HI Jack,
Presto can read a number of different formats. Many people including myself use the filterbank format. This is what I would recommend as a number of people here will probably be able to help with this.
Generically speaking a filterbank file has a header, followed by binary data where such binary data represents a spectrum for each sample period. The binary data can come in either 8-bit integers or 32-bit floats as defined in the header. How do you get to such a file? This is where I would ask other people to step in with practical implementations. I personally have been standing on the shoulder of giants as I am using software from the professionals to get the data from our pulsar spectrometer. I am also using SDRs for pulsar observations and in this case I am using a Gnu radio based implementation from others to create the filterbank file.
There are two different approaches which are used in the amateur community: One approach is to record the raw IQ data and then do the FFT and writing of the filterbank file in a subsequent offline step. The disadvantage here is the large amount of data which is generated. The other approach is to do the FFT on the fly and then write the filterbank file directly . Here one needs to take care that the processing power is good enough to do the FFT without loosing samples. I will leave it at that as others are more into this subject.
Unfortunately there is no document that I am aware of which describes the filterbank format. If you are interested in the details and want to do something by yourself you may want to look into pysigproc (
https://github.com/devanshkv/pysigproc) which has routines to work with filterbank files and headers.
Best regards,
Wolfgang