Wave-data Handler Analyser and Modeller
From EARWiki
Data Model for Wav
- Data is collected as discreet WAV files.
- We want these to be a logical continuous sample with a start, an end and everything in between.
- This whole logical object has the same set of parameters such as sampling rate, etc.
- Code should be written that allows subsequent code to reference WAV files in this manner with simply a sample ID.
- Sample IDs should have some other glue code that allows the ids to be referenced to the site and sampling characteristics for the particular sample.
This can also be useful when we decide we want to parallelize analysis. Function calls for things like make_spectrogram() can distribute the process in units of single wav files to multiple computers in order to perform analysis. The parallelization can be included transparently to the user if we have the logical framework set up in the beginning for combining experimental samples.
Below shows how code can logically combine WAV files in such a manner that each individual wav file is not important to subsequent processes.
######## ######## ######## ######## # WAV1 # # WAV2 # # WAV3 # ... # WAVn # ######## ######## ######## ######## t0 tn t0 tn t0 tn t0 tn ########################################### # Logical object # ########################################### t0 tn
