[music-dsp] Decible Question

Jon Watte hplus at mindcontrol.org
Mon Apr 1 21:27:02 EST 2002


Many things are measured in decibels. Basically, 20 decibel of change
mean a factor of 10 change. Frequently, voltage for signals is measured
in decibels; often the reference (say, 1V) is said to be "0 dB" and
most signals will be measured in negative dBs, as in "how far off from
nominal". Amplification (gain) is also measured in dB, as relative
change. A signal (voltage) at -20 dB (with a voltage of 0.1 V, say)
might have 20 dB of gain applied to it, which will make it end up at
0 dB, or with a voltage of 1 V in this case. In this case, you CAN look
at a single sample and tell what voltage (level) it is in dB.

THEN there's acoustical decibels, as measured by the occupational
safety people and speaker salesmen. These basically translate to sound
loudness relative to some norm, and come in a bunch of different
flavours (dBA, dBB, dBC). They also come in a bunch of different time
weightings; ideally, which weighting and timing window you're using
should be specified with the measurement, but that's often neglected
(and there are standards, although there are several to choose from :-)

Cheers,

			/ h+

> -----Original Message-----
> From: owner-music-dsp at shoko.calarts.edu
> [mailto:owner-music-dsp at shoko.calarts.edu]On Behalf Of O'Donnell Group
> of Jacksonville
> Sent: Monday, April 01, 2002 4:01 PM
> To: music-dsp at shoko.calarts.edu
> Subject: [music-dsp] Decible Question
>
>
> ok, so here's another topic I have not found to make itself easy to
> understand. Decibles. A measure of loudness, as defined in
> webster. So, how
> do i write an algorithm to compute it?
>
> I've thought about this, and it puzzles me. I think its probably
> impossible
> to look at one sample (16-bit numbers in c++) and determine the decible
> level, otherwise it would be easy to understand. So the method
> must involve
> taking a set of data and computing the decible level at a given
> point. Since
> dsp effects are designed to primarily operate on real-time signals where
> there is no knowledge of future data, the set must only include
> the current
> piece of data, and past sets of data.
>
> Anyway, if anyone has any resources, or thoughts. I have found this topic
> particularly under-covered in the literature I've been perusing.
>
> Thanks,
> -Philip
>
> _________________________________________________________________
> Chat with friends online, try MSN Messenger: http://messenger.msn.com
>
>
> dupswapdrop -- the music-dsp mailing list and website: subscription info,
> FAQ, source code archive, list archive, book reviews, dsp links
> http://shoko.calarts.edu/musicdsp/
>


dupswapdrop -- the music-dsp mailing list and website: subscription info,
FAQ, source code archive, list archive, book reviews, dsp links
http://shoko.calarts.edu/musicdsp/




More information about the music-dsp mailing list