TOPIC:

Well, this has been a lot of fun getting this system set up, and I've done just about everything. I'm looking to post what I have done, my results, and solicit suggestions for any further tuning.

My system is for a bedroom that I am converting to a home studio. I am using a pair of 3-way active mains with two subwoofers.  Right now, the room itself is still in the "before" stage, so it is not acoustically friendly. I have a hardwood floor with no rug, a large window on one side, no room treatments, and the mains are backed up against a dresser that will eventually not be present. Also, right now I don't have my stands for the mains built yet, so the mains are sitting on the subs. This positions them about 1-1/2' below my ears, which is not ideal. The room is roughly 12'6" x 13' x 8'.

My process was: 1) With the help of REW, generate filters to flatten the curve for each driver and then generate crossovers; 2) Run MSO to get the 2 subs optimized; and then 3) run Dirac Live.  Dirac Live made a huge difference to the way the system sounds -- wow! I am really looking forward to seeing how much more improvement I get after I get some room treatments, my stands, and re-run. My only nit in the meanwhile is that Dirac decided the right speaker was much louder than the left, so it dropped the right by 7dB. The end result was that the left was a hair too loud and the vocals all seemed to be sitting too far left of center in the soundstage.  After some experimenting, I dropped the level on the left by 1.5dB, and now I'm really happy with how it sounds.  

I ran REW again at the listening position, and the results are below. The curve on the top is pre-Dirac, and the curve on the bottom is post-Dirac. 1/12 smoothing. 
1) How does this look?  Are there more tweaks I should make?
2) I will say that the final curve is not as tight as Dirac Live predicted. There seems to be about a +/-3db window (more in some places).  Is it realistic to hope for an even tighter range in-room without some significant room treatments?
3)  I used the default target curve, and it predicted an overall downward slant throughout the entire frequency range. I'm not seeing that here -- after 300Hz, it looks flatter than the target curve showed.  Does this look right?

Thanks much!

Kelby wrote:

My only nit in the meanwhile is that Dirac decided the right speaker was much louder than the left, so it dropped the right by 7dB.

You probably already did, but be sure that the first measurement is in the center of the listening area as that sets levels and delays. It's possible that if you had extraneous noise) car driving past) that might put the calculation off. I guess I'd try a single-point calibration to see if it was reproducible.

1) How does this look?  Are there more tweaks I should make?

Not sure why you have such a large bump at 30 Hz.

2) I will say that the final curve is not as tight as Dirac Live predicted. There seems to be about a +/-3db window (more in some places).  Is it realistic to hope for an even tighter range in-room without some significant room treatments?

The Dirac Live correction is based on nine measurements whereas there you are measuring in one location (I assume). If you do a single-point calibration you should see that "after curve" quite flat. (This is only for testing, not listening).

3)  I used the default target curve, and it predicted an overall downward slant throughout the entire frequency range. I'm not seeing that here -- after 300Hz, it looks flatter than the target curve showed.  Does this look right?

I had noticed at one point that but didn't spend time looking into it... wonder if it's related to single point vs multi?

Just curious what is the speaker configuration? (drivers, open/closed box, etc - just curious )

Thanks, John!

I will try calibrating again.  There were some background noises, so maybe I can get better results.

The 30 Hz bump was all Dirac.  The f3 of my subs was around 35 Hz before Dirac, so that's interesting. Maybe another calibration will change it.

My mains are Scanspeak D3004/6010 Be illuminator tweeters, 12MU mids, and 18W Revelator woofers. They are in a ported cabinet.   The subs are Scanspeak 13" Discovery subs with a Scanspeak 12" passive radiator in each.

Thanks again!

I re-ran calibration today with much less background noise. I also tweaked the curve a bit and like it a lot better. I got much better results.  The left/right balance was much better this time.   Dirac improved 40-75 Hz radically, bringing it into line with the rest of the curve.   
The imaging was immensely improved. Here is the new REW measurement, with both mains and subs.

It seems some users are getting "problems" during measurement by having the microphone too close to a reflective surface such as the chair/couch you are placing the the microphone on. The big symptom of this is SPL imbalance between channels.

Dirac tends to not correct for dips in SPL.

How did you "tweak" the target curve? I tend to think lowering the entire curve is OK but don't increase it.

If all your speakers are pushed-up against a dresser, then i would think that speaker positioning is a problem. Can you at least repo the subs?

How do the REW measurements compare to Dirac's "predicted response"?

Great points and suggestions.

The speakers/subs are right against a dresser now, which I think is a problem. I've got to build some stands to get the speakers positioned better, and that will allow me to move the subs around as well. I am also going to build a lot of room treatment; right now, there is none. As I plod through those tasks, I'll re-calibrate along the way.

Right now, I'm very happy with how they sound, but I know there is still a lot of room for improvement.

Cool