I just purchased a new receiver, and, as usual, it comes loaded with a grab-bag of DSPs. Some are good, some are totally useless ("SciFi" anyone?). I was just curious about whether there was any kind of software program that would allow you to run the unaffected audio signal into it, then the audio with the "enhancement" added, and show you what exactly the DSP is doing. This would be handy for evaluating any kind of audio signal processing. I'm sure there's some sort of "reverse-engineering" legal issue around here somewhere, but I just can't believe that there isn't anything that can do this. Any information that could help would be totally appreciated.

In general, any analogue signal is converted to digital. i.e. sampled, as step 1. If the input was digital to begin with, it may well be resampled to a different sample rate before processing.

Then some kind of transform(s) is applied in the digital domain. Examples are equalization, reverb, echo, flanger, frequency or time shift, convolution, and selective filtering. Some of these might be - more or less - undone if you know the exact parameters used. When something is removed, however, regenerating it is essentially impossible unless you have the unchanged original to borrow from. If there were multiple operations upon the data, you will most likely have to undo all of them before you can do the restoration.

To output to speakers, or any analogue destination such as a tape deck, the digital samples are converted back to an analogue signal.

An audio editor such as Adobe Audition has quite a few tools to let you analyze digital audio data in a variety of ways. However, you would have to have the experience to interpret what the program shows you in order to decide what might have been done to get to the state you are viewing. It would still often be rather difficult.

Yes. I suspect that you could get some idea about what's happening with a collection of test signals and a FFT analyzer.



See the list of analytical software I just posted over in the RMAF thread.

The difference between the original and the final versions can be determined easily enough but several rules (each with its own complexity) may have been applied to create that difference. In order to infer what the rules are can be complex and it is quite likely that there are insufficently many degrees of freedom from the data in the two samples to work them out.

IYSWIM

For linear processing, you ought to be able to capture the impulse response easily enough.

Interpreting it would be "interesting", though, generally, especially if any reverb or something of that sort is added.