Audyssey and System Set Up
Before diving into the set-up of the new Audyssey MultEQ system, I thought it would be prudent to interview Audyssey regarding it operation and calibration to ensure I would maximize its potential in my system.
Interview with Chris Kyriakakis from Audyssey on MultEQ
Audioholics: How does the Audyssey system differentiate between first arrival and reflected sound?
Chris: MultEQ does not "differentiate". It uses long impulse response measurements that account for the combined effects of first arrival and reflected sound. The challenge in doing that is to create filters that are short enough in length to be practical for consumer applications. MultEQ uses novel signal processing methods that are based on psychoacoustics to reduce the filter length without sacrificing correction accuracy.
Audioholics: How is this system different than others on the market that typically use a variable PEQ?
Chris: There are two fundamental differences:
(i) M ultEQ is not a parametric method that uses IIR filters for a number of bands. Instead it uses FIR filters. The advantage is that this allows both time and frequency domain correction, whereas IIR-based PEQ methods can only correct the magnitude response
(ii) M ultEQ combines the measurements from several listening locations in the room. Furthermore, it does so not by averaging, but rather by a weighted combination of the responses that results in a much more precise representation of the room problems. Spatial averaging methods suffer from the fact that a peak at a certain frequency in one seat might be a dip at the same frequency in a nearby seat. Averaging will smooth the peak and dip into a response that is nearly flat and thus tell the filter to not perform correction at that frequency. M ultEQ clusters responses throughout the listening area using fuzzy logic rules. This method assigns "importance" weights to each response and combines them appropriately. As a result, peaks are moved down and dips are moved up for all locations.
Audioholics: What are the recommended room conditions to obtain accurate calibration results (i.e. NRC level, RT60 time, etc)?
Chris: MultEQ adapts to noise in the room by measuring it. The test signal is then repeated at a higher SPL level until the desired SNR is achieved to guarantee good response measurements. RT60 has been taken into account in the design of the Audyssey target curve.
Audioholics: What are the max/min boosts of the Audyssey system?
Chris: These are determined by careful consideration of the gain structure in the host system. In the Denon AVR-5805 the maximum boost is 12 dB and maximum cut is 20 dB. MultEQ is designed to not attempt to drive loudspeakers beyond their capabilities just to correct a room mode.
Audioholics: How does this system avoid digital and analog clipping the Denon AVR-5805? IE (a 6dB boost at particular frequencies will result in doubling the power demand, etc).
Chris: Same answer as above. The system headroom is known and we apply limits to the filter gains so as to not exceed this. It is important to know what, if any, processes come after MultEQ so that we accommodate for those gains as well.
Audioholics: If your primary listening positions are less than 8, can you simply calibrate remaining positions by placing the mic in close proximity to the primary spots already measured to increase sampling accuracy?
Chris: Actually, the best use of the 8 positions allowed in the Denon AVR-5805 is to sample the listening area at regular intervals. It is not really intended to be "one measurement per listening position".
Audioholics: Should the system be recalibrated any time room parameters are varied? (ie. furniture moved/changed, etc)
Chris: Yes
Audioholics: What happens to the system when higher sampling rate audio signals are presented?
Chris: After the filters are calculated, a version of the filter for each channel is created for each of the supported sampling rates in the host system. So, a 96 kHz signal is filtered with filters optimized for 96 kHz.
Audioholics: Is the Audyssey system calibrated to work optimally for a Specific SPL level (ie. THX reference level)?
Chris: No, MultEQ operates optimally at all listening levels.
Audioholics: This system accounts for absolute phase of the speaker system. What can cause absolute phase problems? (ie. speaker placement, acoustics, or a phase inversion somewhere in the amplifier gain structure, etc).
Chris: Absolute phase can be flipped in the electronics or the wiring to the speakers (internal and external). It is a common problem in in-wall installations. Speaker placement and acoustics would not cause a phase flip. It can also be flipped in the microphone used to make the measurement, but then that would show up for each speaker in the system.
And Now on to the Setup...
Setting up the Audyssey system is a bit more involving than the typical run of the mill auto calibration systems, but as we learned in our first look of MultEQ this isn't your typical run of the mill room correction system. Be warned that calibrating your system with Audyssey (if done correctly) certainly isn't as speedy as fast food service. But if done, right it you will savor it like a good steak dinner. I decided to be patient - as my love for quality food is only equaled by my love for audio.
To begin, there are some basic guidelines you must follow to ensure accurate results:
- Be certain the noise floor of the room is as low as possible (recommended below 45dBA). This means turn off air conditioning, phone ringers, ceiling fans - even the fish tank.
- Map out the eight most common listening positions at a minimum of 2 feet apart. Use all 8 calibrations even if you don't have eight seated positions by placing the mic midway between the primary positions, or sampling different heights for line arrays or ESL type speakers.
- Place the mic on a tripod at seated ear level (not on the sofa). If you run speakers with long vertical baffles such as electrostats or line arrays, it is advised to calibrate at various vertical mic positions for the main listening areas so Audyssey can better map the response of the speakers.
- Don't make any sudden noises during calibration.
- Don 't run this calibration with small children around or any ex naval war veterans.
Editorial Note About Positioning the Denon Microphone ( Model# DMS-305) for Audyssey Set-up
It is strongly advised that the microphone be placed on a fixed tripod pointed straight up in the air for each listening position, or for greater accuracy pointed towards each speaker under calibration for every mic position. Because of the hockey puck shape of the Denon mic, it doesn't have a true omni directional response, thus any obstacle in its path or underneath it may adversely affect its ability to accurately read your loudspeakers. Placing the microphone on a couch or sofa may, depending on the material of the furniture, cause too much absorption or reflection, causing the Audyssey filters to excessively boost or cut certain frequencies - adversely affecting the calibration results.
I placed the hockey puck shaped microphone (optional accessory DMS-305 MSRP: $65) at my primary listening position and followed the prompts from the OSD.
Dealing with Error Messages During Calibration
Any avid reader of Hitchhikers Guide to the Galaxy would tell you, " Don't Panic" if an error message pops up. Typical error messages include: L/R: None, L/R: Phase, Ambient Noise is too High. I personally received the L/R Phase error and was actually surprised to see it. Typically these systems will incorrectly identify a speaker out of phase if one of the drivers within the speaker is wired out of phase (this is typically done on the bass portion of 3 way systems to achieve proper integration at
the crossover points). Oddly all of the drivers in my speaker system were wired in phase so I was unsure as to why I was getting this message. I decided to entertain the Audyssey system and wired my front speakers out of phase. To my surprise running Audyssey again revealed that my center channel was now out of phase. Still experimenting, I wired it out of phase just like I did with the fronts and reran the first step of Autocalibration. Bingo, no more phase error message. Recalling the article on MultEQ we wrote, I quickly realized that the Audyssey system was getting hung up on the absolute phase of each speaker which, to my knowledge, is an ability no other current room correction system possesses
There are two probable conditions which could cause this; 1) the speaker is electrically wired out of phase, or 2) the signal between the preamp and power amp has somehow been inverted. Since my situation didn't fit either of these two criteria, I was again perplexed by this, but decided to proceed with my front three speakers wired out of phase to make Audyssey happy.
After calibrating for all eight microphone positions and allowing the AVR-5805 to process the information (this took about 5 minutes) I eagerly popped in a known reference disc to take a quick listen. I was a bit disappointed to hear very little bass impact or slam and wanted to get to the bottom as to why this was occurring. I took a quick frequency response measurement with my LMS measurement device and noted almost a 10dB suckout at the crossover point. I quickly realized that by changing the phase of the front three speakers without doing so with the subwoofers I had caused this suckout. I was puzzled as to why Audyssey didn't pick this up, especially since it allegedly can delineate absolute phase for all speaker groups. After running extensive phase sweeps between channel pairs using the Avia setup disc I determined that wiring all speakers (including the subs) in phase was the best course of action for my application and I rewired the system accordingly.
As a side note, I temporarily hooked up three two-way bookshelf speakers in place of my main front and center channels and reran the Audyssey test. It didn't report any phase errors. I could only conclude the unique driver array of my RBH Sound T-2 System was confusing Audyssey with respect to determining phase. When this occurs, Audyssey recommends wiring your speakers normally and skipping this test which is exactly what I did.
The OSD showed what at first appeared to be dismal one octave equalization curves. How could this be and where are the subwoofer curves? This system is far more sophisticated than that, isn't it? Thus I polled Chris on this and he told me that " in order for the correction curve to be accurately represented pictorially it would have required significant amounts of processing power from the Denon. Instead they opted for a basic display solely for illustrative purposes. Although it doesn't show curves for the subwoofer channel, the Audyssey system does function on that channel. Again, Denon didn't provision for a display on their OSD for this ." Satisfied and relieved with these answers, I proceeded with the calibration with an overwhelming anticipation to finally hear the results of this system in all of its glory in my own home theater environment.
A quick listen revealed a more expansive soundstage, but also an overly energetic upper frequency response. The first time I ran Audyssey, I tried to take the easy way out and plop the mic down on the head of my couches at eight different listening positions. The end result was the system over EQ'ed my speakers, making them sound too aggressive and bright. The second time around I carefully held the mic and pointed it at each speaker for the primary four listening positions and then mounted it on a tripod for the other 4 positions firing straight up. In addition, since my speakers feature a truncated line array with a very large vertical baffle, I took a couple of measurements 2-3 feet above the two primary seated listening positions so Audyssey could better map the response of my front speakers. The end result was a much more natural and non obtrusive sound. Even the crude graphical depiction on the AVR-5805 showed less boost at high frequencies (especially in the 8kHz region).
Editorial Note on Post Calibration of Audyssey for Advanced Users / the Anal Retentive or Both
I highly recommend that experienced users check the EQ and calibration results with an Audio Analyzer or, at the very minimum, an SPL meter. I found the auto channel balances to be off by as much as 2dB if the mic wasn't carefully placed away from reflective surfaces, and nearly dead on (within +-1dB) if greater care was applied in the calibration process. Using the internal test tones of the Denon AVR-5805 won't be entirely accurate if you plan on using any of the EQ settings since the test tone mode bypasses all post processing. M y suggestion would be to use calibration discs such as Avia or DVE with your desired EQ mode engaged during the procedure to fine tune your response. Personally, I prefer a little bass boost in my response and nudged the subwoofer levels up about 4dB for more impact. If you only have 5.1 channel test disc such as Avia, you cannot calibrate the surround back channels, but you can swag it by interpolating the results of the 5 channel calibration between auto setup and the Avia or equivalent.
Note: Once Audyssey is engaged, a little green light will illuminate on the front panel display of the AVR-5805. If you alter any parameters of auto calibration, the light will change to red. If you wish to re-engage the auto calibration settings you can do so in the Auto Set-Up results parameter menu. The AVR-5805 even provides for three independent memory banks which can store auto calibration results and receiver settings and are accessible from the front panel.
