Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Best format for voice? (Read 3280 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Best format for voice?

Hi,
The discussions in this forum usually refer about audio quality for music or hard-to-encode samples, but have there been any tests or anything about encoding for voice or a capella songs? Since pure voice doesn't contain as much information as a normal song which usually has some three instruments + voice, it makes sense that these samples should be transparently encoded at much lower bitrates than normal music. Are there any codecs that are specially tuned for voice? Or which of the usual encoders provides the best compression for voice recordings?

~Dologan

Best format for voice?

Reply #1
Psytel AACENC has a -ihsc switch (Improved Human Speech Coding)

Might increase quality in voice-only tracks.

Best format for voice?

Reply #2
I've always found that MPC performs quite well for voice myself.  It actually has a feature specifically for encoding voice with higher fidelity in certain cases (vowels specifically, which can be difficult to handle properly for many psymodels) than some other codecs... called CVD (Clear Voice Detection).

Vorbis recently added some techniques which should add for somewhat similar functionality if I understand Monty correctly.  The last time I had a discussion with him about this he mentioned the new impulse frames that Vorbis supports (which drastically improved pre-echo handling in RC3 over earlier builds).  Supposedly these impulse frames also surprisingly occur quite a bit during speech...

PsyTEL AAC also has something which should create a similar effect to the above 2 mentioned codecs called IHSC (Improved Human Speech Coding).

I'd suggest to try any of those and use whichever happens to be your favorite codec for whatever purpose.

If you are needing high compression rates (low bitrates), I'd look specifically at Vorbis and AAC.  If you are looking at higher fidelity I'd probably lean a bit more towards MPC.  At any rate, they should all perform very well.

Best format for voice?

Reply #3
Quote
Originally posted by Dibrom
If you are needing high compression rates (low bitrates), I'd look specifically at Vorbis and AAC.  If you are looking at higher fidelity I'd probably lean a bit more towards MPC.  At any rate, they should all perform very well.


I don't agree. The best is ACELP.net codec (by Sipro Labs). You can find it in Windows Media Encoder

Bye, dB

Best format for voice?

Reply #4
That's for very low bitrate. There's lots of options down there too besides ACELP. Most of us probably use GSM everyday on our cellphones. It's available as a codec too which you can use with WME at 13kbps

Best format for voice?

Reply #5
Quote
Originally posted by dB
I don't agree. The best is ACELP.net codec (by Sipro Labs). You can find it in Windows Media Encoder


The original tone of the post suggested voice recordings at a fidelity level suitible for music encoding.  Voice encoding such as used in low bitrate transmissions (cell phones, etc) is something entirely different, that's why I suggested Vorbis or AAC.

Best format for voice?

Reply #6
Quote
Originally posted by Dibrom

The original tone of the post suggested voice recordings at a fidelity level suitible for music encoding.  Voice encoding such as used in low bitrate transmissions (cell phones, etc) is something entirely different, that's why I suggested Vorbis or AAC.


Sorry Dib, you're right! 


Bye, dB

Best format for voice?

Reply #7
Some interesting results from encodes that I did today using mpc and vorbis.  I ripped Dr. Dre - 2001 and used EAC's wav editor to separate a few end-of-track skits from the songs they were originally attached to.  These skits were almost all human voice, and the encoders handled them in markedly different manners.  I used mpc-standard, average bitrate for most music around 170-175, and the voice skits came out around 150.  This didn't surprise me.  For ogg, I used -q 4.99, which usually averages around 130 for music.  But two of the voice-skit files came out at 142 and 146, interestingly.  I've found that in general, what pushes bitrate up in ogg vbr doesn't necessarily push up bitrate in mpc vbr, and vice versa.
God kills a kitten every time you encode with CBR 320

 

Best format for voice?

Reply #8
if you wnat low bitrates with voice you canot use aac/mpc or whatever hi-fi codec

these codec are tuned for entires sound spectrum and to give hifi quality sound at high bitrates


you need a voice tuned codec.

windows media encoder comes with acelpt.net
and thsi codec is tuned for voice. gives great VOICE ONLY quality at 16kbvits and lower (16kbits is highest bitrats this codec goes)

howevr this is NOT hifi but it gives cler and "listen-able" voice


i used it for backup of story tapes
Sven Bent - Denmark

 
SimplePortal 1.0.0 RC1 © 2008-2021