[GUI Plugin] TTS to MP3: Create MP3 audiobook using Windows TTS

[wiso]

Originally Posted by jackie_w

Unfortunately, due to unexpected issues at home, I am unable to provide any support for this plugin for the forseeable future.

Hopefully the current version will continue to work OK in calibre 5.xx

May the force be with you and yours. best wishes.

[KindleAZWorKFX]

Quote:
Originally Posted by jackie_w View Post
Unfortunately, due to unexpected issues at home, I am unable to provide any support for this plugin for the forseeable future.

Hopefully the current version will continue to work OK in calibre 5.xx
May the force be with you and yours. best wishes.

So any updates on using natural voices from windows 11 in TTS to MP3 , I almost typed TTS to MP4 format which it is not, I noticed that read aloud in edge has a lot more I guess online only TTS voices but they cannot be accessed by this plugin in calibre 64bit version, I have not tried the 32bit version but I am guessing the voices are limited, also did this plugin support stop with windows 10? or does this plugin work correctly under windows 11 too? not sure why but the learn how to add more voices button links to a page that seems to talk more about windows 10 than even windows 11 at all, it doesn't seem like it's been updated even on the new OS system. I guess the process it almost the same between the systems that it may not matter in any case. for example, under voices in read aloud under Microsoft edge on windows 11, like the voice called "Microsoft Aria Online (Natural)" its voice is not usable by this plugin.

https://support.microsoft.com/en-us/...rs=en-us&ad=us

[jackie_w]

Originally Posted by KindleAZWorKFX

So any updates on using natural voices from windows 11 in TTS to MP3 , I almost typed TTS to MP4 format which it is not, I noticed that read aloud in edge has a lot more I guess online only TTS voices but they cannot be accessed by this plugin in calibre 64bit version, I have not tried the 32bit version but I am guessing the voices are limited, also did this plugin support stop with windows 10? or does this plugin work correctly under windows 11 too? not sure why but the learn how to add more voices button links to a page that seems to talk more about windows 10 than even windows 11 at all, it doesn't seem like it's been updated even on the new OS system. I guess the process it almost the same between the systems that it may not matter in any case. for example, under voices in read aloud under Microsoft edge on windows 11, like the voice called "Microsoft Aria Online (Natural)" its voice is not usable by this plugin.

https://support.microsoft.com/en-us/...rs=en-us&ad=us

At the moment I'm not able to devote time to major enhancements of this plugin. However, I will be releasing a new version which allows it to continue working when calibre moves from v5.x to v6.x later this year.

The list of Voices accessible to this plugin should be the same list as is accessible in the calibre Viewer's ReadAloud feature. To the best of my knowledge that will be any Voices under a specific Windows registry key. See the first paragraph of post #2 in this thread. I don't think any of the MS Online (Natural) Voices appear under that key.

Do you see them stored elsewhere in your registry? If I open an HTML file in the MS Edge browser on Win10 I can see/use some MS Online Voices to listen to the browser text in real-time. But that's a long way from being able to automatically record to a local .WAV file in much-faster-than-real-time, which is how the plugin currently works.

I wouldn't expect it to be any different in Windows 11, but I can't be sure because I don't have access to a Win11 PC. This is unlikely to change in the near future.

Kovid Goyal, the creator of calibre, has far better in-depth technical knowledge than I do. It might be worth asking a question in the calibre sub-forum whether he thinks it's ever likely to be possible to access MS Online Voices in calibre ReadAloud. If his response is positive then maybe this plugin could piggyback on whatever he does in main calibre. I suspect neither he nor I would be prepared to develop anything which requires special logins or paying money to MS for the privilege of using their best Voices.

[jackie_w]

• Maintenance upgrade for PyQt6/calibre6 compatibility.

The update should be available within a couple of hours via the normal calibre plugin update system (or directly from post #1).

Please note: this will be my final contribution to this plugin.

[Venia Legendi]

.. for the time you spend, all the best.

[QuDE]

Originally Posted by jackie_w

• Maintenance upgrade for PyQt6/calibre6 compatibility.

The update should be available within a couple of hours via the normal calibre plugin update system (or directly from post #1).

Please note: this will be my final contribution to this plugin.

Your effort is much much appreciated! it's been a blessing to be able to listen to my Light Novels through my long rides!

I'm curious to know how other TTS voices are, if possible could you send me the .reg for the other voices?

Thanks again and I wish you the absolute best

[KindleAZWorKFX]

Quote:
Originally Posted by jackie_w View Post
Maintenance upgrade for PyQt6/calibre6 compatibility.

The update should be available within a couple of hours via the normal calibre plugin update system (or directly from post #1).

Please note: this will be my final contribution to this plugin.

I Also will thank you for your effort is much appreciated!

I myself been off these forums doing other stuff at home these days not related to calibre stuff.


So, under Accessibility > Narrator in Windows 11, not sure about windows 10 but you can now download with Add natural voices a much of English voices like Microsoft Aria (Natural) English (United States) but they seem to be limited to Narrator.

unless we can add them into the regedit pages for text to speech, it would be unofficial but I would be nice if we could get this TTS to MP3 to output the files to the natural English; anyway,

I think this plugin doesn't or might not be updated much anymore, but since we can access the voices somewhat outside of edge now, on windows 11 it should be possible to get the text to speech , or maybe if Microsoft adds them into the normal TTS system when we could also get the eBook Reader that calibre comes with to use them with read aloud feature it has, outside of the TTS to MP3 plugin here.

I am not sure if windows 11 plans to release them outside of Narrator app in the future or not, in any case I put a feedback hub suggestion to add them into text to speech I don't know if they will, but it would be nice if we had better sounding voices for free in calibre. Under Suggest a Feature in Feedback Hub, again I don't know if that will help get them to do so, but it would be nice if we could get them working in the next few years in calibre.

Link to my post on Feedback hub Requesting that they add the natural voices to text to speech for accessing them in calibre eBook reader's read aloud and TTS to MP3 plugin.

My Feedback hub Suggestion link for this.
https://aka.ms/AAp53w5


These are the natural voices I am able to download for Narrator, if they can be downloaded I think they will work offline too, but I haven't tried that yet.




The current natural voice I have set for this Narrator app on my desktop.



as you can see you cannot select the natural voices yet in text to speech.

[jackie_w]

Originally Posted by KindleAZWorKFX

So, under Accessibility > Narrator in Windows 11, not sure about windows 10 but you can now download with Add natural voices a much of English voices like Microsoft Aria (Natural) English (United States) but they seem to be limited to Narrator.

unless we can add them into the regedit pages for text to speech, it would be unofficial but I would be nice if we could get this TTS to MP3 to output the files to the natural English; anyway,

I think this plugin doesn't or might not be updated much anymore, but since we can access the voices somewhat outside of edge now, on windows 11 it should be possible to get the text to speech , or maybe if Microsoft adds them into the normal TTS system when we could also get the eBook Reader that calibre comes with to use them with read aloud feature it has, outside of the TTS to MP3 plugin here.

Last time I looked the newer "natural voices" were not physically present as files in the local copy of Windows. I believe everything is done via sending the text to a MS server and the server sending back the audio, i.e. everything is done via the internet. Whilst this remains the case I don't really see a possibility of harnessing online-only voices in a local calibre plugin like this.

[Quoth]

Exactly.

Also it's a serious privacy / security issue. On Android I only use local TTS files.

[KindleAZWorKFX]

Originally Posted by jackie_w

Last time I looked the newer "natural voices" were not physically present as files in the local copy of Windows. I believe everything is done via sending the text to a MS server and the server sending back the audio, i.e. everything is done via the internet. Whilst this remains the case I don't really see a possibility of harnessing online-only voices in a local calibre plugin like this.

Originally Posted by Quoth

Exactly.

Also it's a serious privacy / security issue. On Android I only use local TTS files.

No, the Natural voices are local TTS files on Windows 11, they are just so far limited to being used by Narrator's voice online and offline.

when you select add natural voices box it downloads them, if you go offline and have it set to Any Microsoft (Natural) - English US or UK, they work with the internet disconnected now, but they only show up under Accessibility > Narrator, text to speech only has Legacy voices for it, so it's not true that they are online only anymore-

So, if you can make say Natural voices work as Legacy voices by changing a few lines in regedit it might be possible to get them to work outside of Narrator, but since I am not a expect in regedit myself, I am not sure I can do that on my own.

again when you click on Add Natural Voices in the (Add) button, that I shared in that shows the details in the image which shows each natural english voices and shows in image each ones download size, I removed each voice in powershell and redownload them afterwards to make a screenshot of the file sizes today each one is small and mb in size and they all work offline in narrator app under windows 11. included the file sizes as a quote here, used (a919) Free Ocr for Windows Desktop to change the text in the graphics into text just to make they easy to post if you for some reason cannot see my image posts in your browsers.

Microsoft Neerja (Natural)
Female, English (India). 32.3 MB

Microsoft Prabhat (Natural)
Male, English (India), 32.3 MB

Microsoft Ryan (Natural)
Male, English (United Kingdom), 20.1 MB

Microsoft Sonia (Natural)
Female, English (United Kingdom), 20.1 MB

Microsoft Aria (Natural)
Female, English (United States), 12.2 MB

Microsoft Guy (Natural)
Male. English (United States), 12 MB

Microsoft Jenny (Natural)
Female, English (United States), 11.4 MB

after you download them in Narrator, in PowerShell app using this command, you can find the files names

Get-AppxPackage -Name "*Voice*"

Name : MicrosoftWindows.Voice.en-GB.Sonia.1
Publisher : CN=Microsoft Windows, O=Microsoft Corporation, L=Redmond, S=Washington, C=US
Architecture : X64
ResourceId :
Version : 1.0.3.0
PackageFullName : MicrosoftWindows.Voice.en-GB.Sonia.1_1.0.3.0_x64__cw5n1h2txyewy
InstallLocation : C:\Program Files\WindowsApps\MicrosoftWindows.Voice.en-GB.Sonia.1_1.0.3.0_x64__cw5n1h2txyewy
IsFramework : False
PackageFamilyName : MicrosoftWindows.Voice.en-GB.Sonia.1_cw5n1h2txyewy
PublisherId : cw5n1h2txyewy
IsResourcePackage : False
IsBundle : False
IsDevelopmentMode : False
NonRemovable : False
IsPartiallyStaged : False
SignatureKind : Store
Status : Ok

Name : MicrosoftWindows.Voice.en-US.Aria.1
Publisher : CN=Microsoft Windows, O=Microsoft Corporation, L=Redmond, S=Washington, C=US
Architecture : X64
ResourceId :
Version : 1.0.8.0
PackageFullName : MicrosoftWindows.Voice.en-US.Aria.1_1.0.8.0_x64__cw5n1h2txyewy
InstallLocation : C:\Program Files\WindowsApps\MicrosoftWindows.Voice.en-US.Aria.1_1.0.8.0_x64__cw5n1h2txyewy
IsFramework : False
PackageFamilyName : MicrosoftWindows.Voice.en-US.Aria.1_cw5n1h2txyewy
PublisherId : cw5n1h2txyewy
IsResourcePackage : False
IsBundle : False
IsDevelopmentMode : False
NonRemovable : False
IsPartiallyStaged : False
SignatureKind : Store
Status : Ok

Name : MicrosoftWindows.Voice.en-US.Guy.1
Publisher : CN=Microsoft Windows, O=Microsoft Corporation, L=Redmond, S=Washington, C=US
Architecture : X64
ResourceId :
Version : 1.0.5.0
PackageFullName : MicrosoftWindows.Voice.en-US.Guy.1_1.0.5.0_x64__cw5n1h2txyewy
InstallLocation : C:\Program Files\WindowsApps\MicrosoftWindows.Voice.en-US.Guy.1_1.0.5.0_x64__cw5n1h2txyewy
IsFramework : False
PackageFamilyName : MicrosoftWindows.Voice.en-US.Guy.1_cw5n1h2txyewy
PublisherId : cw5n1h2txyewy
IsResourcePackage : False
IsBundle : False
IsDevelopmentMode : False
NonRemovable : False
IsPartiallyStaged : False
SignatureKind : Store
Status : Ok

as you can see it downloads files and stores them on your pc for the voices, but the cache is that they are setup so only Narrator can use them so far, that's why I think if you can make them show up as a Legacy voice you should be able to use them with Calibre's eBooks viewer read aloud offline since it works with Narrator offline too. of course, Narrator can be used to read eBooks I think I just have not figured out how Narrator works as a reader in such software. So if and when they work as Legacy voices then you should be able to use them with TTS to MP3 also, since they are locally stored and are not online only compared to edge's pdf read aloud which I think still using them as online only, in Narrator they are downloaded in windows 11 and do work offline, since I took my pc offline today Wi-Fi Connection my our home router, and as far the voice feature in Narrator with these voices do seem to work.

and if you want to remove them just use this type command which I put the link to that page here, for whatever reason if you want to uninstall them it has to be done in powershell.

https://www.thewindowsclub.com/how-t...tor-on-windows




Get-AppxPackage -Name "*Voice*"

Remove-AppxPackage -Package "MicrosoftWindows.Voice.en-US.Jenny.1_1.0.8.0_x64__cw5n1h2txyewy"

How to Remove Natural Voices in Narrator on Windows 11

To remove a Natural Voice in Narrator on Windows 11, you will have to make use of the
command line. So launch Terminal or PowerShell as administrator and run this command

first:

Get-App><Package -Name "*Voice*"

This will give you the details of the Voice packages installed on your computer.

( Image from website went here) it does not translate images within images from (a919) Free Ocr for Windows Desktop and I had to correct one line since OCR is still not perfect and can make mistakes reading images)



Locate the one you want to remove. Let us say it is Jenny, Note down the PackageFullName

for it. lt is:

MicrosoftWindows.Voice.en-US.Jenny.1_1.0.8.0_x64__cw5n1h2txyewy

Then you have to run this command:

Remove-AppxPackage -Package "MicrosoftWindows.Voice.en-US.Jenny.1_1.0.8.0_x64__cw5n1h2txyewy"

You will see a message flash, and then the Voice package will be removed.

My Feedback hub Suggestion link for this so the natural voices get setup as legacy voices post.

https://aka.ms/AAp53w5

I don't know if any of this information is helpful in getting natural voices working as legacy voices, if its possible, I search one entrie of Microsoft Sonia in regedit to see all the places its stored in. first image should be of the 3 legacie voices in text to speech, then the rest of it should be where narrator is storing some of the info about the locally stored natural voices on windows 11. first link is a page on windows 10 and text to speech voices if that helps.

https://www.ghacks.net/2018/08/11/un...-more-of-them/

__________________
My Current Desktop pc info from CPU-Z website.



https://valid.x86.fr/sgnydq

[KindleAZWorKFX]

"Deleted." - No Delete button under Edit, it says Edit/Delete Message on it?

[jackie_w]

Originally Posted by KindleAZWorKFX

No, the Natural voices are local TTS files on Windows 11, they are just so far limited to being used by Narrator's voice online and offline.

when you select add natural voices box it downloads them, if you go offline and have it set to Any Microsoft (Natural) - English US or UK, they work with the internet disconnected now, but they only show up under Accessibility > Narrator, text to speech only has Legacy voices for it, so it's not true that they are online only anymore-

I wasn't aware of that, so thanks for the info. I've just downloaded a few English UK/US natural voices to test for myself in MS Narrator. That's the easy bit done!

Originally Posted by KindleAZWorKFX

So, if you can make say Natural voices work as Legacy voices by changing a few lines in regedit it might be possible to get them to work outside of Narrator, but since I am not a expect in regedit myself, I am not sure I can do that on my own.

All the Legacy voices (including Cortana speech assistants) install themselves under directory

C:\Windows\Speech_OneCore\Engines\TTS\...

Unfortunately all the Narrator offline Natural voices are installed as MS Store apps, i.e. stored somewhere under

C:\Program Files\WindowsApps\...

The Win registry entries for installed Legacy and offline Natural voices don't seem to bear any resemblance to each other, so it's anyone's guess whether "changing a few lines in regedit" is likely to be successful. Unfortunately, messing around intelligently with MS Store apps and the Win registry is way outside my comfort zone.

I'm not an expert in regedit either but with a bit of cautious regedit tinkering (Export > manual edit .reg file > run new .reg to import new keys), I've always been able to (so far) make any voice which installed itself to the Legacy voice directory, available to the calibre Viewer's ReadAloud option and to the "TTS to MP3" plugin.

If calibre Viewer's ReadAloud option can't see the Natural voices then this plugin won't see them either, because all it does is call calibre's existing Windows speech engine utilities.

There is also another unknown. I'm under the impression that the existence of a speech engine's "speak out loud" function doesn't necessarily guarantee it also has a "speak to file (WAV)" function. This plugin requires the latter to create MP3s. Fortunately for us, the existing calibre/Windows speech engine does have both for Legacy voices. Who can say whether the speech engine used by Narrator to access offline Natural voices includes a "speak to file" function?

If I find out any more useful info I'll post in this thread ... but don't hold your breath!

P.S. I'm not sure what relevance your post #146 has?