Monday, June 11, 2018

Graphical audio spectrum visualizer: Trick Spectrum from Microphone, MP3 players, Win10, in Visual Basic 6.0


Here we have a superb app made by Krivous Anatoly Anatolevich. Here is his description on VBForums: The sound is analyzed through a standard recording device, i.e. you can select the microphone and view its spectrum, or you can select stereo mixer and view a spectrum of a playback sound.
Captures via my Microphone

This visualizer allows to adjust the number of displayed octaves, transparency of background and amplification.
You can also load a palette from an external PNG file with 32ARGB format. It also supports the following effects: "blur" and "burning". You can view a spectrum of a signal represented in the two modes: arcs (rings) and sectors (pies). If you use the ring view an octave is mapped to radial coordinate and a semitone to angle. The separated harmonics are placed along the same line; color represents an intensity. The sectors view maps the amount of signal to the radial coordinate, the frequency in octaves to the color, the frequency in semitones to the angular coordinate.
This idea was suggested to me by Vladislav Petrovky (aka Hacker). His idea was a little different.

Initially it creates the buffers for sound and buffer bitmaps. Further it starts the sound capture process and waits when a buffer will be filled. When a buffer has been filled it begins processing. Firstly it performs the Fast Fourier Transform in order to transform a signal to the frequency domain form. Before performing it applies the Hamming window in order to reduce distortions because a signal has discontinuity at the edges of a buffer. When a signal has been translated to the frequency domain the buffer contains complex value that represent the vectors. The module (length) of a vector implies the energy of signal in that frequency and the argument (angle) implies phase of a harmonic in that frequency.


Screenshot of the menu
We need the energy of frequency although the phase information allows to determine the frequency more accurately considering the phase difference. I don't use phase information in this project. The drawing method is different for each appearance mode. In order to boost the work it uses the precalculated coordinates named MapData. This array contains the angles of arcs and sectors for the current appearance mode. When coordinates has been calculated it calculates the amount of frequency for each FFT bin figuring out the length of a vector. This value is uses as the index in the color palette after converting the value to a range from 0 to 255. Further GDI+ draws the necessary primitives depending on the appearance mode. Note that all drawing occur onto the buffer bitmap not on window. I specially have not mentioned about the Release procedure that animates the background. This procedure applies an effect to the buffer bitmap before signal processing. It uses the Fade property that determines the speed of the disappearance of previous drawing bitmap. It just decrease the alpha value of the entire bitmap. When you use an effect it also works with the bits of the buffer bitmap and decreases the alpha value. For instance, if the blur effect has been selected it averages the near pixels (analog of low-pass filtering) then it decreases the alpha value for all pixels depending on Fade property. Eventually it draws buffer bitmap onto the main window. Thus it draws the energy of the spectrum of signal in the polar coordinates. It can be used as the start point for the notes or chord recognition.


Download from VBForums

Download from me



Here is a video of the app:










No comments:

Post a Comment