saveMatrixAsMtx

csoundengine.csoundlib.saveMatrixAsMtx(outfile, data, metadata=None, encoding='float32', title='', sr=44100)

Save data in wav format using the mtx extension

This is not a real output. It is used to transfer the data in binary form to be read by another program. To distinguish this from a normal wav file an extension .mtx is recommended. Data is saved always flat, and a header with the shape of data is included before the data.

Header Format:

headerlength, numRows, numColumns, ...

The description of each metadata value is included as wav metadata at the comment key with the format:

"headerSize: xx, numRows: xx, numColumns: xx, columns: 'headerSize numRows numColumns ...'"

This metadata can be retrieved in csound via:

itabnum ftgen 0, 0, 0, -1, "sndfile.mtx", 0, 0, 1
Scomment = filereadmeta("sndfile.mtx", "comment")
imeta = dict_loadstr(Scomment)
ScolumnNames = dict_get(imeta, "columns")
idatastart = tab_i(0, itabnum)
inumrows = dict_get(imeta, "numRows")
; inumrows can also be retrieved by reading the table at index 1
; inumrows = tab_i(1, itabnum)
inumcols = tab_i(2, itabnum)
; The data at (krow, kcol) can be read via
kvalue = tab(idatastart + krow*inumcols + kcol, itabnum)

; Alternatively an array can be created as a view:
kArr[] memview itabnum, idatastart
reshapearray kArr, inumrows, inumcols
kvalue = kArr[krow][kcol]

Parameters:

• outfile (str) – The path where the data is written to

• data (numpy array) – a numpy array of shape (numcols, numsamples). A 2D matrix representing a series of streams sampled at a regular period (dt)

• metadata (dict[str, str | float] | None) – Any float values here are included in the header, and the description of this data is included as metadata in the wav file

• encoding – the data can be encoded in float32 or float64

• title – if given will be included in the output metadata

• sr (int) – sample rate. I

Return type:

None