The old with the new - Dirac with ST-120

I saw the post asking about a DAC and it triggered me to write this post.

A little over a year ago I built a ST-120. The ST-120 was the golden bullet
I was looking for. It sounds amazing and could not be happier.

After a few months with ST-120 I purchased a NAD C658 streaming DAC.
Though included, Dirac was not available at the time of purchase so I used the C658
like any other preamp/dac.

I spent several months listening to this setup feeding a pair of GoldenEar Triton Fives.
I was very happy with the setup; anyone who cared to listen were completely blown
away with the sound quality. My tin eared wife even noticed the quality differences.

Though I was extremely happy with the sound I tried Dirac when it became available,
it was included with the NAD so it cost me nothing.

If you are unfamiliar with Dirac, it is not simply a digital EQ, it is a digital
modelling/room correction system that does most of its work in the time domain.

In the simplest form, it requires 9 measurements per speaker to create a model of
time/frequency at/around the listening position. If you own a sub, add 9 more
measurements per sub.

Once collected the user can create custom response curves (the EQ portion) for the
main speakers and separate curve for subs. Dirac then creates a custom time and
frequency correction based on those settings.

Yes, I said EQ, but more importantly I also said time.
Timing information is adjusted for of all frequencies over the entire frequency range
per speaker; thus delivering balanced, coherent audio to the listener.

As I explained to somebody, its like this:
Living in Chicago we are accustomed to the noise and light all around us. The hum
of cars, trains, sounds reflected off all the infrastructure around us; the dogs and cats
and the neighbors wind chimes. The inability to see the stars or a mile into the
distance is a fact of life due to congestion and light pollution.

When you go to bed way out in the country you are suddenly aware of the quiet.
It all goes away. It can be disconcerting, or uncomfortable. If you look out the
window you can see all the stars you've never seen before. You can hear all the
nature you've never heard before. The noise is gone.

Dirac mitigates the noise by accounting for room response, reflection, timing issues,
etc. It allows the listener to hear the music, amplifier, and speakers by managing and
accounting for the horrors of most listening spaces.

For me, I have reached a level of listening pleasure I never thought I could reach.

I have the old - tube amplifier technology and all that it brings.
I have the new - streaming DAC with all the convenience of instant access to music.
I have the future - Dirac mitigating the influences of my listening space on the music.
Dirac does not take away from the ST-120, it allows it to be heard clearly.

For the record, 12 months ago I thought all the talk of digital room correction was crap,
experience has proven otherwise.

Dirac sounds like an interesting possibility if I can get a Dirac capable 2 channel DSP that I can insert between my streamer and my DAC... I've also been reading about REW+rephase but it sound like they are harder to use.

MiniDSP may work for you. If you were considering a new streaming solution, the NAD C658 is a DAC, preamp, and has a decent phono stage.

I've used REW to take measurements but nothing more. I feel like you need a Phd to understand that application.

I experimented with MiniDSP 2x4 HD for a few weeks and I did not like it. I also purchased the UMIK-1 calibration microphone (which I still have and use with the excellent REW SW). I really like the concept of what the miniDSP was trying to do with room correction by developing custom digital filters based on measured room responses and placing them where they would make the most difference, but it just didn't seem to make things sound better. In fact, it was just as likely to mess up the sound as to improve it in my experience. I wanted to like it, but couldn't get results. I may take another look at it one day. On the other hand, as multipath time of arrival processing matures I think these devices are going to make a huge difference, but still maybe only for one listening position. Watching with interest for now...

UPDATE: I just checked the miniDSP website and I see a lot of talk about Dirac Live. It looks like this application does not run on the miniDSP 2x4HD, just the more expensive devices. From what I gather only the Dirac capable devices run the time of arrival based room correction, so I guess I never experienced that yet. The 2x4HD is just a set of derived digital filters placed at the peaks (and optionally nulls) of the measured room frequency response. I think the only time correction was once delay setting for each channel.

I may have to try one of these more advance processors...

This year as Axpona I spoke with one of the engineers and they said (paraphrased), "Dirac should not be considered an equalization device though it does work in that realm, it does most of its work in the time domain".

While it does work to manage peaks and valleys (usually the result of room issues) it more importantly measures and corrects timing issues of source material reaching the listener. The timing issues may result from:
- distance between listener and speakers
- signal phase and delay in crossovers
- disparity between subwoofer and full range distance
(Dirac will be releasing a bass management system which can process the signal for several subs, adjust timing, phase, and, I believe, mitigate some mode issues in a room via active cancelling, etc)

If two drivers deliver a signal 1/3 out of phase it will result in some nasty peaks and valleys, I am sure; thus time domain correction if of paramount concern.

My system consists of:
- NAD C658 streaming DAC/Preamp (Dirac installed)
- ST-120 tube amp
- Behringer NU1000DSP (digital poweramp for subs)
- GoldenEar Triton Fives
- Two 12" sub woofers, each in their own sealed cabinets. (The C658 has true stereo subwoofer output. Single sub = summed mono, dual sub = L and R outputs)

Correction for a single position measurement (one person sitting in the sweet spot) Dirac requires nine measurement locations.
Listening position one (where you head would be), and eight additional locations forming a 1 meter cube around position 1.

In my case, each position results in five measurement sweeps - Left speaker, Right speaker, Subwoofer 1, Subwoofer 2, L+R+S1+S2; 45 measurements in all.

This allows Dirac to greater understand the physical relationships of speakers to listening position. This also refines the relationship of source to first, second, third, etc, reflections.

Armed with this information Dirac creates a filter for correction. In my case, the NAD can store 5 correction models thus allowing custom curves. I generally have the following stored:
- Totally flat (No EQ)
- 3.5db boost @ 40hz (3.5db @40 ramping down to 0db @ 300, for relatively loud listening
- 6db boost using same range as above, for moderate volume
- 9db boost using same range as above, for low level volume (cocktail party)
- Usually something ridiculous just to dink around with

I am still experimenting with curves, those above are the most recent. Usually people start bass boost at 80hz and above, I wanted to see what long gradual ramps sound like.


*I do not have time to reread this post, hopefully it is not impossible to read. I'll look at it again and refine later.