Text To Speech Goes In As A Tech Preview For Qt 5.8

[phoronix]

Phoronix: Text To Speech Goes In As A Tech Preview For Qt 5.8

With Qt 5.8 that's due to be released next week there is the new Qt Speech as a "tech preview" of text-to-speech for this tool-kit...

http://www.phoronix.com/scan.php?page=news_item&px=Qt-5.8-Text-To-Speech

[juno]

Can anyone recommend a tts system with high quality voices that sound more natural? Several apps have tts support (e.g. Firefox) but it sound so bad it is unusable.

[hansg]

On my Amiga, back in 1985, I booted from a floppy, opened a CLI ("Command Line Interface", as we called it in those days), and typed

> say "Hello, world!"

and it would say those words out loud. I'm glad to see that QT is slowly closing the gap.

Oh, and why is there no example of how it sounds? Surely that is one of the most important attributes of a TTS system.

[juno]

Oh, and why is there no example of how it sounds? Surely that is one of the most important attributes of a TTS system.

They are not implementing tts themselves, it depends on the native tts solutions of the system. So speech dispatcher (which sounds horrible to me) or Microsoft's (they used to call him "Sam", maybe it's now Cortana?).

[Chewi]

On my Amiga, back in 1985, I booted from a floppy, opened a CLI ("Command Line Interface", as we called it in those days), and typed

> say "Hello, world!"

> say "pooooooooooooooooooooooooo"

Those were the days.

At work, we use CereProc but it's not free.

[linuxgeex]

On my Amiga, back in 1985, I booted from a floppy, opened a CLI ("Command Line Interface", as we called it in those days), and typed

> say "Hello, world!"

and it would say those words out loud. I'm glad to see that QT is slowly closing the gap.

Oh, and why is there no example of how it sounds? Surely that is one of the most important attributes of a TTS system.

And on an Amiga, if you yanked a removable media while an IO operation was underway it would pop up a notice politely demanding that you re-insert it so that it could finish the operation. How many of you have corrupted USB media devices (probably unknowingly due to an intermittent failure) on Linux systems?

Will the Linux Kernel, or any desktop environment on Linux, perform reliable IO to a removable media? Nope. Is it important? Apparently not. Heck, we're only just starting to get somewhat reliable Audio on our desktops lol.

[linuxgeex]

BTW for those of you who are interested to know what the QT TTS speech quality is like, here's what Michael and the original QT Dev felt was meaningless for you to know:

"Use say() to start synthesizing text. It is possible to specify the language with setLocale(). To select between the available voices use setVoice(). The languages and voices depend on the available synthesizers on each platform. On Linux by default speech-dispatcher is used."

[Chewi]

And on an Amiga, if you yanked a removable media while an IO operation was underway it would pop up a notice politely demanding that you re-insert it so that it could finish the operation. How many of you have corrupted USB media devices (probably unknowingly due to an intermittent failure) on Linux systems?

Will the Linux Kernel, or any desktop environment on Linux, perform reliable IO to a removable media? Nope. Is it important? Apparently not.

On the flip side, I once triggered a bug in DiskMaster2 that started rapidly deleting all the files from my hard drive. I panicked and rather than just hitting Escape, I hit Ctrl+A+A (Amiga's Ctrl+Alt+Del equivalent). Yep. I'd corrupted the entire partition.

[DanL]

BTW for those of you who are interested to know what the QT TTS speech quality is like, here's what Michael and the original QT Dev felt was meaningless for you to know...

It's an abstraction layer, and the article clearly states that. It doesn't have its own speech quality. (And it's Qt, not QT.)

[cfeck]

Can anyone recommend a tts system with high quality voices that sound more natural? Several apps have tts support (e.g. Firefox) but it sound so bad it is unusable.

The mbrola voices are acceptable (for stuff that is open). Of course, anyone developing better speech synthesizers, keep them closed.