The Elf wrote:Errr... Very interesting. Wonder what it means!?
If you compare the Sweep plots for the DAWs I mentioned it will become obvious.
Basically, imagine a sweep tone in a 96kHz system. It can range up to 48kHz. If that signal is sample-rate converted down to a 44.1k system, everything above 22.05kHz should be removed, and the resulting sweep plot should show a single clean line rising from left to right and then stopping.
If the SRC produces aliasing, you'll see an aliased image of the sweep coming back down, and possibly other aliases cris-crossing in the background. The brighter those lines, the worse the aliasing. Check out the SADiE 6 plot... :shocked:
The 1kHz tone plots are another way of showing the amount and level of those aliases and other spurious junk. The lower the level of the white 'grass', the better...
The Passband and Transition plots just show the shape and turnover frequency of the reconstruction filtering in 'overview' and 'closeup' views. An ideal filter response would be a flat line ending in a vertical 'brickwall' at 22.05kHz. In practice that's not possible, so the filtering will start at a lower frequency and round into a near-vertical dive.
The Phase response plot shows the amount of phase shift introduced by the filter. As most are 'linear phase' filters there is no phase shift and you get a flat line. But a few are 'minimum phase' and in those you will see the phase lag at high frequencies (eg. Ferocious Sampler-minimum phase plots).
Similarly, the Impulse response plots shows the, er, impulse response. With linear phase filters you will see pre- and post-ringing, and the main spike being delayed in time (to the middle of the time window). With minimum phase filters there is no pre-ring, but the post-ring is usually longer and the spike is earlier (to the let of the time window) because there is less filter latency. If the filter is more gentle, the ringing will be shorter, and steeper filters have more ringing.
Again, compare the Ferocious Sampler standard and steep plots, looking at the impulse and transition band plots.