Things were working fine for the first few days after my purchase, but now my copilot just does whatever he wants replies to himself, hears his own keywords and goes in a loop, it has become completely unusable. I did a complete speech training session prior to use.

12 hours ago, raelonmasters said:

Things were working fine for the first few days after my purchase, but now my copilot just does whatever he wants replies to himself, hears his own keywords and goes in a loop, it has become completely unusable. I did a complete speech training session prior to use.

The co-pilot doesn't suddenly go in a loop talking to himself.

A speech profile is a collection of 3 things.

A specific headset

A specific user voice

Specific ambient noise conditions.

Change any of the above and you should make a new ad-hoc speech profile in Windows Control Panel and train it from scratch.

Do not overdo speech training though. One session per topic using our speech trainer should be enough in Windows 10.

Start->Multi Crew EXperience->Speech training.

For instance, if you want to mpress the audience and decide FO voice will come out of speakers, then of course some of that speech will feed into the mike and may cause Fo to hear himself occasionally(but not go in a loop).

It's not impossible to use it that way, but need to understand the speech profile is trained against an almost silent background so won't filter out engine sounds FO speech and such. You would need another profile where diffeent ambient noise conditions are taken into accout.

At the very least, would need to econsider speakers positioning so that they don't feed into the mike and reduce volume level.

 

4 hours ago, FS++ said:

The co-pilot doesn't suddenly go in a loop talking to himself.

A speech profile is a collection of 3 things.

A specific headset

A specific user voice

Specific ambient noise conditions.

Change any of the above and you should make a new ad-hoc speech profile in Windows Control Panel and train it from scratch.

Do not overdo speech training though. One session per topic using our speech trainer should be enough in Windows 10.

Start->Multi Crew EXperience->Speech training.

For instance, if you want to mpress the audience and decide FO voice will come out of speakers, then of course some of that speech will feed into the mike and may cause Fo to hear himself occasionally(but not go in a loop).

It's not impossible to use it that way, but need to understand the speech profile is trained against an almost silent background so won't filter out engine sounds FO speech and such. You would need another profile where diffeent ambient noise conditions are taken into accout.

At the very least, would need to econsider speakers positioning so that they don't feed into the mike and reduce volume level.

 

I wish he didn't go and just start talking to himself but that's exactly whats happened, I haven't changed any of the above variables, I don't use external speakers, i have a razor gaming headset i use, which had no problem in the beginning, and still works perfectly with voice attack. I sit alone in an office so external noise isn't a concern. I will attempt retraining.

Yah just did retraining, I ask for weather, he gives weather then triggers on something trying to get runway information, then asks which runway captain, then goes on to as for an identifier, then pauses to wait, then continues asking if im not interested, and on and on it goes, him carrying on a conversation. Ideas? other then the above?

20 minutes ago, raelonmasters said:

Yah just did retraining, I ask for weather, he gives weather then triggers on something trying to get runway information, then asks which runway captain, then goes on to as for an identifier, then pauses to wait, then continues asking if im not interested, and on and on it goes, him carrying on a conversation. Ideas? other then the above?

You may be coming from a different type of voice control, and need to understand the following

MCE is a fully interactive Crew Simulation. Tens of thousands of speech commands are available at any given time for speech reco. Unlike VoiceAttack which doesn't care about item status and just triggers things, MCE ensures switching is done correctly. Not to mention speech commands such as "obtain taxi clearance" or "before start checklist" to trigger checklist reading. And don't forget Voxkey could add a few more (could transfer the ones you have in VoiceAttack so you don't need to run 2 instances of the speech engine).

Start->Multi Crew Experience->Voxkey.

There are other things that are triggered by events. For instance FO requests runway in use when you make a flap selection, arm spoilers or other. This is to help him handle ATC better.

Critical items are safeguarded against accidental trigger. Fo will typically prompt for confirmation. Unless you agree, after keeping silent for the next 10 seconds or so, he'll ignore the critical commands. Try "shut down engine one", "gear up" on ground (VoiceAttack would attack the switches and levers immediately 🙂 and make you very nervous).

You can customize Fo behaviour by going to "General" tab and clicking <More options>.

If you find FO too chatty, disable option "Talks a lot" and make all your flows non-Verbose for silent execution.