I have to start off being clear that I find the response correction and curve editing capabilities of Dirac Live as executed by the DataSat RS20i to be the best I've used yet. Any issues or limitations explained below get into matters of how bass management is executed alongside the response correction that the Dirac software allows. It should also be noted that not all Dirac implementations are equal, as available processing power does play a role in how resources are allocated toward correction.
Let's first look at what Dirac does with the subwoofers. In the earlier measurements I posted
when we did the initial connection, setup and alignment of the system, I used a 2channel DSP
to double check the potential of the various subwoofer locations and investigate what we might encounter when blending them all together. We used 2 of the available subwoofer channels on the RS20i to separately feed the mono subwoofer channel to the front 4 SubMersive HP's behind the screen on the Sub1 channel and the rear stacked pair of F2's on the Sub2 channel.
Here we see 4 curves showing Dirac turned off for the front 4 SubMersives in RED
, 2 rear stacked F2s in BLUE
, and then with Dirac engaged for each with the 4 front SM's in MAGENTA
and rear F2's in CYAN
You can see the very impressive smoothing of the response. What I found so impressive was that this was a measurement at the main seat but Carl fed Dirac 9 separate measurements spanning the front & rear of the main row along with 2 in the money seats. This was not just a correction of one location.
Next I took a look at how the 2 combined. While the issue is much less severe here, we did find a similar result when the rear subs were turned on with no delay as seen in GREEN
, vs. the result of adding the 7ms of delay I had settled on in our previous measurement session seen in YELLOW
Let's look at what happens when the delay on the rear stack of F2's is adjusted: 0ms
These measurements show we have a good range to work with depending on what we find elsewhere.
Starting with the 7ms delay setting at the center of each row we have MAIN
Remembering that the calibration points were not taken fully in the front and rear rows (maybe another time), I first looked at the rear row as the upper bass notch is more easily affected by delay adjustments (shorter wavelengths are affected more per time adjustment
There's room for significant improvement there, although we can't pick & choose a graph from each, but have to get the best set...
Similar spot check of the front row: 0ms
Looking at the 3 sets of measurements I settled on a 3ms delay to the rear subs which give this set of measurements MAIN
Looking at the deviation from the middle row we have MAIN
Here's a final look in GREEN
at what we settled on for the summation of the subwoofers as measured in the main row with the 7ms starting point
and 0ms delay
shown for comparison.
You can see we ended up with a starting response that is +/-3dB from 70Hz extending down below 10Hz to the lowest 5Hz extension setting Dirac allows.
This was with a starting flat target curve in Dirac. All adjustments to the subwoofer response was made to the subwoofers as a group. I believe I ended up shelving up the low end about 4dB around 35Hz and raised the subwoofer level 3-5dB over that of the main speakers. I didn't get measurements after we did the tweaking as everything was adjusted from the flat starting point.
Next up the sub-center channel integration...