A tool to generate voice files using various TTS providers

erskyTx runs on many radios and upgrade boards
ersky9x was a port of er9x for use on the sky9x board.
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

As the 12 month free tier for Amazon Polly will be expiring in a couple of days, I've created a website to host TTSAutomate updates from now on.

TTSAutomate is still free to download and use, however I've added an no obligation option to purchase 'support' through PayPal in case anyone would like to contribute to the costs of running it.

I'm still working on the website. I hope to have a lot more content there shortly about how to use TTSAutomate, with plenty of example PSV files, plus links to OpenTX, er9x and any other projects that may benefit from the voices that TTSAutomate can produce.

The website is live at TTSAutomate.com

User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: A tool to generate voice files using various TTS providers

Post by jhsa »

I am in, just need to get home tomorrow. This is a wonderful project, and a life saver for us who make our own voice files..

Thank.you

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

Thanks João. It's a bit of a gamble on my behalf, but I'd rather try to keep the tool alive and hopefully cover the cost, instead of abandoning it 😎

I'll be adding an acknowledgement section to the website shortly, and I'd like to put your name in there for your assistance in troubleshooting, beta testing and helping with translations. Is that ok with you? Do you have a website or something I can link in there?
User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: A tool to generate voice files using various TTS providers

Post by jhsa »

Thanks, that is fine... Any testing you need done, please let me know..
I don't have any website, but thanks anyway :)

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

I linked to your YouTube tutorials :)

For some reason, my site has trouble displaying the accented character in your name, so I'll have to investigate that.

User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: A tool to generate voice files using various TTS providers

Post by jhsa »

:mrgreen: thanks, I still have to finish the Ersky9x video manual.. will also have to edit some of the videos as there are some new features and options implemented :)

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

I had planned to complete the audio editor functionality, however it's pretty complex (even more than the rest of the program, that's why it's been sitting unloved for 12 months) so I've put that aside for now. I'm adding a trim silence feature, which will automatically trim silence at the start and end of the generated voices below a configurable dB threshold (defaults to -48dB)

I'm also looking at using Squirrel to handle installation, deployment, and checking for new versions, as the current installer I'm using (Microsoft's own installer project type) isn't supported well in visual studio 2015 or 2017.
User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: A tool to generate voice files using various TTS providers

Post by jhsa »

Nice features, thanks..

João

My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

I tried out squirrel and I have a couple of concerns:.
  • It silently installs new versions of the software without acknowledgement by the user
  • There is no mechanism for having portable vs installed versions (and no mechanism for associating PSV files with the software)
  • There's no way to indicate to the user what has changed in the latest update.
On the plus side, it's very fast and installs without interrupting the use of the software.
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

Hrmm, it turns out that you can actually ask the user whether they want to apply the latest update, and also show release notes. Still working on whether associating with file types is possible.
User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: A tool to generate voice files using various TTS providers

Post by jhsa »

Thank you..
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

A question on automatically trimmed files; Would you like to see the folder layout as follows?

Code: Select all

MP3
|_SOUNDS
|_System
|_User
|_Trimmed
	|_SOUNDS
	|_System
	|_User
Wav
|_SOUNDS
|_System
|_User
|_Trimmed
	|_SOUNDS
	|_System
	|_User
  
i.e. trimmed as a subfolder of the wav and mp3 folders

or would you prefer the distinction between trimmed and not trimmed further up the folder tree?

Code: Select all

MP3
|_SOUNDS
|_System
|_User
 Wav
|_SOUNDS
|_System
|_User
Trimmed
	|_MP3
		|_SOUNDS
		|_System
		|_User
	|_Wav
		|_SOUNDS
		|_System
		|_User  
User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: A tool to generate voice files using various TTS providers

Post by jhsa »

I think I would prefer the second.. :)

Also, is that SOUNDS directory for OpenTX? I am asking because Ersky9x doesn't seem to have it.
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

Probably yes, I just opened the most recent export that I have
User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: A tool to generate voice files using various TTS providers

Post by jhsa »

Ahh, ok.. thanks.

I don't know if you already did it, but what about an option to select between Ersky9x and OpenTX? This would allow to save the right file structure, ready for the SD Card.. :)

Thank You

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
budavaril
Posts: 56
Joined: Wed Feb 17, 2016 9:34 am
Country: United Kingdom
Location: London

Re: A tool to generate voice files using various TTS providers

Post by budavaril »

First I have to say BIG Thanks about this tool to Caffeine.

After nearly 3 years I found the right information here in the forums and on Youtube (thanks João) and finally I was able to change the firmware on my 9XR Pro. (Until now I used with r212 because I misunderstood and thought the 3position switch support has been removed.) So a week ago I flashed to r222b6 and I love my "new" controller. :D

I "playing" now 4-5 days ago because when I thought I'm done, I also found that, the voices now possible to setting easier in this new firmware (Thanks Mike..) I never understood how I can adjust voices and what's the contact between voices and "safety switches".. so Until now my controller all voices was the factory default "Landing gear" messages :lol:

Wasn't easy, but found in the forums that informations, the folders structure and filenames has been changed, but with this tool and João's Phrase file I found the final solution.

I just have a little problem that, the generated voices are not enough loud. I saw in a
previous post the volume should be much louder than my. (And my original voices was louder as well)
Is it possible to increase the volume somehow? Because now I can hearing in the quite room, but if have prop noise or wind outside I definitely can't hearing anything. :(
I generated Amazon Polly/Sally.
Volume_dB.PNG
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

The volume of voices can be adjusted using the SSML prosody tag as per the following link: https://docs.aws.amazon.com/polly/lates ... rosody-tag

Using SSML is bit advanced but very powerful.

An example:

<prosody volume="x-loud">this is the maximum volume</prosody>
budavaril
Posts: 56
Joined: Wed Feb 17, 2016 9:34 am
Country: United Kingdom
Location: London

Re: A tool to generate voice files using various TTS providers

Post by budavaril »

That's good news.. And a little bad news.. I checked that, but I'm afraid those things are very "advanced" to me... :roll:
Unfortunately my limit much lower level than find how I can using this options in your software.
I can modify text based config files (like config a Samba or Apache server-MySQL server under OpenWRT or Ubuntu..etc), or a bit modify some simple Arduino sketches, but I have no any idea in this case to be honest.. :cry:
Caffeine
Posts: 170
Joined: Wed Jul 27, 2016 5:29 am
Country: Australia

Re: A tool to generate voice files using various TTS providers

Post by Caffeine »

If you put:
<prosody volume="x-loud">
At the front of your phrase to speak and:
</prosody>
At the end, it should work
budavaril
Posts: 56
Joined: Wed Feb 17, 2016 9:34 am
Country: United Kingdom
Location: London

Re: A tool to generate voice files using various TTS providers

Post by budavaril »

Wow... That's working. Thanks.
This way make sense to play with another SSML functions to try to make the best result for this "not really the best" speaker.. 8-)
Thank you.

(Before ad After)
before-after.PNG
budavaril
Posts: 56
Joined: Wed Feb 17, 2016 9:34 am
Country: United Kingdom
Location: London

Re: A tool to generate voice files using various TTS providers

Post by budavaril »

Done. I worked 2 days on it, but now successful.
The "X-loud" was much better, but still less volume than original (factory) files. I found on that website another "dynamic compressor" effect (drc) and that was a bit more understandable, but similar loud than "X-loud".
So finally I combined both function and that I think perfect.
Was a bit headache, because I didn't want paste each TAGs each line (332 line 664 paste) so I tried editing used Excel, but that (because the code format) can not save worksheet data as csv file without put a plenty of double quotes (") in to the text. :cry:
I found a half solution (a VBA code) and with that I could Export without double quotes, but just with "TAB" separators. :evil:
With another csv editor, what couldn't do that job than Excel (fill and joining multiple columns) I exported the file and changed the separator from "TAB" to "|" so I got a proper phrase file.
Compared the effect: (1.Normal _ 2.X-loud _ 3.drc _ 4.X-loud+drc)
Compare Norm-Xvol-drc-Xvoldrc.PNG
Can using in same file those all levels: (same sentences repeated in 4 effects)
Combi Norm-Xvol-drc-Xvoldrc.PNG
And here is the modified Phrase file:
ErSky9x Phrase file LOUD budavaril.zip
(3.12 KiB) Downloaded 261 times

Post Reply

Return to “erskyTx (was ersky9x)”