| From Jurgen Konings
| Sun, 26 Aug 2007 01:05:40 +0200
| Subject: [Audacity-help] program startup parameter
| Ok i changed from 32bit float to 16 bit PCM. But there is also 24bit.
| What is the difference in these qualities ?
| Is 16 bit much worse than 32 bit or is the difference unhearable ?
As I said, 16 bit quality of recording may well not be perceptually inferior
to 32 bit quality for many listeners, even if you compared recordings by
simply playing them back in Audacity. This would be especially true if you
did not edit them. Bear in mind that to play the recorded audio in computer
media players, you are almost certainly going to have to either truncate it
to 16 bit anyway for export as an uncompressed WAV or AIFF file, or
compress it in other ways (e.g. by exporting as an MP3). Not many computer
media players can play 24 bit WAV or AIFF files let alone 32 bit, and if you are
burning the WAV/AIFF files to audio CD the files are required to be 16 bit by
the format specification.
Even if you have to export a 16 bit file, there is a theoretical advantage in
recording at 32 bit float, then dithering down to 16 bit ("High quality dither"
on the Quality tab of Preferences), rather than avoiding the downsampling
altogether by recording in 16 bit. Whether the user could genuinely hear the
difference in a double blind test on two exported 16 bit files from a 32 bit
and 16 bit recording respectively is another matter.
Higher sample sizes (e.g. 24 bit against 16 bit) allow more bits in each
sample to be captured and thus allow more dynamic range to be captured -
see a brief explanation here:
http://audacity.sourceforge.net/manual-1.2/tutorial_basics_1.html
32 bit quality in Audacity is especially beneficial because it is floating point
data (hence "float" in "32 bit float"). Normalised floating point values are
quicker and easier to process on computers than fixed integer values and
allow greater dynamic range to retained even after editing. This is because
intermediate signals during audio processing can have very variable values.
If they all get truncated to a fixed integer format, you can't boost them back
up to full scale without losing resolution (i.e. without the data becoming less
representative of the original than it was before). With floating point,
rounding errors during intermediate processing are negligible.
The (theoretically audible) advantage of this is that 32-bit floating point
format has no noise floor. For example, with fixed integer data, applying a
compressor effect to lower the peaks by 9 dB and separately amplifying back
up would cost 9dB (or more than 2 bits) of signal to noise ratio (SNR). If
done with floating point data, the SNR of the peaks remains as good as
before (except that the quiet passages are 9dB louder and so 9dB noisier
due to the noise they had in the first place).
If you are concerned, do a test recording of a short passage in 32 bit float,
make some edits involving scaling the volume (e.g. compressor and FFT Filter)
and export to your intended format e.g. 16 bit WAV or MP3. Record the same
passage in 16 bit quality, make exactly the same edits with the same
parameters and export to the same format, and compare the two exported
audio files.
Gale Andrews
.
Outbound message virus free.
Tested on: 8/26/2007 3:06:02 AM
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>
http://get.splunk.com/--
Audacity-help mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-help
