Mercurial > hg > aimc
annotate wiki/AIMCDataFormat.wiki @ 552:42f8ed31b2b1
(none)
| author | Ulf.Hammarqvist@gmail.com |
|---|---|
| date | Sat, 07 Apr 2012 10:00:56 +0000 |
| parents | bc75da283e83 |
| children |
| rev | line source |
|---|---|
| tom@434 | 1 #summary AIM-C simple data file format. |
| tom@434 | 2 |
| tom@434 | 3 The module [http://code.google.com/p/aimc/source/browse/trunk/src/Modules/Output/FileOutputAIMC.cc FileOutputAIMC] provides data in a simple format which can be read by other programs. |
| tom@434 | 4 |
| tom@434 | 5 The data is stored as 32-bit floats with the native byte-ordering of the machine the file was written on. Before the data, there is a simple header. |
| tom@434 | 6 |
| tom@434 | 7 == Header format == |
| tom@434 | 8 The first 20 bytes of the file contain the following fields: |
| tom@434 | 9 |
| tom@437 | 10 * 32-bit unsigned integer: frame count - total number of frames in the file (these can be SAIs, filterbank output frames or other things (eg. features), but all frames in the file are of the same type) |
| tom@438 | 11 * 32-bit float: frame period (the interval between adjacent frames in milliseconds) (TODO(tom): this is currently not set properly). |
| tom@436 | 12 * 32-bit unsigned integer: channel count (number of frequency channels in the output) |
| tom@436 | 13 * 32-bit unsigned integer frame length (length of each frame in samples) |
| tom@436 | 14 * 32-bit float: sample rate (audio sample rate in Hz) |
| tom@434 | 15 |
| tom@434 | 16 |
| tom@434 | 17 == Reading the files using MATLAB == |
| tom@435 | 18 Use [http://code.google.com/p/aimc/source/browse/trunk/matlab/AIMCread.m the MATLAB AIMCread function] to read the files. |