These are essentially different rolloffs.
Let us say that the mid control is set to zero ohms. In that case, the .01uf cap is irrelevant. The gain of the stage is then essentially set by the ratio of the bass control (50k) and the 3k resistor to Vref. At max resistance, the gain of the stage is 53/3 or 17.7x. But with the control set to 50k, the treble rolloff begins around 38hz. Reduce the pot resistance to 10k, and the gain is reduced, but the rolloff begins around 194hz.
So the general principle is that as the bass control is "increased" there is a gain boost but it is selectively applied to only the lowest content.
Okay, let's set that knob to zero ohms, rendering the .082 cap irrelevant, and look at the other pot and cap. With the mid control set to max resistance, we have a stage gain of 28/3 = 9.3x. The treble rolloff, however, begins around 637hz. Drop that mid control by half to 12.5k and the gain drops to a little under 5.2x , and the rolloff starts around 1273hz.
For both those caps/pots, the treble rolloff is at 6db/oct, so there is still some treble coming through, just not very much.
The "Tone" control also produces a general treble rolloff at 6db/oct, but has no impact on the level. In contrast, the bass and mid pots in the feedback loop compensate for what they take away, by adding some gain. A much simpler (and more cost-effective) version of the circuit might skip the bass and mid controls, and replace them with a 12k fixed resistor and a 3300pf cap (stage gain = 5, treble rolloff beginning around 4020hz), and just use the tone and presence controls to vary the bite.
There, make more sense to you now?
