Maximum practical profile size

[Astan]

I've been using VoiceAttack for several years now, and my "General Desktop" profile has increased in its size considerably, now hosting several hundreds of (voice-activated) commands, some of them also multi-segmented. I presume that a large number of commands in a profile (whatever that "large" number would be) may alone cause VoiceAttack to operate sub-optimally.

Does it help to split one large profile into two or several smaller ones, to increase the app's overall performance (or: the accuracy of voice recognition)?

[Pfeil]

If you hover over the command count at the top of the "Edit a Profile" window, how many derived commands do you have?

[Astan]

Thanks for the (tool) tip, it says 2820. Is that still within the frames of "acceptable"?

Also, I just launched VoiceAttack with the parameter "-showloadtime", and I can read:
"Profile load time: 305.4098 ms (Speech engine: 223.5968 ms)".

[Pfeil]

That is a very minimal amount of command phrases. There are users with hundreds of thousands, even millions of commands.

If you are experiencing recognition issues, that amount of commands is not going to be a factor.

There are a number of tips in this topic that may be relevant


The load times on my machine are nearly identical, with just over 4200 derived commands, for what that's worth.

[Astan]

Many thanks for your insight, Pfeil. I bookmarked the thread you linked to.

I started wondering about this, because I recently noticed slowdowns and issues with cursor movement which sometimes occur with VoiceAttack (/ Windows Speech Recognition) running in the background.

The issues go away when I close the app (I don't use WSR outside of VoiceAttack). I already started troubleshooting with following the steps described in this topic on resetting the recognition profile.

This goes off topic from here, as it looks like it was a false track on my end. I'll start a new post to discuss this further (weird performance issues).