Closed pitermach closed 1 year ago
Yes, unfortunately it is a bit difficult to hear that amount of pauses between utterances.
@djphoenix any thoughts on this issue? I would personally really appreciate it if this got implemented. Reading with such long pauses is quite uncomfortable. With other VoiceOver voices such as Eloquence, it seems like these pauses scale depending on the speech rate and they get reduced if you are using a faster speed, however ESpeak doesn't seem to do that, so it would be quite nice if this in app setting could be added to reduce them.
Thanks from my side for all your work as well.
Hi there, of course yes, there are a plans to bring new version and this feature is one of key changes. Unfortunately I have no estimates right now. I think, it may be in next two-three weeks to get it on TestFlight.
Hi guys, new TestFlight for iOS is out now, check this out.
MacOS version is coming soon too.
Thank you, this is great! Looks like the beta is on Mac OS as well. I set the slider to 20 here and the pauses are much shorter.
Slightly unrelated, but would it be possible to add proper labels to the sliders? At least on my iPhone, I'm getting no label information and with the growing amount of slider, this gets a bit complex.
@leonardder sure, I will check accessibility.
Not sure if I'm doing something wrong, but I can't tell any difference when setting the SSML multiplier to 0 or 100. I've also tried restarting VoiceOver. What am I doing wrong?
Here, with the rate at 375 and the multiplier at 10 I hear a very noticeable reduction in pauses. My vo rate is at around 55 percent
Ah, this was my mistake. For anyone who couldn’t perceive the change, the best way to perceive it is to lower the slider, and read a paragraph with 2 sentences in it. yOu will notice that if you set the slider to 0, there is almost no pause after the period.
On 28 Feb 2023, at 09:45, pitermach @.***> wrote:
Here, with the rate at 375 and the multiplier at 10 I hear a very noticeable reduction in pauses. My vo rate is at around 55 percent Sent from my iPhoneOn 28 Feb 2023, at 09:13, Parham Doustdar @.> wrote: Not sure if I'm doing something wrong, but I can't tell any difference when setting the SSML multiplier to 0 or 100. I've also tried restarting VoiceOver. What am I doing wrong?Sent from my iPhoneOn 28 Feb 2023, at 08:34, Yury Popov @.> wrote: @leonardder sure, I will check accessibility.
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: @.***>
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: @.***> — Reply to this email directly, view it on GitHub https://github.com/espeak-ng/espeak-ng-ios-app/issues/40#issuecomment-1447787330, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACWCGZREB4BUGCNJ6W7ZE3WZW3LRANCNFSM6AAAAAATDUDUTI. You are receiving this because you commented.
For me in voice over pauses between utterances are not beeing reduced.
@djphoenix Thanks! Here I can also notice reduced pauses with the new SSML slider.
Works great on my side as well. Thanks for your wonderful work!
First of all thanks for all your work on this Espeak port so far!
Before the changes made to Espeak's sonic support, increasing the rate in the app above 450 made it possible to reduce the long pauses that VoiceOver on both Mac and iOS requests from Espeak (see the discussion in #6). However, when Espeak was updated so that SSML pauses stay consistent even if Sonic is used for speed up, this stopped working and now the pauses between utterances stay very long even if the rate is set very fast.
So, my suggestion for this is adding a reduce pauses slider to the app settings which would, well, reduce the pauses requested by SSML before the speech is generated. The way I imagine it is a percentage based reduction, so IE if the slider would be set to let's say 50% and VoiceOver requests a 1000 MS pause, it would get reduced to 500 before being passed to the synth. This way the user can decide if they want the original behavior with very pronounced pauses or if they want them to be shorter to make the synth speak smoother.