Spectrogram

Revision as of 17:28, 19 December 2018

Description

Linear spectrogram for AviSynth 2.6.

Requirements

• AviSynth 2.6.0 or greater
• Supported color formats: all colorspaces are supported but output clip will always be Y8.

• Microsoft Visual C++ 2017 Redistributable Package (vc_redist.x86.exe or vc_redist.x64.exe)

• FFTW 3.3.5 (fftw-3.3.5-dll32.zip or fftw-3.3.5-dll64.zip)

*** libfftw3f-3.dll needs to be in the search path (C:\windows\system32 or C:\Windows\SysWOW64)

*** be aware to have your whole processing chain in matching bits: either 32 bit (=x86) or 64 bit (x64).

*** Spectrogram will not run or load without it.

Syntax and Parameters

Spectrogram (clip, int width, int height, bool "transpose")

clip   =

Input clip; audio must be float so use ConvertAudioToFloat() if needed. Video can be any colorspace but the output will always be Y8.

Frame rate of the spectrogram will always default to the frame rate of the input clip. If the input clip does not contain video then the frame rate will default to 24FPS.

int   =

Width of the spectrogram.

int   =

Height of the spectrogram. For best performance it's recommended that the height is a value of 2^n.

bool  transpose = false

When set to true the spectrogram will be rendered vertically; recommended for larger dimensions.

Spectrogram(x, y, transpose=true).TurnLeft() is identical to Spectrogram(x, y, transpose=false) but faster. For better performance use the optimized FTurnLeft() function from the FTurn plugin.

• Note: the first 3 parameters are unnamed and do not have a default so they must be specified.

Examples

The following script creates a scrolling spectrogram from just an audio file. The length is automatically adjusted based on the duration of the audio and the desired frame rate. The output clip will be Y8, 512x256, and 59.94 FPS, of course this is all adjustable.

WavSource("sample.wav") # for other types of files you can use FFAudioSource or LWLibavAudioSource or any other suitable audio source filter
audio = last

fps = 60000
den = 1001
length = Round((AudioDuration(audio)*fps)/den)
BlankClip(length, width=512, height=256, fps=fps, fps_denominator=den, pixel_type="Y8")
AudioDub(last, audio)
ConvertAudioToFloat() # if needed

Spectrogram(512, 256, transpose=true)
TurnLeft()

Changelog

Version      Date            Changes

r7           2018/12/19      - order channels like Waveform does
r6           2018/12/13      - add x64 bit build and a VS2017 solution to the archive
r5           2015/07/04      - increase cache sizes to accommodate larger dimensions
                             - add "transpose" parameter for better performance

r4           2015/03/11      - fix a bug; the cache was actually completely broken

r3           2015/03/10      - fix a bug that caused a crash
                             - it's now ~40-50% faster; more profiling and fixing a hot spot

r2           2015/03/09      - implement cache; increase speed by profiling

r1           2015/03/08      - initial release

Archived Downloads

Source code is included.

Back to External Filters ←