Thanks for all these quick responses! I suspect that some of my problems mi=
ght be resolved just by using "autogain" in the display parameters when loo=
king at the spectrograms.
What I'm trying to do is extract individual calls that occur in a train of =
3-6 calls. I need to put these calls into separate wav files yet measure th=
eir original intercall intervals accurately in the original file. I don't w=
ant my interval measures to be affected by the differences in overall call =
volumes. If I have to do it all individually some calls might appear shorte=
r/ longer simply due to overall amplitude. Due to background noise I intend=
labeling each call by hand so as to not confuse Avisoft's automatic measur=
es.
I will go through all these answers & options and let you know how it worke=
d out. I'm using Avisoft in Bootcamp on my mac, so will try that Sample Man=
ager program too.
About CORRELATOR -- are the comparisons not affected somewhat by amplitude?=
I realize that the program examines the spectogram as a whole, but it feel=
s like calls should be at least within a small amplitude range instead of a=
ll over the place... Is this true, or can I compare soft and loud spectrogr=
ams (of the same call type) without fear? I want to look at certain individ=
ual call parameters but running the correlator in addition to this certainl=
y can't hurt.
Thanks for all the input so far!
--- In "Raimund" <> wrot=
e:
>
> Steve Pelikan wrote:
>
> > Could you perhaps compare spectrograms instead using some sort of corre=
lation coefficient? (This has worked pretty well for me in the past)
>
> Yep, simple spectrogram cross-correlations can work surprisingly well (so=
metimes much better than certain sophisticated patented digital signal proc=
essing algorithms...). In Avisoft, this kind of correlation is available in=
the separate CORRELATOR application or in SASLab Pro in the commands "Anal=
yze"/"Specials"/"Detect and classify waveform events..." or "Analyze"/"Spec=
ials"/"Scan for template spectrogram patterns..."
>
> Regards,
> Raimund
>
|