Quote from: Defender on 2024-06-21 14:46:02

Now the dust seems a bit settled with your foo_uie_text component,
can I expect a bugfix for foo_vis_spectrum_analyzer in regard to windows scaling (125/150/175/225/250/350) peakmeter issues?

https://hydrogenaud.io/index.php/topic,125031.msg1045171.html#msg1045171

Thx

Eventually. It may surprise you but I do have a life besides foobar2000 development.

Quote from: pqyt on 2024-06-21 15:07:12

Quote from: Defender on 2024-06-21 14:46:02

Now the dust seems a bit settled with your foo_uie_text component,
can I expect a bugfix for foo_vis_spectrum_analyzer in regard to windows scaling (125/150/175/225/250/350) peakmeter issues?

https://hydrogenaud.io/index.php/topic,125031.msg1045171.html#msg1045171

Thx

Eventually. It may surprise you but I do have a life besides foobar2000 development.

And as a developer I know that when you are "out" of the code for a period of time fixing old bugs becomes more difficult.

International Programmers’ Day celebrates the positive changes that programmers make to improve our everyday lives. It is usually held on January 7, but is also popularly celebrated on September 13, or on September 12 in leap years.

Is there a way to choose where in the DSP chain the analysis happens? I would love to have the analysis happen before application  of any DSPs in the chain as a configurable behavior.

And visualization would have no meaning if it worked on data prior to DSPs, it could be entirely invalid in regard to what is actually heard.

DSP chain can do anything to the signal.

By going to Common in the Artwork section and activating the "Show artwork on background" flag, if you do not have the panel active the artwork will not appear.

for (let i = 0; i < averageSpectrum.length; i++) {
  averageSpectrum[i] = (data[i] + cumulativeIdx*(isFinite(averageSpectrum[i]) ? averageSpectrum[i] : 0))/(cumulativeIdx+1)
}
cumulativeIdx++ // being incremented by one every time you calculate the infinite average spectrum

Suggestion: Infinite average spectrum/cumulative average graph
[attach type=image]31580[/attach]

Like the RMS spectrum in foo_enhanced_spectrum_analyzer, but with infinite averaging that eventually converge to the particular curve when you let the analyzer running for long enough with an equation:

Code: [Select]

for (let i = 0; i < averageSpectrum.length; i++) {
  averageSpectrum[i] = (data[i] + cumulativeIdx*(isFinite(averageSpectrum[i]) ? averageSpectrum[i] : 0))/(cumulativeIdx+1)
}
cumulativeIdx++ // being incremented by one every time you calculate the infinite average spectrum

Of course, since infinite averaging never decays, this infinite average spectrum should be reset on seek or when playing new track(s) alongside the main spectrum and peaks like foo_enhanced_spectrum_analyzer

Please no. I just want a simple long overdue bugfix instead if new features.

Strange:  there is a stream Nordic Chillout Radio (http://176.9.86.211:8023/stream) that according to Foobar v.2.2 x64 Properties is being sent at a sample rate of 7350 Hz--I find that hard to believe unless Foobar is resampling it internally (I did not have that sample rate in my "resample only" list in my active DSP SoX resampler), because the stream plays normally, although does not send metadata.  The stream's "icy-sr" header does show "7350" though that could be bogus. 

Anyway, what this has to do with foo_vis_spectrum_analyzer is that no spectrum is shown with this stream, although the Peak Meter component is active.  Why would that be??

This is one of those things that the sending server will probably fix tomorrow, but what would cause the no-spectrum?

I'll see if I can reproduce it in Debug mode.

I've noticed that the component looks odd when playing .dsf files. The movement is very slow compared with other file types. If it dances for those, then the motion is more like the homecoming queen waving from the parade float. The built-in components don't look nearly so sluggish.

I have foo_input_sacd and foo_dsd_processor installed to play them with.

I've noticed that the component looks odd when playing .dsf files. The movement is very slow compared with other file types. If it dances for those, then the motion is more like the homecoming queen waving from the parade float. The built-in components don't look nearly so sluggish.

I have foo_input_sacd and foo_dsd_processor installed to play them with.

I also get the slow motion effect when playing DSD files and have DSD+PCM output selected in the SACD plugin (V. 1.5.11), interestingly all FB2K vsualizations run at the correct speed.