April 2013

Premiere Review!
Audiolics Anonymous Chapter 163
Dirac Live Room Correction Suite
The closest I've ever been to being in a concert hall.

  While I usually write my columns at the most a month before publication so the reader can obtain the newest information long before the print magazines can accomplish it, this one is being written in late January in the middle of a snow storm with 25 mile an hour winds and a temperature approaching 20 degrees F. Publication will be mid Spring when your fancy may turn to things other than audio, such as cleaning up the yard. Happily, in a few days I'll be leaving for three months in the Caribbean where I'll hopefully be relaxing on the beach not thinking about winter. Of course I'm taking my Smyth Realiser and headphones along with my OPPO BDP 105 and my hard disks so I'll be able to recreate my music room's sonics while sitting on the deck overlooking the Caribbean. Unhappily I won't have the ability to evaluate anything audio. Before I leave though, I must discuss a new software program from Dirac, a Swedish company, their Room Correction Suite, as I've found this to have given me the biggest improvement in my system's sound in at least two years.

Back at AA Chapter 157 the column discussed  the XTZ Room Analyzer Pro, a software program with a +/- 1 dB over the frequency spectrum measurement microphone for room correction. It was used to correct the frequency response and time alignment of my 7 loudspeakers and subwoofers, using active crossovers and adjustment of the speaker drivers, with pretty good accuracy. But even with it, I could not remove several anomalies in speaker and room interactions, meaning standing waves and crossover problems. At the same time a set of XTZ MH-800 desktop speakers were obtained for use with my secondary sound system, reviewed at AA Chapter 153. With these came the Dirac Audio Processor Controller software program which allowed a computer to do 12 different algorithms for adjusting the output  of these speakers for different room placements.

You may be asking at this point, why would one need the Room Correction Suite? Your speakers and equipment may cost north of $100,000, so they should put out a fairly flat frequency response, the room is probably decked out in all sorts of room correction paraphernalia, and either you are of the school that abhors tone control or have a modern preamp-processor with the Audyssey or similar inexpensive room correction system.

Well I have to tell you that no speaker is perfect, and in any media room no matter its size, configuration and correcting apparatus, there will be at best  +/- 5 to 10 dB peaks and troughs in frequency response, plus manufacturers tune the sound of every piece of equipment in the chain to what they consider to be proper sound. Plus, until this program, I've been very dis-satisfied with the results obtained with the other correction programs out there; the curves obtained being what the manufacturer considered to be correct.

About a month ago, Dirac contacted me about a new program of theirs, the Room Correction Suite. For 487.50 Euros now, about $649 and in the future 650 Euros, one  gets both the Calibration Tool and Dirac's Audio Processor Controller. That's a lot of money for a program when one considers that one can get the XTZ Room Analyzer Pro with microphone, or several other programs out there, some for no charge, to do room and speaker analyzing, with correction if one has active crossovers. But for those with normal passive crossover speakers or those with built-in amplification, all one can do is move the speakers around or play with the room acoustics to improve both the speaker and its room interactions.

What the Room Correction Suite can do is not only give you close to perfectly flat frequency response curves on up to 8 loudspeakers in your room, but will then allow you to adjust the frequency response of each individual speaker to any parameters you wish. See this link for technical descriptions of how it functions. Why would one want other than a flat response?  Because in a concert hall you are not being bombarded by perfectly flat sound, but  a response that depends on the hall you are in, your position in the hall and where and how the microphones are picking up that sound for the recording. Then you have the recording engineer's take on what the recording should sound like. So believe it or not, most people do not prefer ruler flat response from their recordings. That's one of the reasons each loudspeaker sounds different as each is tuned to what their developer prefers for sound.

So what equipment does one need to use the software? First off, one needs a reasonably fast computer for the computations and the final processing. Second one should use as good a measurement microphone as possible with decent cables. Dirac recommend the XTZ system, an Earthworks M23 or a Behringer ACM 8000. Happily I still have the XTZ unit. Why these? Because the program can use correction filters supplied by the microphone manufacturer to make the microphone perfectly flat for the program. Helpful is a portable keyboard and mouse and a decent microphone stand to be placed at the point where one's ears are for listening, a good connection from computer to preamp such as HDMI, S/PDIF or USB, and of course a great stereo system.

What are the steps? Turn on the program, and the First Step, Sound System, allows one to determine how many speakers there are and their configuration, plus the sample rates used by your system up to 44, 48, 88 and 96 kHz. Unhappily the program will not do 176 or 192 kHz, but Dirac states that such support will come during the first half of 2013.

Step Two, Mic Configuration, chooses which recording device or microphone to use and loads its calibration file if any. The Third Step, Output and Levels,  sets up the input gain and output volume of each speaker and allows one to name each speaker and make sure it's in the proper place in the program. One puts the microphone at the ear position and adjusts the gain for each speaker to fall within about -6 to -20 dB. so no distortion will occur.

Unlike most room correction software which are set up to measure one set of room volume parameters by having you move the microphone to set areas the program wants, Step Four, Measurements, allows one to tell the program whether one wants correction for a single specific room position such as a chair for one person listening, or an area such as a couch, or correction for the entire room. Single position placement would be the norm for most audiophiles as the larger the correction area, the less accurate will be the correction for the sweet spot.

One then puts the microphone in the ideal listening position and presses the start button. A frequency sweep then runs for each loudspeaker. The program then analyzes the data and has the person move the microphone to the next position where the frequency sweep and analysis occurs. When completed, one goes on to Step Five, Filter Design. The frequency and time delay of each speaker and driver is graphed out. At this point, one can make corrections through crossover adjustments, driver distance, speaker position, etc., then go back to the Measurements page, cross out the previous frequency sweeps and check the corrected speaker response like with other programs.  Or, and here's where the program starts to shine, by hitting the optimize button, the computer analyzes the curves and computes the correction curves for each speaker to make them flat both in the frequency and time domains. One then saves the corrected curves.

One then opens the Dirac Audio Processor Controller, loads the corrected curves to it, set the computer's sound controller to use the Dirac Controller, then set the Dirac controller's output to whatever output from the computer that you use, and play your music. One can turn the processor on or off to determine whether the curve's sound output is exactly what you want. But here's where I disagree with the program developers. As you can see above, what they do is produce corrected frequency curves that slope down by about 3 dB from 20 to 20 kHz. I asked them about this and this is their reasoning:

The automatic curve has a slope or tilt of -0.5 dB per octave. Its range is calculated from the measurements. In our experience, a target curve that is slightly tilting down towards high frequencies (like the auto-target) is often preferable - a flat target often sounds too bright. A loudspeaker with a flat on-axis response will usually have a slightly tilting in-room response. So, the tilt in our target is simply there to take into account that the loudspeaker is operating in a room. This may not fit well in all rooms as they have different acoustical treatments. This being said, you can of course almost always get a bit better results than the auto target when tweaking manually if you want to put some time into it.

On listening to these corrected curves, compared to my uncorrected speakers after I had set the crossovers up as flat as possible, the sound was clean but somewhat dull and shaded, due to that decreasing slope. You may find differently on your speakers. Yet here's where the program shines. The curves produced have a line through them with four moveable points present on the graphs, or by double clicking on the target line adding as many as you wish, to adjust the frequency curves. One can then move those points to correct in a smooth fashion the frequency curve to just about any shape one wants. Thus the program corrects narrow frequency anomalies and this corrects for wide frequency slopes. This would include increasing or decreasing  the lowest and highest frequencies reproduced and adjust the entire slope to produce anything from perfectly flat to sloping up or downward at either end, to increasing or decreasing the output of an area of the curve.

So what has the program accomplished on my system that couldn't be done with any other program?

1. Due to a discrepancy between the low and mid drivers frequency ranges there was a 6 dB drop-off in 600 to 800 Hz range, due to the woofer horn's frequency range, which was easily corrected as it wasn't due to a room standing wave and I had plenty of amp power to overcome it.

2. A 1 millisecond difference between the arrival time of the mid and tweeter and 2 millisecond difference between the woofer and mid on the main speakers were erased.

3. A 30 Hz peak in the output of the main speaker subs was removed, along with several other dips and peaks due to room standing waves.

4. The frequency response corrections allowed the surround speakers, although different from the main speakers in build, to sound more similar.

5. The subwoofer channel speaker's output were made flat from 20 to 120 Hz.

6. In one simple measurement, the program allowed the speakers to sound the best they ever had either by themselves or using several other programs available.

7. By being able to manipulate the program's curves, I was able to voice my speakers in such a way that I actually felt the closest I've ever been to being in a concert hall. My system ended up sounding best by putting a 3 dB smooth upward slope from 20 Hz to 20 kHz on the main speakers with almost flat frequency response on the surround speakers. Through manipulation of the curves you'd be able to do the same to optimize your speaker-room interaction.

So what are the drawbacks to the program?

I know that $600+ plus the cost of a measurement microphone is a fairly high price for a computer program, but I guarantee you that the results would be equivalent to meg-bucks updates to your room or equipment. I also know that there are many of you out there who feel it's an anathema to use a computer for high end playback, never mind using it to do tone control, then performing major corrections to speaker output. All I can say is you don't know what you're missing.

Obviously it can only be used with the computer as a source, which is not a problem if you have all of your music as computer files, and use your computer's internet connection for downloading music files or radio programs. But if your main sources are analog such as phono or the analog outputs of a digital player, all the program can do for you is allow for easy adjustment of speaker placement and crossovers. There is a work-around and that would be to either get an excellent soundcard with 5 to 8 analog inputs and superb A/D converters, digitize your analog and run everything through the computer. The hooker for the program though is that the highest bit rate is 96 kHz and not the audiophile preferred 176 or 192 kHz (which will occur later this year). As no computer that I know of can accept an HDMI input, which is the preferred method for multi-channel audio, and high end audio equipment still do not have FireWire or Thunderbolt output, there is no other workaround. Otherwise, this program is the best I've found for room and speaker correction and highly recommended.

If you wish to try the program, there is a 14 day free trial period and can be downloaded at this link but I suggest you don't unless you have the cash to pay for it at the end as you probably won't want to go back. As for further recommendations, all I can say is that Bentley, Rolls Royce and BMW use it in their cars, and Theta and Naim in their high end products.