TOPIC:

Hey everyone

I recently purchased the 2x4HD + Dirac upgrade (which makes the DDRC-24).

After going though the whole measurement process, and loading the filters in the right slot, I noticed a decrease in stereo-imaging.

My current set-up:
- SONOS Connect
- MiniDSP DDRC24
- Dayton APA150 > Boston A25
- Dayton SPA500 > Dayton UM10-22 (sealed, ± 0,7 cb ft.)

The improvement on the low end is out of this world, absolutely amazing and definitely a keeper. This turns the UM subwoofer into one of the best subs I've ever heard (and I'veard a lot )

Unfortunately, my A25's sound better without Dirac switched on.

Dirac adds a delay of 0,8 ms to the left (or right, I'm not sure right now) speaker, which I think only makes thins worse.
So I thought adding a 0,8 ms delay to the other speaker would fix things. Well, unfortunately, it doesn't. It only makes it worse.
Also it adds about half a dB to one of the speakers...

When listening to music the voice should appear to come from the center, which it does in "direct" mode (only the Sonos and APA150). Witch Dirac switched on, this effect goes away.

I measured the distance to speakers so that the sweet-spot measurement was in the exact center of the speakers, at my listening position.

Here's some more information about the configuration:
- Speaker gain: -11 dB
- Speaker crossover: 100 Hz, LR12
- No PEQ

- Sub gain: -17 dB
- Sub crossover: 100 Hz, LF24
- No PEQ

Any ideas? Maybe I did something wrong while measuring...

Thanks in advance!

Your posted image shows significant differences in response curves between left and right through the critical midrange frequencies. Which means there are also differences in left right phase relationships that define a stereo image.
The best way to tune for stereo imaging is to voice one speaker until it sounds right to you ( forget the sub for this). Then use the measured response result of that one as the reference curve for generating a matching set of compensating filters for the other. If tuning a multi-channel system with LCR front, tune the center channel first then match left and right to that. Do this at moderately low volume levels.
Use subtractive filters only starting with the largest and highest peaks, progressing to smaller peaks. With PEQ's use the fewest and lowest Q filters you can get by with to achieve the end result. Don't worry about small deviations or dips. The important thing is that the left and right match each other as closely as possible, not that they independently match an abstract idealized curve.

Disclosure: I am new to the world of FIR filter tuning but have decades of experience in tuning custom stereo arrays by more conventional means. With minimum phase 1/6th octave filters on 1/3rd octave analog EQ's I tuned the channel with the roughest response curve first. Then using sine wave signals from a precision generator set to each band center frequency I would match the other channel by tuning to bring the sine wave localization to center between the two. This corrected both frequency and phase response between the two channels to a very high degree, including compensation for the effects of early reflection room acoustics on each speaker system.

Razz, thanks a lot! Haven't even looked at it that way...
I'm gonna try it right away

I sort of got to a solution:

What I did was measuring the individual responses of both speakers (with REW at the listening position), which resulted me in adding about 3 dB to the right speaker in the 500-800 Hz region. Also I played around with the delay while listening to some music. I eventually added 19 ms to the right speaker, with Dirac already adding 17 to the left speaker.

I'm not completely satisfied with the results yet, but that may be because I caught a cold, so my ears are kinda blocked at the moment and it's hard to locate sound (noted this too when listening through earbuds).

For now, I think playing around with some EQ settings and the delay I'll be able to get it to sound right once I feel better

Anyway: Thanks a lot

Glad I could help. Let me know how things progress.

@ gijs98

Ok, let's try to clarify here.. Can you maybe share your project? (Dirac Live project + XML inside your DDRC-24).
You add "19mS" to your system to be able to get correct audio? Something really doesn't add up here.. We're most likely missing some information..

By default, for 99% of customers, when they put the microphone at the center stage (it must be the sweet spot for the first measurement), they never have to do such extreme latency adjustment (or no latency adjustment as a matter of fact)... We're talking of moving the speaker by 6.5m here by doing an adjustment of 19mS... Something isn't right. Let's just get more info pls. Feel free to start a tech support ticket (support.minidsp.com)

DevTeam

I forgot a 0.... so it's 0,19 ms... pardon the confusion

Which would be about 6-7 cm (2,5" I think)

I was also wondering, if Dirac adds 0,19 ms to the left input signal: does that mean that the subwoofer (which is the left and right inputs summed together) gets one of the signals with a delay? Does that introduce some "extinction" in the signal (if that's the right word for it...)?


One other note, which just crossed my mind, when I used the Dirac Live software for Mac OSX (so not the MiniDSP version), I was literally blown away by the results. My speakers sounded better than ever. So I thought I should get the same results with the MiniDSP.

Now with the DDRC24, I'm not really impressed yet. So far the main speakers sound better without Dirac.
So I still uploaded the project files, hoping I can get the same results as before.

Such a shame we can't see what Dirac actually does (in terms of eq'ing etc..)


Anyway: many thanks for all the help!

(If this goes further that what's usually discussed on the forums, I'd be happy to start a tech support ticket!)



** edit: for some reason it won't upload a .dprj file... Here they are anyway: wetransfer.com/downloads/b8f0482f9214ab9...0171003200741/2b87de

So I ended up putting the speakers in an entirely different (larger!) room. It sounds so much better there. I get the feeling the other room is simply too small. Also Dirac doesn't add any delay to one of the speakers, so I guess that's a big plus

Anyway: I'm completely happy now