Subject: Re: [ecasound] the -ev effect thing
From: Kai Vehmanen (k_AT_eca.cx)
Date: Mon Apr 28 2003 - 21:54:59 EEST
On Sun, 27 Apr 2003, Jeremy Hall wrote:
> of all the samples for this channel, x% of them peek at 0, 2, 4, 8, 16,
> 32, 64 etc
To be precise, x% of samples have value in the range (1-2dB, 2-4dB, etc).
You can see the whole algorithm in:
libecasound/audiofx_analysis.cpp:
void EFFECT_VOLUME_BUCKETS::process(void)
> some even shoed -infin which shouldn't be, unless it is considering the
> noise floor infin.
'inf' is samples with value abs(sample) <= 0.001f.
> Now what do the pos and neg mean? and I'm assuming 0db is unity and -2 etc
> is reduction by say 2db
As the counters are for individual samples, there are separate value range
buckets for positive and negative sample values.
> If I am understanding this output correctly, it is allowing me to "see"
> the graph curve.
Yup, not very academic, but the output is quite useful to check what kind
of signal is going through.
-- http://www.eca.cx Audio software for Linux!
This archive was generated by hypermail 2b28 : Mon Apr 28 2003 - 21:57:13 EEST