good mp3 setting for speech?

Clocker

Golden Member
Sep 17, 2000
1,353
0
76
I have some lectures i want to convert from wave to mp3s. what is a good setting. normally i use alt-preset-standard. but that is for music. I want to maximize space so I am thinking of using 32kps to 64kps variable. Is this too low? What does everyone else think.

 

mugs

Lifer
Apr 29, 2003
48,920
46
91
That's not too low. I've seen speech as low as 24 kbps I think. With that kind of content, sound quality is not a huge issue.
 

Chaotic42

Lifer
Jun 15, 2001
34,958
2,110
126
I just tried convering a 320kbps speech MP3 to 24kbps. The drop in quality was obvious, but it was still very understandable. It would be fine for lectures.
 

bunnyfubbles

Lifer
Sep 3, 2001
12,248
3
0
There are other codecs you could use instead of mp3 that are much better for things such as speech at very low bit rates

But as far as 32 to 64kbps VBR that's about what I use for compressing CD audio books (pretty much exclusively speech) - much nicer being able to fit a couple dozen discs into one mp3 disc.
 

McCarthy

Platinum Member
Oct 9, 1999
2,567
0
76
Depends how clean the source is too. For example, when I record a radio show off an internet stream - even thought it's already compressed, it has very little noise. Re-encoding to 24 will work fine. That's "this is ok to listen to, nothing more" quality. To get that same listening experience when recording a local AM radio show off the air I choose 40kbps.

So really it comes down to what sounds good to your ear. Try a encoding a few minutes at each bitrate and listen - whatever is still listenable without having to concentrate to make out what's being said is what you should go with.

If you plan to keep these for decades to listen to again, go up another step or two as by then you'll be completely out of the habit of listening to low quality audio and your ears won't be any better than they are now.
 

biggestmuff

Diamond Member
Mar 20, 2001
8,201
2
0
Uh, why don't you convert a speech or a one minute section of a speech at various bitrates using two or three different codecs, then let your ears tell you what's the best?