Dictation start takes long

[timo]

Hello,

First of all, I want to thank you guys, the program works like a charm and is really a great help. I purchased it today only and hope to increase my productivity at work. One problem I have is, that when I want to start a dictation ("Start Dictation Mode") it takes about 2.5-3 seconds from the command being recognized until the dictation mode starts. Disabling the clearing of the cache didn't change anything. Any idea as to why this is the case?

Best regards,

timo.

[Pfeil]

Does the delay still occur if you create a new speech recognition profile?

[timo]

Yes, it still occurs. Similarly to this, when I stop the dictation mode, it is also delayed by the same amount.

[Pfeil]

Did you restart VoiceAttack (or your computer entirely, which is worth trying as well) after creating the new profile?

Does the Windows Speech Recognition application appear to be working correctly?


Just in case, try a system file scan

You can also try reinstalling VoiceAttack itself (configuration and profile data should be preserved)

[timo]

Hey Pfeil,

thanks for the reply.

The windows speech recognition application works perfect. The actual recognition of commands in  voice attack also works perfect. Just starting and ending the dictation has the delay, which interrupts the workflow a lot.

I tried all the things you suggested, still the same problem.

[Pfeil]

Unfortunately, this is likely a more deeply-rooted issue with your (speech recognition) system.

While the system file scan can repair various important components that could cause all kinds of issues, experience shows that repairing the speech recognition system itself requires repair-installing Windows.

This can be done by following these instructions, and choosing the "Keep personal files and apps" option, which should preserve your files, applications, and settings.

Note that making backups is recommended, and that this is at your own risk.

Also note that the instructions stop at "Your PC will restart several times during the reinstallation."; do not follow the instructions in the "Clean install of Windows 10 using installation media" section below that, unless you want to wipe your system of all data entirely.

[timo]

Ok, at least good to know, where the issue is coming from.

Thanks!

Best regards,

timo.

[timo]

Hey it's me again.

So I installed a fresh Windows 10 Pro today on my machine, but I run into the same issue. Could it be a hardware related issue? I tried it on my laptop and there was no problem.

Best regards,
Timo.

[Pfeil]

"fresh" as in repair-install, or as in clean, completely-from-scratch install (from the installation media, not using the system builder's recovery system, if applicable)?


If it occurs on an installation that has no real configuration changes, and no manufacturer bloatware utilities installed, that would be quite strange.

Theoretically I suppose it could be a driver issue, but given that this issue is extremely rare, that would have to be something fairly exotic, perhaps a specific combination that causes it.


What soundcard are you using? Do you have a USB headset?

[timo]

"Fresh" as in clean, completely-from-scratch install. No manufacturer utilities installed as the install was made with the windows 10 media creation tool.

I am using an external audio device: Focusrite Solo 3rd Gen. I already tried changing the "sample rate" and "buffer size" of the driver and also not using it entirely (I also have a non-USB headset on my onboard soundcard). No luck...

[Pfeil]

Did you happen to try starting the dictation mode before installing/plugging in the Focusrite device?

[timo]

Yes it seems to be independent of the audio interface/it's driver.

What could other components/drivers be that can cause issues?

[Pfeil]

The sound card would have been the most likely candidate, as obviously the speech recognition engine uses that device. Other than that there is no real method to narrow it down, that I know of, especially considering that aside from drivers, just about everything else should be pretty much standard after a clean install.

What language and region are you installing Windows in? US English?


If you start VoiceAttack with the "-showloadtime" argument, using an empty profile (I.E. no commands in it at all), what are the times shown in the log?

[timo]

Language and region are German and Germany respectively. I did install US English for VA though. The same setup is used by my laptop and it works like a charm.

Times shown in VA:
3:43:49.646 Profile load time: 94,9782 ms  (Speech engine: 21,9794 ms)

It's not a dealbreaker, as I can also adjust to not using dictation. Thanks for the support so far.

[Pfeil]

That looks fairly normal (mainly I wanted to make sure the initialization of the speech recognition engine doesn't also take several seconds).

I would be curious to know whether the issue occurs with both the German and the US English speech recognition engine, though it very well might, both considering that you're using the same setup on another machine, as you mentioned, and that there are other users also using a different engine language than originally installed without reporting this issue (though not everyone necessarily uses the dictation mode).

Also out of curiosity, if you disable speech recognition (by enabling the "Disable Speech Recognition" option on the "Recognition" tab of the VoiceAttack options window), does the "Dictation functionality not available." log message take a while to appear after attempting to start the dictation mode as well?