Author Topic: A test piece using text to speech  (Read 468 times)

hutch--

  • Administrator
  • Member
  • ******
  • Posts: 8366
  • Mnemonic Driven API Grinder
    • The MASM32 SDK
A test piece using text to speech
« on: February 27, 2021, 04:22:15 PM »
If I could impose on anyone to try this out. Its a short test video using text to speech software and I wanted to see if anyone had a problem with it. My only complaint is it requires larger files as speech is a lot slower than written text. This one uses an American voice but I have some funny ones available, a very good Indian female voice and a passably good female Australian voice.

I have removed the link so I can delete the test video on the Youtube channel.

« Last Edit: March 02, 2021, 10:24:25 AM by hutch-- »
hutch at movsd dot com
http://www.masm32.com    :biggrin:  :skrewy:

Gunther

  • Member
  • *****
  • Posts: 3722
  • Forgive your enemies, but never forget their names
Re: A test piece using text to speech
« Reply #1 on: February 27, 2021, 06:13:39 PM »
Steve,

it sounds good. Don't forget to try the Indian female and the Australian female.

Gunther
Get your facts first, and then you can distort them.

Biterider

  • Member
  • ****
  • Posts: 705
  • ObjAsm Developer
    • ObjAsm
Re: A test piece using text to speech
« Reply #2 on: February 27, 2021, 06:57:34 PM »
Sounds very clear


Biterider

hutch--

  • Administrator
  • Member
  • ******
  • Posts: 8366
  • Mnemonic Driven API Grinder
    • The MASM32 SDK
Re: A test piece using text to speech
« Reply #3 on: February 27, 2021, 08:08:21 PM »
Thanks guys, I bought a recent TTS package called Talkia that uses a number of Google and Amazon voices and they are a lot better that the best of the SAPI 5 voices I played with using a TTS app called Balabolka. Apart from the better voices it has an important capacity, the ability to add times pauses to the sound track so you could set up phrases and paragraphs properly.
hutch at movsd dot com
http://www.masm32.com    :biggrin:  :skrewy:

daydreamer

  • Member
  • *****
  • Posts: 1644
  • building nextdoor
Re: A test piece using text to speech
« Reply #4 on: February 27, 2021, 08:55:39 PM »
It is a nice clear voice
Does it,to still be clear voice, you need to use less compression ratio, compared to bg music?
SIMD fan and macro fan
why assembly is fastest is because its switch has no (brakes) breaks
:P
only in 16bit assembly you can get away with "Only words" :P

hutch--

  • Administrator
  • Member
  • ******
  • Posts: 8366
  • Mnemonic Driven API Grinder
    • The MASM32 SDK
Re: A test piece using text to speech
« Reply #5 on: February 28, 2021, 03:56:12 AM »
Sound is not particularly demanding like video but if you want to compress speech, Audacity works fine, just set the bitrate to 8k and the size will drop to about 1/16th of the size and still sound the same. Music cannot be compressed the same way as it has a lot more data in it. The action to get the video size down is the bitrate and on these mainly text videos, I have it down to 512k.
hutch at movsd dot com
http://www.masm32.com    :biggrin:  :skrewy: