Re: [ecasound] lossless conversion broken?

From: Dan Muresan <danmbox@email-addr-hidden>
Date: Sat Oct 30 2010 - 21:22:01 EEST

Hi Julien,

> áNo it shouldn't be reason for distortion, as float can be much more precise
> than int. So only the other way round float -> int can introduce distortion

Well, a 32-bit float stores a 23+1 bit mantissa and an 8-bit exponent,
so the resolution is *at most* the same as for 24-bit PCM. In any
case, if there's an internal int->float conversion, there must also be
a float->int conversion when writing the output file.

Theory aside, as my wav -> wav experiment showed there can be 1-bit
round-off errors, and for a low-level signal those would count more
than for a full-range signal. I like ecasound's versatility and
power-features (double-buffering, realtime priority), but I prefer to
"stay lossless" as far as I can down the processing chain.

-- Dan

Nokia and AT&T present the 2010 Calling All Innovators-North America contest
Create new apps & games for the Nokia N8 for consumers in U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store
Ecasound-list mailing list
Received on Sun Oct 31 00:15:02 2010

This archive was generated by hypermail 2.1.8 : Sun Oct 31 2010 - 00:15:02 EEST