Please critique my Dirac EQ project

[klettermann]

After much effort to optimize room placement I bit the bullet and went down the DSP/MiniDSP/Dirac rabbit hole. I feel like I'm at the point where I can use the software reasonably well but still have a long way to go in terms of interpretation and where to go with it. Kind of like learning the mechanics of driving but no understanding of navigation. Hopefully somebody will help me get to that next level. Here goes.....

Below shows work on FR. Scans were taken during different sessions so levels don't exactly match. Regardless, the Dirac EQing made a pretty obvious improvement. The tweaked version seems better still, though it did introduce another problem I'll get to later.

Freq Resp Scans (1/6 smoothing)
Purple: Uncorrected room FR
Red: Default Dirac correction, no tweaking
Blue: Tweaked version of the red Dirac-corrected scan.
Light green: Room noise with HVAC running. I'll work on that at some point.
Light blue: Room noise as-quiet-as-I-can-make-it. No idea what the 4kHz bump is, I'll try to find that too.

xx

Now we come to the spectrograms. Below is Dirac EQ'd with the tweaked version below that. Notice the ringing (??) at about 100Hz. Seems odd, I have no idea how messing with EQ would introduce that. But does it look OK apart from that? Last is the RT60 response for the tweaked version, not surprising I guess. .

Dirac EQ'd

Dirac FR Tweaked

Any comments, critiques, interpretations, suggestions or ideas of next steps - if any - would be terrific. Thanks in advance and cheers,

 

[kemmler3D]

Looks pretty good to me, but I agree the ringing at 100hz is concerning. I wonder if something is just resonating in the room?

 

[Oddball]

Looks pretty good as noted. But then you are cutting off a whole lot of bass response and SPL potentially if that matters to you. Having a different curve with higher bass could lift your overall SPL with potential dissociates around 100 hz dip.

 

[goat76]

What speakers do you have and does their direct response really need that much correction above 500 Hz?

 

[Oddball]

I have great speakers and 19 of them, but room is challenging - see the signature. But even without the room, I do believe that full range EQ is the starting point until actually proven that is is detrimental.

Also, my idea is to calibrate response to at least 100bB across the range, so cutting so much of low end response would not be my choice.

 

[OCA]

The graphs suggest that the original Dirac correction would likely sound good. It appears to have effectively avoided altering beyond the room transient, resulting in a balanced overall sound. In contrast, the over-EQ’ed version is likely to sound constrained or 'throttled.' Those peaks at 50/100Hz seem to stem from phase discontinuities related to AC ground hum, and are typically inaudible. However, trying to correct these may introduce quite audible artifacts. Contrary to the common belief here, it’s essential to listen to the results first instead of relying solely on the graphs. The more you flatten that response, the worse it is going to sound

 

[klettermann]

Thanks to all for the comments. This is very helpful! I'll try to fill in some of the blanks.

A bit of context: Speakers are factory restored Magneplanar MGIIIa's (tweeters inside) driven by Mark Levinson No. 332. This has proven to be an very nice combination. There's also a pair of Heco Aurora 30a subs. Nothing special about those except how amazingly cheap they were. Still, they seem to work fine. The room is L22 x W12.2 x H7 ft and partially treated - carpeted, soft absorbent couch and corner ACS bass traps.

Actually the original Dirac correction DID sound good, very good. That said though, so did the "over-Eq'd" filter. I wasn't aware of any obvious problems with it, though maybe I would have with more listening. Anyway, I did discover the cause of the 97Hz ringing. For whatever reason there's a very, very sharp notch at 96Hz that I tried to eliminate. The very narrow, sharp boost seemed to create the ringing artifact. I got rid of it and the ringing disappeared. The change doesn't seem to have any impact that I could tell. I'll post the improved pix tomorrow.

As for the bass, I'm fine with the balance I've got. It's very, verry clean, defined and clear. No muddiness, quite amazing really. The original Dirac EQ was actually a bit much for me, that's why I toned it down a bit. At some point I'll put together a 'bass cut' filter for late night listening or trance/house stuff when somebody is upstairs. The room is pretty soundproof, but the kick in such content can still bleed upstairs. Cheers,

 

[theREALdotnet]

It would be better to use the same smoothing on all plots, to make them comparable.

That said, the 100Hz dip is most likely a floor-bounce SBIR effect (for speakers/woofers about 85cm above the floor). It cannot be treated by EQ, any attempt to do so will result in a worse outcome, as you've seen. If it bothers you audibly you could fill it in by adding a sub closer to the floor.

 

[goat76]

I have great speakers and 19 of them, but room is challenging - see the signature. But even without the room, I do believe that full range EQ is the starting point until actually proven that is is detrimental.

If its possible to have different presets of room calibration in your system, wouldn't it be fairly easy for you to set up a blind listening test. One with the full-range EQ based on the in-room response, and another preset that has just the bass region equalized based on the in-room response and the rest of the frequency range based on a gated measurement.

You see, both can be ‘full-range’ adjustments even if one is based exclusively on the full-range in-room measurements, and the other is based on both the in-room measurement for the bass area and the gated measurement for the rest of the frequency response.

 

[klettermann]

It would be better to (1) use the same smoothing on all plots, to make them comparable.

That said, the 100Hz dip is most likely a floor-bounce SBIR effect (2) for speakers/woofers about 85cm above the floor). It cannot be treated by EQ, any attempt to do so will result in a worse outcome, as you've seen. If it bothers you audibly you could fill it in by adding a sub closer to the floor.

1. Absolutely, fully agree, my bad. The purple is indeed 1/12, the other 2 are both 1/6.

2. Admittedly I don't fully understand floor-bounce SBIR, so that's another path in my ongoing education. In any case, the subs are on the floor, not elevated. The speakers are another matter. Though they're floor standing, they're also ~H6 x W2ft panels, so there's a substantial radiating surface at 85cm above the floor. In any case, removing the 95Hz correction attempt got rid of the ringing.

Thanks and cheers,