For DAC measurements:
I would like to see the transition band of the amplitude frequency response, plus a few side lobes. Some DACs have rolloff behavior that is quite bad,
with an example here having a super-high transition band.
I would also like to know if phase is linear or not. I encountered a research context in which phase is important, and a minimum phase DAC caused trouble. Phase and amplitude frequency responses are hard to find for many DACs.
Lastly, I would also like to see the impulse response, but this is much less important than the other two. I know the impulse response is easy to draw wrong conclusions from without signals experience. For example, this graph may be a curiosity:
This is an "intermediate phase" impulse response. It is clearly asymmetric. What might the frequency response be? If you've seen other intermediate phase impulse responses, you may guess that the phase will suck, as it is trying to sacrifice phase for latency or pre-ripple reduction.
Both amplitude and phase are flat to Nyquist (at +-1). So the guess is wrong.
This is just a toy example; the window is suboptimal, with Fourier uncertainty of 0.75, compared to the optimum of 0.25. But it is much better than SoX's "intermediate phase" impulse, because SoX seems to have a math or design error. This is a fun feature that the impulse response is able to demonstrate. SoX's intermediate impulse response also reveals the source of its problem; it has nearly no pre-ripple even at -p 45.
The above graphs can be generated with this code in Wolfram Cloud:
Gam[x_] = Piecewise[{{x^1 E^(-x/10) , x > 0}}, 0];
shift = (1)*10;
SincS[x_] = Sinc[(x - shift)];
Plot[Gam[x], {x, -1, 80}]
Plot[SincS[x], {x, -1, 80}, PlotRange -> Full]
Plot[Gam[x] SincS[x], {x, -1, 80}, PlotRange -> Full]
FourierGamma[z_] = FourierTransform[Gam[x], x, w]
(*ReImPlot[Log[FourierGamma[w]],{w,-4,4}]*)
FourierSinc[z_] = FullSimplify[FourierTransform[SincS[x], x, w]]
(*ReImPlot[Log[FourierSinc[w]],{w,-4,4}]*)
FourierProd[z_] = FullSimplify[Convolve[FourierGamma[w], FourierSinc[w], w, z]]
ReImPlot[Log[E^(-I z shift) FourierProd[z]], {z, -4, 4}]
