Rockbox.org home
release
dev builds
extras
themes manual
wiki
device status forums
mailing lists
IRC bugs
patches
dev guide



Rockbox mail archive

Subject: Re: High kbps/Low kbps power consumption

Re: High kbps/Low kbps power consumption

From: Jon Drukman <jsd_at_cluttered.com>
Date: Thu, 05 Jun 2003 18:12:04 -0700

At 04:54 PM 6/5/2003, stevem_at_gator.net wrote:
>Now for my even dumber question: What is so bad about "transcoding" from a
>higher-encoded mp3 to a lower kbps, as opposed to ripping right from an
>original audio file?
>
>(I can imagine a few things, but you sound like you could explain
>definitively...)

i assume you're familiar with the concept of psycho-acoustic masking... the
very quick explanation is: your brain simply can't hear some frequencies
when other frequencies are present. do a web search for more info.

the idea of mp3 encoding is to look at all the frequencies in a slice of
time and decide which ones your brain won't hear. if you've already done
that once, then the second time you are taking a file that is missing
several frequencies, and deciding which ones to throw out. a good mp3
encoder might be able to make different assumptions about which things to
throw away based on the bitrate that it is shooting for. however you've
already thrown away a lot of stuff... so the encoder is starting with a
compromised sound spectrum.

also consider the dreaded "artifacts" that are added (quantization noise or
outright errors). the second encoder has no way to tell that those sounds
are unwanted. it is going to try to preserve them! in the best case it
will preserve them exactly as is. in the worse case it will add artifacts
on top of the artifacts. either way it will waste bits on noise that
should be used on signal.

transcoding: just say no.

-jsd-
Received on 2003-06-06

Page template was last modified "Tue Sep 7 00:00:02 2021" The Rockbox Crew -- Privacy Policy