On ‎12‎/‎10‎/‎2018 at 7:14 PM, RALF9636 said:

I can confirm that I experience the same.

It also seems to be dependent on the used voices because for me it is much worse with Ivona voices (which work perfectly outside the ATC readbacks) than it is with Cereproc voices (but still noticeable also with the latter).

 

Ralf, does this mean that the cereproc voices are working as best as possible? I´ve tested an ivona voice for the trial period of 30 days and it worked much better on the readback than the default Win10 TTS. But because those voices consuming some money and cereproc doesn´t offer a trial period, I would know your opinion if I buy a cereproc TTS would make me more happy as I already am with MCE and PF3 together

10 hours ago, BerndB said:

Ralf, does this mean that the cereproc voices are working as best as possible? I´ve tested an ivona voice for the trial period of 30 days and it worked much better on the readback than the default Win10 TTS. But because those voices consuming some money and cereproc doesn´t offer a trial period, I would know your opinion if I buy a cereproc TTS would make me more happy as I already am with MCE and PF3 together

It is correct that the default Microsoft voices are the worst by far. I never use them. As i said Cereproc voices are faster with the readback than the Ivona voices. Also my Ivona voices often say "oh" (like in "flight level three two oh") instead of "zero", which never happened with my Cereproc voices. Finally Cereproc voices tend to sound better when I try to reproduce some german words or procedure names.

And I think Cereproc still has a new year sale going on with a 50% discount.

So, yes: Go for it!

You can start with Giles and Heather, but that is just a personal choice. I think the quality of all the Cereproc voices is the same.

Edited January 21, 20197 yr by RALF9636

On ‎1‎/‎21‎/‎2019 at 8:36 PM, RALF9636 said:

It is correct that the default Microsoft voices are the worst by far. I never use them. As i said Cereproc voices are faster with the readback than the Ivona voices. Also my Ivona voices often say "oh" (like in "flight level three two oh") instead of "zero", which never happened with my Cereproc voices. Finally Cereproc voices tend to sound better when I try to reproduce some german words or procedure names.

And I think Cereproc still has a new year sale going on with a 50% discount.

So, yes: Go for it!

You can start with Giles and Heather, but that is just a personal choice. I think the quality of all the Cereproc voices is the same.

it´s done! I´ve purchased the Giles voice and it works very nice with the ATC. The latter is still noticeable but only for the readbacks. Anything else is really natural. I hope, well, i´am pretty sure that the MCE-Team will find a solution in the future for this little issue left. Thank you for the tip with the new year discount. Otherwise I wouldn´t seen this.

reg.

Ivona voices have a user-editable pronunciation dictionary which affects all apps. 

4 hours ago, jabloomf1230 said:

Ivona voices have a user-editable pronunciation dictionary which affects all apps. 

Thanks!

I will take a look at that.

 

On ‎1‎/‎22‎/‎2019 at 6:36 AM, RALF9636 said:

Also my Ivona voices often say "oh" (like in "flight level three two oh") instead of "zero", which never happened with my Cereproc voices

 

On ‎1‎/‎23‎/‎2019 at 10:57 PM, jabloomf1230 said:

Ivona voices have a user-editable pronunciation dictionary which affects all apps

Team

This issue is curious to say the least & still not solved with Ivonna lexicon (one might need to know the character string at fault to succeed with the lexicon).

For me, it is only an issue with MCE saying transponder codes / speed / heading / flight level.

It is not an issue for other TTS voices with MCE nor is it a problem for any voice (Ivonna included) with VoxATC or Textaloud (noting that VOX uses the 64bit voicesets, but MCE & Textaloud use 32bit).

The "bottom line" nuisance is that I prefer the same Ivonna voice as my co-pilot in both VOX & MCE.

You mean that if you add the digit 0 to the Ivona lexicon and have it set to translate it to the word "zero", it doesn't work in MCE? Sounds like an MCE issue. If you open the Windows speech control panel applet (not the Win 10 settings app which only shows MS voices) and switch to an Ivona voice, it will pronounce 3500 as "thirty five hundred" and 3 5 0 0 as "three five zero zero". 

Control Panel\Ease of Access\Speech Recognition\Text to Speech

I don't use MCE, but VOXATC has the numeric translations built in to its code, including the digit 9 being pronounced as "niner". You can see how it works by looking at the file voxgram.xml. It defines a character set as "DIGITS".

Does MCE have a language/region setting? You can also see how VOXATC does that for "." defining it as "point" by default, but as "decimal" in the UK, in VASettings.xml. There are no regional settings for either zero or any other digits though.

3 hours ago, jabloomf1230 said:

You mean that if you add the digit 0 to the Ivona lexicon and have it set to translate it to the word "zero", it doesn't work in MCE?

not for me.

3 hours ago, jabloomf1230 said:

Does MCE have a language/region setting?

can't see it in their folders.

I wonder if it may be caused similarly by that "ongoing hassle" found in VOX & P2A when Ivonna is being used in parallel with Cereproc voices.

7 hours ago, vadriver said:

not for me.

can't see it in their folders.

I wonder if it may be caused similarly by that "ongoing hassle" found in VOX & P2A when Ivonna is being used in parallel with Cereproc voices.

Jabloomf1230 is right, should be possible.

Might be to do with the speech recognition panel you're editing voice lexicon under.

The one you access via Windows Control Panel is a 64 Bit app (on Windows x64) and will edit the 64 Bit voice.

Since MCE is a 32 bit voice it can only load 32 Bit voices.

To see those, go to C:\Windows\SysWOW64\Speech\SpeechUX\ folder and run "sapi.cpl wwhich is the 32 bit version of the speech recognition panel.

Note that making MCE a 64 bit app would be a simple recompile same source code to 64 Bit and check everything works as expected.

It's just that we don't see a need for that as it barely uses even a quarter of its massive 4 GB VAS Windows assigns to it. We're not doing maps or huge databases to justify the move.

And the bonus, old 32 TTS voices can be used as well, whereas a 64 Bit app can only load 64 Bit voices.

Sometimes small is beautiful.😛

 

11 hours ago, FS++ said:

To see those, go to C:\Windows\SysWOW64\Speech\SpeechUX\ folder and run "sapi.cpl wwhich is the 32 bit version of the speech recognition panel.

gerald

thanks for your thoughts, you too jay.

I had used the 32bit panel / lexicon (?) ..... to no avail.

I'm more sure it's this "issue / conflict" between parallel use of Ivonna & Cereproc voices ..... remove either & no issues.

11 hours ago, FS++ said:

And the bonus, old 32 TTS voices can be used as well

except with P3DV4 + Vox or P2A, one can't use the same co-pilot unless of course you have recent sets which have both 32 & 64 bit such as the Cereproc / Ivonna's which I have (ps .. I use voicescripts to initiate co-pilot functions in VOX & GSX)

11 hours ago, FS++ said:

Sometimes small is beautiful

that said, I'm more than happy with a small MCE

Edited March 12, 20197 yr by vadriver

I gave this a little more attention again and here are my findings.

 

The entries in the Ivona lexicon (under both paths mentioned here before) seem to be irrelevant for this issue.

 

The MCE FO with an Ivona voice pronounces "zero" correctly for example in these situations:

- replying to "what is our callsign?"

- reading back an altitude clearance

- reading a notify command in a voxscript (using the digit 0).

 

The FO pronounces "oh" instead in these situations:

- replying to "what is the wind?"

- notifying the altitude when checking in with ATC

- requesting a different altitude.

 

So for example I say

"request flight level three four zero"

FO then says 

"<callsign> requesting higher altitude, flight level three four oh"

ATC approves and FO reads back

"roger, climb and maintain flight level three four zero"

 

So there must be a difference in how MCE puts out the "0" in these cases.

Again, this does not happen with Cereproc voices, they always pronounce "zero".

 

 

 

 

Edited March 14, 20197 yr by RALF9636

Quote

On 3/14/2019 at 1:39 PM, RALF9636 said:

The MCE FO with an Ivona voice pronounces "zero" correctly for example in these situations:

- replying to "what is our callsign?"

- reading back an altitude clearance

- reading a notify command in a voxscript (using the digit 0).

 

The FO pronounces "oh" instead in these situations:

- replying to "what is the wind?"

- notifying the altitude when checking in with ATC

- requesting a different altitude.

 

I can confim this issue because i have the same behavior of the MCE FO with Ivona voice. 

By the way Gerald, is there anything new from your Team About the issue that only the ATC-readbacks by the FO are extremly slow in case we use TTS-Voices?? May be there is a connection between These two issues?!?

I just want to make sure that you will not be lazy and chilling around instead of working as harder as you can in serving us with resolving issues. (sorry please, just kidding)

You know that MCE does make a huge difference in the Immersion. It´s so cool and real with the FO at your side. After i updated to the act. Version even my Purser is Talking to me. First time is was scared, because i said "Cabin Crew, please prepare for takoff" while taxiing, just for fun because i never got to hear from them, but that time suddenly my Purser said with her sexy voice "you´ll be happy to hear that the cabin is Ready for takeoff". That´s so nice. Next time i will ask to bring me a coffee into the flightdeck and lets see how MCE will handle this...LOL