for one to  make use of the screen reading tool 
individually.
But again, I am not an actual user.  those choosing the tool might not 
care.  Still, speaking personally, the lack of flexibility without 
complexity would deter my considering the program, which is a shame since 
it may be the only Linux option for hardware speech.
Karen



On Mon, 12 Oct 2020, Samuel Thibault wrote:

> Karen Lewellen, le lun. 12 oct. 2020 13:39:22 -0400, a ecrit:
>> If volume is restored, where is inflection now?
>
> There is no default keybinding for it, but it could be added. I'm
> however not sure whether it would be so useful as a keybinding? It can
> be configured through the file in
> /sys/accessibility/speakup/soft/inflection
>
> Samuel
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>
>

Re: So, where'd the volume controls get to?

[Samuel Thibault]

Karen Lewellen, le lun. 12 oct. 2020 14:19:40 -0400, a ecrit:
> one desires the flexibility of adjusting how much or how little
> inflection is provided..without having to be a computer programmer and
> edit that way.

Sure, I just mean that there aren't so many shortcuts possibilities out
there. It's up to the speakup community to decide whether assigning a
default shortcut is desired.

Samuel

Re: So, where'd the volume controls get to?

[Karen Lewellen]

I can respect that stance. However there are as many ways to  decide what 
is  desired in one's computer tools as there are individuals using them.
better to build an existing buffet of options than have to add things in 
one  at a time later.



On Mon, 12 Oct 2020, Samuel Thibault wrote:

> Karen Lewellen, le lun. 12 oct. 2020 14:19:40 -0400, a ecrit:
>> one desires the flexibility of adjusting how much or how little
>> inflection is provided..without having to be a computer programmer and
>> edit that way.
>
> Sure, I just mean that there aren't so many shortcuts possibilities out
> there. It's up to the speakup community to decide whether assigning a
> default shortcut is desired.
>
> Samuel
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>
>

Re: So, where'd the volume controls get to?

[Tyler Spivey]

Where is the genmap utility discussed in the guide (spkguide.txt in my 
kernel source)?
That way we could just add our own key bindings. Is the guide outdated?

On 10/12/2020 11:31 AM, Karen Lewellen wrote:
> I can respect that stance. However there are as many ways to  decide 
> what is  desired in one's computer tools as there are individuals using 
> them.
> better to build an existing buffet of options than have to add things in 
> one  at a time later.
> 
> 
> 
> On Mon, 12 Oct 2020, Samuel Thibault wrote:
> 
>> Karen Lewellen, le lun. 12 oct. 2020 14:19:40 -0400, a ecrit:
>>> one desires the flexibility of adjusting how much or how little
>>> inflection is provided..without having to be a computer programmer and
>>> edit that way.
>>
>> Sure, I just mean that there aren't so many shortcuts possibilities out
>> there. It's up to the speakup community to decide whether assigning a
>> default shortcut is desired.
>>
>> Samuel
>> _______________________________________________
>> Speakup mailing list
>> Speakup@linux-speakup.org
>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>
>>
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Re: So, where'd the volume controls get to?

[Didier Spaier]

Hi,

spkguide.txt is indeed outdated.
As an aside in the 5.9 kernel tree it has migrated here:
Documentation/admin-guide/spkguide.txt
And the file  sysfs-driver-speakup is now there:
Documentation/ABI/stable/sysfs-driver-speakup

Maybe one can find genmap in an old kernel tree?

Didier

Le 12/10/2020 à 20:42, Tyler Spivey a écrit :
> Where is the genmap utility discussed in the guide (spkguide.txt in my kernel source)?
> That way we could just add our own key bindings. Is the guide outdated?
> 
> On 10/12/2020 11:31 AM, Karen Lewellen wrote:
>> I can respect that stance. However there are as many ways to  decide what is  desired in one's computer tools as there are individuals using them.
>> better to build an existing buffet of options than have to add things in one  at a time later.
>>
>>
>>
>> On Mon, 12 Oct 2020, Samuel Thibault wrote:
>>
>>> Karen Lewellen, le lun. 12 oct. 2020 14:19:40 -0400, a ecrit:
>>>> one desires the flexibility of adjusting how much or how little
>>>> inflection is provided..without having to be a computer programmer and
>>>> edit that way.
>>>
>>> Sure, I just mean that there aren't so many shortcuts possibilities out
>>> there. It's up to the speakup community to decide whether assigning a
>>> default shortcut is desired.
>>>
>>> Samuel
>>> _______________________________________________
>>> Speakup mailing list
>>> Speakup@linux-speakup.org
>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>
>>>
>> _______________________________________________
>> Speakup mailing list
>> Speakup@linux-speakup.org
>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Re: Flushing issue (was: So, where'd the volume controls get to?)

[Gregory Nowak]

On Sun, Oct 11, 2020 at 05:16:40PM +0200, Samuel Thibault wrote:
> It seems the buffering is happening inside alsa's dmix then. Could you
> try the attached patch on pcaudiolib? (which is the one that configures
> the buffering)

I just got around to trying this patch, and it seems to be working for
me under ALSA. If necessary, I can install debian configured for
pulseaudio in a virtual machine, and try it under pulseaudio as
well. I don't want to reconfigure my sound setup on my production
machines for pulseaudio. Thanks Samuel for your fixes.

Greg


-- 
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.

--
Free domains: http://www.eu.org/ or mail dns-manager@EU.org