Scrolling Spectrogram visualization

Hi,

I'm looking for a scrolling spectrogram visualization, or will try to develop one if I can't find one.

Specifically, it would have frequency (kHZ) in the vertical dimension, and time in the horizontal. So if you had an input (or file) of say, a flute (relaively few harmonics, just the fundamental frequency) playing a scale, you would see a series of short horizontal lines scroll accross in a successively higher position in the window.

This may sound very basic, and not an appropriate application for WinAmp, but what I'm looking for is really a traditional Spectrogram on Steroids! An entertaining 3D representation of this scrolling time/frequency doamain pitch-specific signal. Mind you, this is different than a spectrum analyzer representation which an equalizer-style view. (I've attached a jpg to show the type of output).

If there isn't already a preset to do this, I'd be interested in creating one. Could someone tell me what the best development environment would be for this? AVS, MilkDrop, Other?

Thanks for your help,

SteveW

Well, Render//Timescope is a good way to start.

I'm pretty sure that it's basically getspec(i,0,0) in an SSC, use Color Map accordingly.

As for 3D... Well, good luck! Maybe you can modify UnConeD's Turing Carpet.

This has been asked for before at least a couple of times, try searching the forums.

if you learn avs it will allow you to easily perform what you are asking. Try looking at some of the tutorials.

If you want someone else to make it for you its unlikely anyone will be willing, or you'll get some noobs effort which wont be very good.

Thanks, guys. I've viewed the various threads that have to do with 'Spectrogram' and found them somewhat useful. I'll go through the tutorials now, and maybe mess with creating some presets to see if this might work for me.

suggestion

this may not be 3d. but it works.
try learning how to make the scope scroll untill it hits the edge and repeats.

for the codes them selves try this.

y=2*i-1;
x=t; use when you learn how to make a repeating code.
red=getspec(i,0,0);
green=red;
blue=red;
this will make a spectrum image like you want. suggested presets are The Earthquaker rerelease 1 pack. Earthquaker does impressive spectrum based presets so try him.

Be warned though, if you want a GOOD spectrogram, no Winamp plugin will cut it. There is a limitation on the sample rate (576 samples per frame) recieved by plugins so that there isn't enough quality of data to 'rip' audio through a plugin (I think this is the reason). This also means that the spectrum has low resolution, in addition I think that the scaling on the AVS spectrum is linear rather than logarithmic.. I can't remember exactly.. but there is something fishy about it.

Good luck with that anyway...

There is one thing

There is a program i use thats of professional quality. its called Ahead Nero Wave Editor. the program gives you three ways to view your sound data. Wave form or Osciliscope in terms.
Spectrum like what you want. Or the personal fave of mine is the Wavelet form which displays a more acurate form of the sound then the spectrum. each color used in the wavelet tells you interesting things. the brighter the orange the higher the volume. the redder the more amplitude. it gives you dynamic post processing and sound reduction(not very useful),Transpose,8 channel Equalizer,delay,reverb,echo, and time correction. the program usualy comes with Nero burning rom. it may come with this program at http://www.oldversion.com they have old versions of nero that may have the program i described.

Just for the record, you can do wavelet transforms in AVS

A gift

I've been working on this.
enjoy.

Cool vis, Nanakiwurkz. This could provide some very useful example code for me to work with.

Thanks!

avs or winamp can only portray data in 576 samples per second

Its not per second its per frame

There is a limitation on the sample rate (576 samples per frame) recieved by plugins so that there isn't enough quality of data to 'rip' audio through a plugin

This isnt strictly speaking true, its entirely dependant upon the framerame you will get though the plugin. The sample rate of most audio tracks (CD,MP3,etc..) is 44KHz (human hearing range is up to approx 20KHz so its roughly twice that, nyquist etc...). So you have (typically) 441,000 sample per second, 576 samples per frame, meaning you need approximately 766 frames per second to avoid missing any samples. Not really feasable in AVS, but possible for a.n.other plugin. Whether or not it would be practical to reconstruct the original signal in realtime from arbitrary 576 sample segements is of course another issue...

Originally posted by PAK-9
Its not per second its per frame

who said it was per second. i can't see that anywhere above. 🙁

Its in the comment of Nanakiwurkz' preset