Pause duration problem with different voices

[dunkmyster]

When inserting pauses, ie. {{Pause=10}}, I have found that my AT&T and Nuance voices pause for approximately the right time, but the Scansoft and Acapela voices don't. I noticed this problem was reported 6 years ago: viewtopic.php?f=13&t=6213. Is there any solution yet? Thanks.

[DavidW]

I have the same problem with pauses being omitted.

The problem is voice dependant. If I use the Microsoft Windows 10 English voices (Hazel Desktop and Zira Desktop), the pauses are inserted correctly. If I use my IVONA voices (Amy, Brian and Emma - version 1.6.70, licensed directly from IVONA in the days when that was still possible - I am not entitled to further updates without buying these voices again), I hear an 'mmm' sound where the pause should be, but there's no pause. If I use my Cereproc voices (Heather and Stuart - version 4.0.1, which I believe is the latest version), I do not get a pause and have no audible indication of where the pause should be. The problem is present whether I route audio to my speakers or to an audio file.

I'm using TextAloud version 3.0.110 on Windows 10 Pro, which I believe is the latest version. I'm very happy to test a fixed TextAloud version 3 build or a TextAloud version 4 beta. I'm using TextAloud to read revision notes for my upcoming law exams, so the lack of pauses is a huge problem!

I will also send this via the "Report A Problem" option in TextAloud.

[Jim Bretti]

If I understand correctly, the first post in this thread has to do with the accuracy of longer pauses with some voices, for example {{Pause=10}}. I know we've seen this issue with Acapela, the only solution I know of is to break pauses over some limit, lets say 5 seconds, into multiple smaller pauses. So in TextAloud pronunciation dictionary maintenance, set your own pause 'tags', something like this:

Text Matching:Simple Text
Pause_10

Pronounce As: Respell
{{Pause=5}} {{Pause=5}}

So the idea is to define the pause durations you need need as tags, Pause_15, Pause_20, etc, and have dictionary entries that respell these tags as a series of 5 second pauses.

DavidW - I'll reply to your email about your Ivona licenses

[DavidW]

Thanks for your e-mail, Jim. I have used the information you sent to devise a fairly simple test case that seems to reveal two voice dependent bugs in TextAloud and I have e-mailed you full details. The Ivona voices seem to prevent TextAloud noticing newlines, so "Between Paragraphs" rules don't work (bug #1). The Cereproc voices do not appear to respect any pauses, be they explicit pauses in the text or a pause in a rule (bug #2).

dunkmyster's post suggests that bug #2 also happens with Scansoft and Acapela voices (I wrote this the wrong way round in my first e-mail; I've sent a correction); I cannot verify this as I don't have any of these voices, but I hope this is the same issue that I'm seeing with my Cereproc voices.