Skip to main content

Poll

What audio format do you usually use for speech contents?

AMR-NB
[ 0 ] (0%)
AMR-WB
[ 0 ] (0%)
AMR-WB+
[ 0 ] (0%)
Speex
[ 1 ] (4.8%)
WMA Voice
[ 0 ] (0%)
Opus
[ 12 ] (57.1%)
EVS
[ 0 ] (0%)
Codec2
[ 0 ] (0%)
MP3
[ 1 ] (4.8%)
Vorbis
[ 0 ] (0%)
AAC
[ 5 ] (23.8%)
Other lossy compressed formats
[ 0 ] (0%)
Lossless compressed formats
[ 2 ] (9.5%)
Uncompressed formats
[ 0 ] (0%)

Total Members Voted: 21

Voting closed: 2019-03-12 12:58:12

Topic: What audio format do you usually use for speech contents? (Read 5050 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

What audio format do you usually use for speech contents?

What audio format do you usually use for speech contents?

Re: What audio format do you usually use for speech contents?

Reply #1
Is there really any reason to use anything other than Opus, unless you are locked in by legacy devices that only support a limited set of formats, or your connection has extremely restricted bandwidth? That goes for both music and speech content.

Re: What audio format do you usually use for speech contents?

Reply #2
There is. Using it makes no sense until this is fixed.

Re: What audio format do you usually use for speech contents?

Reply #3
It's not as bad as to make it "no sense to use". Especially with speech, one rarely needs to have track boundaries right in the middle of a phrase. Even with music, the glitch on track transition is not always noticeable.

Re: What audio format do you usually use for speech contents?

Reply #4
When I encode audiobooks from CD, I use HE-AAC at CVBR 32kbps.

Re: What audio format do you usually use for speech contents?

Reply #5
Speech + background music: WMA standard, 64 kbps

Re: What audio format do you usually use for speech contents?

Reply #6
If we are strictly talking speech... Opus v1.3.1 @ 13kbps (or Opus v1.2 @ 14kbps) is what I consider THE minimum I would personally use (although if one likes to play it a bit safer you could bump up the bit rate a bit). I can notice the difference from a higher quality source file but it's not that significant of a hit (say going from a 128kbps MP3 and thereabouts to Opus @ 13kbps) and saves you a lot of storage space and it is speech at the end of the day and not music, so I don't mind sacrificing some speech quality to save storage space.

so while I have not played with it too much, I can't imaging needing more than around 32-48kbps tops for speech only on Opus, even for those who prefer to have a bit higher quality speech but are not too concerned with total or near transparency.
For music (especially on-the-go)...
-I suggest Opus @ 96kbps (or... 64kbps minimum, 128kbps maximum). *preferred choice*
-I suggest AAC(Apple) @ 96kbps (q45 TVBR) or 128kbps (q64 TVBR). *secondary choice*
-I use Foobar2000 (/w Encoders Pack etc) to convert FLAC to Opus/AAC(Apple).

Re: What audio format do you usually use for speech contents?

Reply #7
There is. Using it makes no sense until this is fixed.

That will never be fixed. Monty and the other prima donnas are already onto something new. I use and I will keep using Vorbis, which is really good in several aspects, but I will stay away any other work of Monty and Xiph. Only one other successful project they have: FLAC. But they took it over long after it became a success. So it is not their achievement. Every Vorbis user including myself uses Aotuv Vorbis. It seems to be proved and tested over here for years, but they refused to incorporate that hard work. No reasons. Nothing. And now one has to know where to turn to have the source code. It is all disappearing from the internet. 5 more years and it is all gone.

It is a shame how Xiph codecs are managed.

Maybe it is part of the push towards Opus. I did not compared them, but it is possible that Aotuv Vorbis would be at minimum on par with Opus.

Re: What audio format do you usually use for speech contents?

Reply #8
Speech: 80kbps ABR Lame

Got locked out on a password i didn't remember. :/

Re: What audio format do you usually use for speech contents?

Reply #9
I've done a few hundred audiobooks from cassette and CD to LAME  with this setting
-V 8 -m m --vbr-new --resample 22 --lowpass 11 --noreplaygain
The result is easy to  understand, easy to listen to, the files are relatively small, and they play pretty much anywhere.

Being VBR, it sometimes produces a bit rate as high as 256kbps, according to foobar2000 or Razorlame, where I do most of the encoding. Possibly a close ABX could detect some differences from the original or relative to "better" encodings, but I could not care less as there are no playback or listening problems.

 
SimplePortal 1.0.0 RC1 © 2008-2019