search

OSOceanAcoustics / echopype

Enabling interoperability and scalability in ocean sonar data analysis
https://echopype.readthedocs.io/
Apache License 2.0
89 stars 70 forks source link

Reduced sampling rate Power/Angle samples #702

Open marianpena opened 2 years ago

marianpena commented 2 years ago

Hi, I've been testing echopype with EK80 data recorded in CW mode and reduced sampling rate Power/Angle samples; backscatter looks fine but Sv is full of NaN values. I used ep.calibrate.compute_Sv(ed,waveform_mode="CW",encode_mode='power')

Echopype specifies 2 modes ("complex"and"power") but three data formats are possible:

The Stored sampled data for WBTs running CW options can be used to reduce the amount of recorded data when you are using a Wide Band Transceiver (WBT) with CW pulses. • Select Complex samples to use the default data format. • Select Power/Angle to reduce the file sizes. • Select Reduced sampling rate Power/Angle samples to minimize the file sizes.

Is this last one supported?

Thanks!

marianpena commented 2 years ago

Ok I found the error and how to fix it. Temperature, salinity and depth had the same length as ping_time but only the first value was filled (the rest as NaN). I fixed it by specifying: env_params = { 'temperature': ed.environment.temperature[0].values, # temperature in degree Celsius 'salinity': ed.environment.salinity[0].values, # salinity in PSU 'pressure': ed.environment.depth[0].values, # pressure in dbar } and then ep.calibrate.compute_Sv(ed,waveform_mode="CW",encode_mode=encode_mode, env_params=env_params)

leewujung commented 2 years ago

Hey @marianpena ! Thanks for reporting this. I think this is actually a bug in the code, as the dimensions are not cast correctly. 😔 I will re-open this and we'll try to get the fix in soon!

leewujung commented 1 year ago

Hey @marianpena : I have been doing some refactoring of calibration code, which contains particularly parts about cal_params and env_params intake. I think my changes fixed the issues you exposed. Could you point me a sample file and a snippet of code to test? I don't have access to an EK80 system right now to produce reduced sampling rate data. Thanks!