ARNOLD KLING
August 14, 2011
The Top Political Contributors
August 11, 2011
Gender and the New Commanding Heights
August 11, 2011
Jamie Galbraith Makes an Assumption
August 11, 2011
Macroeconometrics: The Science of Hubris
August 10, 2011
Real and Nominal Bond Yields
BRYAN CAPLAN
August 14, 2011
The Effect of Thumb Sucking on Income
August 12, 2011
The Voice of Cold, Hard Truth to All Would-Be Educators
August 12, 2011
Ability, Morality, and Prosperity: A Paper and a Report
August 11, 2011
The Theory of Time and Frittering
August 10, 2011
Male Variance and the Remnants of the Gender Gap
DAVID HENDERSON
August 9, 2011
Hayek in "Unbroken", Part Two
August 8, 2011
Hayek in "Unbroken"
August 5, 2011
James Bovard on the Peace Corps
August 4, 2011
Summers Way Off on FDR and 1941
August 3, 2011
The "Amazon" Tax
SHARE
There are lot's, but I use iTunes. If you encode in mono at 16 kbs, which I do for some talk radio programs, an hour's worth is around 7 MB.
You might try lame, if the licensing is not an issue. http://lame.sourceforge.net/using.html
Something like: lame -b 32 infile.wav outfile.mp3 will result in a quite small mp3. See
http://lame.sourceforge.net/USAGE for more info about various options. Size is of course traded for quality.
Hey. I got all eight right.
As usual, I find open-source based software hard to figure out. The best I could do with one LAME-based tool I downloaded was slightly worse than what I could do with something called Coffee Cup rip-and-burn. That is, the file was slightly larger with the open source stuff. But I have no idea how to optimize the open source stuff.
I think that if one could set the bitrate rate from 32 k down to 8 k it might help. But I don't know how to do that, and it's just a guess, anyway.
You should reduce the sampling rate to 12 kHz in order to get small file sizes. In doing so, you remove all frequencies above 6 kHz, which will not adversely affect the intelligibility of the recording; on the other hand, it will make high-frequency artifacts from the Olympus files less obvious. Using LAME, the following is pretty good for spoken recordings:
lame --resample 12 -b 24 -m m
That'll be a mono file at 24 kbps.
There are several good solutions for low bit-rate audio. A recent test of them was summarized at the following site: http://www.rjamorim.com/test/32kbps/results.html
Most of them are problematic at very low rates, not least because they would require your users to download new software to listen. If you assume they have iTunes, the AAC seems like a good bet.
The program dbpoweramp can make it easier to convert audio formats than some others. It can be used with the aforementioned LAME as well as many other codecs.
"Most of them are problematic at very low rates, not least because they would require your users to download new software to listen."
That's what's true of the dss format in which I make the recordings. You have to download the software, and on top of that the browser tries to read the format as an HTML page, so you have to save the file to disk and then listen to it. That's what I mean by being a user-hostile format.
I tried re-encoding one of your MP3 files as 8 kbps mono AAC using iTunes and I got a pretty decent quality at 1/4th the size. You might want to try that.