Investigations Using DeltaWave

[fas42]

Thought I would kick off a new thread which dealt purely with what one can find out interesting about audio waveforms using Paul's software, with the focus being on what what's happening in the signals, rather than about how DeltaWave is performing, or queries about what could be changed in the way it processes files. This would obviously overlap somewhat with the original thread,

 

 

but that shouldn't distract from the point of posting here - to find out interesting things in the audio waveforms, which might tell us something about why sound replay sounds different, for sometimes non-obvious reasons.

 

Now that Alpha Audio is uploading captures of what is actually coming out of the speakers,

 

 

in lossless form, this gives us another good set of 'meat' to play with - which brings me to the point of this first post: now uploaded are some files which are of the rig with swaps of ethernet cables; to see if audible differences are there ... you know, the stuff that is meant to have zero bearing on how the system sounds, 😉.

 

Listening to the files, of the same piece, they sound quite obviously different - and lo and behold, in DW they are so far from being a good match that there should be zero argument about the fact that they can sound different, if one listens in a certain way. As a trivial problem, there are glitches in the captures, which means if it matches at one point, then right next to that area the mismatch is huge - DW is not yet clever enough to compensate for literally a sample of two, needing to be deleted, or inserted, at random points in the clip 🤪. So it becomes an exercise in finding sections of the clips which are sufficiently 'clean' over a decent length of time - and even there it's a struggle for DW to do its thing, 🙂. A quick glance at a few spots in the Matching window shows major variations in the waveforms - they are so far from being "identical", it's not funny ...

 

So, they are going to sound different - the subjectivists will be happy, because it demonstrates that "cables matter" - but what are we actually seeing here - is it lack of precision in how the captures were made - experimenter error; or, variation in how parts of the rig were performing during each capture, because it was still warming up, say; or, was it that the cable actually caused the waveform to vary, by some mechanism? Or was it a combination of all of those things? My feeling ... is that it's the latter.

 

Which shows, to me, how difficult it is to do this exercise of comparing "different things" - it's trivially easy to show that the waveforms are different, if they are recorded - the hard bit is to try and sort out in that 'mess' what is going on; to have one thing, and one thing only, varying; so that one can actually measure/see what is significant in the exercise, of say comparing ethernet cables.