I am working with a TI PCM1807 ADC (datasheet: https://www.ti.com/lit/ds/symlink/pcm1807.pdf) and try do some signal processing on a FPGA. The PCM1807 output is I2S. I was wondering how to interpret this output. The datasheet says "2s complement". What does this exactly tell me? Has the sample to be treated as an integer or is there something more behind it like a special Q-number (https://en.wikipedia.org/wiki/Fixed-point_arithmetic#Notation) format?
Yes, it's a signed integer, but in signal processing applications, it's usually more useful to treat it as a fraction that represents the signal level relative to "full scale", where full scale is ±1.000.
So if the device is putting out 24 bits, you would divide the number by 223 (i.e., shift the "binary point" left by 23 places). In "Q" notation, this would be "Q23" or "Q0.23". Normalizing the data in this way makes the design of most of the signal processing pipeline independent of the actual number of significant bits in the input data.