Help Request in Using Dirac Trial

[kilroy]

For the past week I’ve been trying the Dirac Live trial. Based on online forum posts here and elsewhere I was expecting great results. Haven’t got there yet. I am mostly fighting a room setup where speakers are placed along a long wall, with the right speaker closer to a corner than I’d like, and the left speaker far from the left wall. Live Calibration does show several bass peaks that I expected to see. Because the room is a shared living space, using room treatments is not really an option.

 

In general the results have sounded flat and uninvolving, with a reduced soundstage width. The bass does seem improved however. Still, when comparing with the filter in and out, the no filter position is preferred; the sound is just more relaxed and natural. (Interestingly, using the frequency spectrum in Live Cal as a guide, I used JRiver’s parametric equalizer to reduce these bass bumps and it was a great improvement.)

 

I’ve made several filters using the sofa pattern, although it’s a single listening position, based on recommendations I’ve seen online. The measurement areas have varied in size. The largest was about 60 feet wide and 30 inches front to back. The first measurement position has varied from about 5 inches from the back measurement row to about 20 inches towards the front. Flavio of Dirac said in another post that the first position should be 50 cm from the back row. Tried that. Another post said the listening position measurements should be within a 1 meter radius from the head position. Tried that too.

 

For vertical positioning the spacing has varied from about 6-8 inches, centered on my ear position in the chair. I’ve also followed the staggered mic position guideline on the horizontal plane, and have measured out and tape-marked the floor to make sure measurement locations are exact.

 

I’m ready to try the single seat measurements, but again the measurement spacing will just be guess work. How small to go? There seems to be an infinite amount of adjustments and trials available, so the fun is fast going out of this experimentation.

 

If anyone can share their experiences and help toward using Dirac to get the best results, I’d appreciate it. Thanks!

[tne]

There is a thread in this forum that Bill Scott started and that I (recently) added to that describes some of our experiences. Check it out. For me, there is a noticeable improvement with the filters engaged. I have two seating positions and made sets of nine measurements for each, with the chair model from the calibration software as a guide to positioning and with the microphone pointing in the horizontal plane. For the first measurement I placed the microphone where the center of my head between my ears would be, then followed the measurement pattern in the software. All measurement points within 0.4m radius from the central point (most closer).

 

Of course speakers, rooms, listening preferences, etc. all differ, so the experimentation that you are doing is very important. Actually it is pretty cool to have a visual representation of the speaker/room response function based on these measurements, no matter how one decides to address any issues that are present.

[kilroy]

tne, thanks. I understand the room etc. may be a factor. So today I will embark on the next leg and do single seat measurements. Wonder if you might share a bit more info on your measurement positions, especially the height differences between high and low mic positions? I'll stick with the small radius and will make all measurements within that 33 inch diameter. One reason I avoided the close pattern was that a post by Dirac said that the effect could be overdone in a small area and not have the best results. But I'll go there and see what happens.

[tne]

^ sure

 

my room is about 23 x 12 ft with about 9 ft ceiling and a carpeted concrete floor. My speakers point down the long axis of the room slightly toe-ed in with a sub-woofer in the middle (my sig has all the equipment details).

 

about half way into the room on the central axis is a large chair with ottoman - this is the upstream position. Further down is a desk and work chair, typically on the central axis and this is the downstream position. The data from the two sets of measurements are shown in post #30 of the thread I mentioned in my earlier post. As you can see there are significant differences in these plots, particularly in the bass region.

 

I used the UMIK-1 with the supplied calibration file and attached to the small tripod. To adjust the height, I used boxes, boards, books. Positions are symmetrical around the central position. The procedure that the calibration software lead you through is straightforward and it on starts with the "sweet spot" which I define as where the middle of my head/ears will be in the seating position being measured. Use the top view and the frontal view to guide you through the positions and sequence. Once again, make sure that the measurements are symmetrical about the central measurement position by using a tape measure (or such). The figures will give you a good idea of how far from the center is optimal. I made all measurements with the microphone pointed in the horizontal plane.

[wanta911]

From my experience and reading countless other user experiences - don't get too hung up on the ancillary measurements, the first "sweet spot" measurement is the critical one. What tne has posted above is a good guide.

 

I experimented with a lot with mic positions all within 2 metres on the horizontal plane and found that the differences were very subtle if at all, as long as the first measurement was in the right spot.

 

Getting the curve that you like is the hard part and it took me a lot of trial and error to find a good base curve that I could tweak into variations depending on the source material. I settled on the B&K curve as per this CA blog by mitchco:

 

What is the tone quality of your audiophile system? - Blogs - Computer Audiophile

 

Using 0dB as the reference point, it is essentially 0dB @30-40Hz, -0.5dB @200Hz, -3.0dB @2kHz & -6.0dB @ 20kHz.

 

I have 4 filters on my MiniDSP/Dirac unit and this is one of them with the other 3 being subtle variations of it.

[kilroy]

From my experience and reading countless other user experiences - don't get too hung up on the ancillary measurements, the first "sweet spot" measurement is the critical one. What tne has posted above is a good guide.

 

I experimented with a lot with mic positions all within 2 metres on the horizontal plane and found that the differences were very subtle if at all, as long as the first measurement was in the right spot.

 

Getting the curve that you like is the hard part and it took me a lot of trial and error to find a good base curve that I could tweak into variations depending on the source material. I settled on the B&K curve as per this CA blog by mitchco:

 

What is the tone quality of your audiophile system? - Blogs - Computer Audiophile

 

Using 0dB as the reference point, it is essentially 0dB @30-40Hz, -0.5dB @200Hz, -3.0dB @2kHz & -6.0dB @ 20kHz.

 

I have 4 filters on my MiniDSP/Dirac unit and this is one of them with the other 3 being subtle variations of it.

 

 

Good to hear about the sweet spot / #1 mic position. On the last filter I put the mic exactly in the middle of the head position, also as tne earlier said. Which, by the way was the first single seat filter, which didn't sound much better / different than the earlier filters.

 

Well, one thing I haven't done is play with the filter settings. I sort of presume (I know...) that the Optimize button means it does the work for me. I'll take a look at that blog and start tweaking some filters.

 

Thanks.

[wanta911]

Good to hear about the sweet spot / #1 mic position. On the last filter I put the mic exactly in the middle of the head position, also as tne earlier said. Which, by the way was the first single seat filter, which didn't sound much better / different than the earlier filters.

 

Well, one thing I haven't done is play with the filter settings. I sort of presume (I know...) that the Optimize button means it does the work for me. I'll take a look at that blog and start tweaking some filters.

 

Thanks.

 

The optimize button will make the adjustments to wherever you set the data points. The default curve that Dirac gives you was OK but not quite right for my speakers in my room and for my taste.

 

You should also take into account that you are used to the way your speakers sound and a big change can sound weird at first. A lot of people only correct the low frequencies and leave the rest of the frequency range untouched to maintain the voice of the speakers they bought. Personally, I ended up doing full range filters but my B&W speakers have a certain character in the form of a midrange hump between 500 & 800Hz. By adjusting this area by +2.5dB over the B&K base curve, it sounds more like the speakers "natural voicing" and this is my most used filter.

 

All of this has taken months of tweaking, I gotta say that I wasn't sure about my purchase in the first few weeks but this is a tool that you need to spend a lot of time with and don't rush the tweaks, listen to different curves for prolonged periods with different source material before making changes and you'll find it to be worthwhile and a lot of fun in the long run.

[hifi_swlon]

From my experience and reading countless other user experiences - don't get too hung up on the ancillary measurements, the first "sweet spot" measurement is the critical one. What tne has posted above is a good guide.

 

I experimented with a lot with mic positions all within 2 metres on the horizontal plane and found that the differences were very subtle if at all, as long as the first measurement was in the right spot.

 

Getting the curve that you like is the hard part and it took me a lot of trial and error to find a good base curve that I could tweak into variations depending on the source material. I settled on the B&K curve as per this CA blog by mitchco:

 

What is the tone quality of your audiophile system? - Blogs - Computer Audiophile

 

Using 0dB as the reference point, it is essentially 0dB @30-40Hz, -0.5dB @200Hz, -3.0dB @2kHz & -6.0dB @ 20kHz.

 

I have 4 filters on my MiniDSP/Dirac unit and this is one of them with the other 3 being subtle variations of it.

 

Not wishing to distract from the OP, but thanks for posting that.

 

Having just ventured into a Dirac Live trial to try and combat my 'bright', 'harsh' sounding lounge (which can't accommodate acoustic treatment), this is just what I need.

 

I'm very pleased with my initial test, just based on one set of measurements and a default filter.

This is exactly what I need to take things forwards.

[Fitzcaraldo215]

Not wishing to distract from the OP, but thanks for posting that.

 

Having just ventured into a Dirac Live trial to try and combat my 'bright', 'harsh' sounding lounge (which can't accommodate acoustic treatment), this is just what I need.

 

I'm very pleased with my initial test, just based on one set of measurements and a default filter.

This is exactly what I need to take things forwards.

 

I have had Dirac for awhile now, and I have not been the least bit tempted to fiddle with the stock target curve. Just because you can change and adjust the frequency balance via the target curve does not mean that you must or should do it.

 

I have a friend who has converted to Dirac, and he is a compulsive tweaker and fiddler who never seems quite satisfied with his sound. He does not seem to know good sound when he hears it. He has massaged and adjusted his target curve ad infinitum by ear. But, clearly, he is totally lost and fiddling for the sake of fiddling. None of his adjustments sound any better or sometimes not very good at all compared to the stock curve in my opinion. But, he has wasted hours and hours doing that with critical comparisons by ear when he might have just been sitting back enjoying the music. To each his own, I suppose.

 

I also agree that in many cases, full range EQ is a must. Yes, the biggest corrected peaks and dips are room modal issues in the deep bass, but I think too often most speakers in most rooms are too bright. Dirac's stock target curve including gentle HF roll off seems to manage that very nicely.

[flak]

Hi Kilroy,

 

one aspect that is often overlooked is the setting of the DSP gain in the DAP... you should adjust it at the lowest value (the nearest to zero) that does not incur into digital clipping.

Please note that you should keep some margin as the ideal setting can be different for different recordings.

 

Good listenings, Flavio

[hifi_swlon]

Hi Kilroy,

 

one aspect that is often overlooked is the setting of the DSP gain in the DAP... you should adjust it at the lowest value (the nearest to zero) that does not incur into digital clipping.

Please note that you should keep some margin as the ideal setting can be different for different recordings.

 

Good listenings, Flavio

 

Orbital - Wonky, is a good test album here

 

Earlier this week I noticed for the first time ever the clipping light on the Devialet flashing - just playing out clean USB (no room correction or filters)

[REShaman]

Hi Kilroy,

 

one aspect that is often overlooked is the setting of the DSP gain in the DAP... you should adjust it at the lowest value (the nearest to zero) that does not incur into digital clipping.

Please note that you should keep some margin as the ideal setting can be different for different recordings.

 

Good listenings, Flavio

 

I have had Dirac for awhile now, and I have not been the least bit tempted to fiddle with the stock target curve. Just because you can change and adjust the frequency balance via the target curve does not mean that you must or should do it.

 

I have a friend who has converted to Dirac, and he is a compulsive tweaker and fiddler who never seems quite satisfied with his sound. He does not seem to know good sound when he hears it. He has massaged and adjusted his target curve ad infinitum by ear. But, clearly, he is totally lost and fiddling for the sake of fiddling. None of his adjustments sound any better or sometimes not very good at all compared to the stock curve in my opinion. But, he has wasted hours and hours doing that with critical comparisons by ear when he might have just been sitting back enjoying the music. To each his own, I suppose.

 

I also agree that in many cases, full range EQ is a must. Yes, the biggest corrected peaks and dips are room modal issues in the deep bass, but I think too often most speakers in most rooms are too bright. Dirac's stock target curve including gentle HF roll off seems to manage that very nicely.

 

 

While I never argue with perception (read as subjective, objective, non-jective), the perceptions and advice provided in the quotations in the posts I identified match my experience since I first employed iRC with Amarra Symphony back in September 2013. And subsequently with the introduction of roon and the fortunate choice I made in adding Dirac Live (Full). In both cases, with either editions, iRC and / or Dirac Live (Full), the enjoyment of music and rendition of sound quality is enhanced without exaggeration or artificial mangling of what one starts with and what regulation iRC and Dirac Live add to the equation. And,

 

This perspective is intended to discount those who do take the time and focus to "adjust" the impulse room correction as they prefer it. I am very happy with the default results with both editions. Highly recommended whichever player or OS X edition employed.

 

Enjoying the music,

Richard

[jjkale]

Flak and Fitz 215,

I am going to run the Dirac Trial soon.

I have full range mains, Def Tech Mythos ST. They are employed in both a 5.1 and a stereo configuration. They are rated down to 16 Hz but I doubt they get very close to that spec.

Does it make sense that I first run a Dirac sweep to best determine where they roll off and then use that for the basis of a stereo curve?

With that information I could then set a crossover freq for 5.1 listening; run a 5.1 sweep and manufacture a first-try 5.1 curve.

Is that reasonable? Or rookie nonsense?

Second question do either of you know if JRiver can change DACs/drivers on fly? For instance play Porcupine Tree 5.1 24/96 and then Soundgarden 2 ch 24/192 without any manual input...

Thanks dudes, I am learning a lot from your expertise.

[flak]

Hi jjkale,

 

for you stereo listenings Dirac Live will measure the output of your full range Def Tech Mythos ST and automatically adjust the target at low frequencies in order to take into account their capabilities (you don't want the correction to boost frequencies below)

Crossover frequency to the LFE in a multichannel home theater configuration is another story... if you have a valid subwoofer you may simply look at the Dirac graphs and choose the crossover frequency that is one octave above where the response falls below the Auto Target in your other speakers that are likely less capable than your main ones (i.e. 80 Hz if you find 40 Hz)

That's a simple approach but you may experiment with more intricate/advanced ones, this is one by the AVS member AustinJerry:

A brief guide to creating custom target curves with Dirac Live

 

Good listenings, Flavio