Room Equalization Technologies 101 – Part 2
Room Equalization in Progress
Introduction
In the last post introducing this topic we saw that virtually all AV receiver and sound processor manufacturers now include automated equalization or room correction in most of their products even those costing as little a $500.00. The good news is that most of them are implemented in almost the same fashion. A microphone is placed at a number of positions around the main listening position (MLP) which measures a swept or blipped tone from each satellite speaker and sub. The resulting data is then analyzed by software that is either contained in the receiver/processor or runs on a laptop connected to it. The resulting data is then programmed into one or more dedicated DSP chips within the receiver or processor. This analysis provides two areas of adjustment/correction:
- Basic
- It adjusts the relative levels of all your speakers, so that the center, mains, surrounds, height and subwoofer(s) are all playing at the same loudness at the MLP.
- It sets the crossover point between your satellite speakers and subwoofer(s). Some systems allow different crossover frequencies for different sized pairs of speakers.
- It determines how far each of your speakers are from your MLP and adjusts the delay settings in your receiver or processor accordingly so that all the direct radiated sound arives at the MLP coincidentally.
- Advanced
- It analysis the measurements to determine what sort of undesirable effects your room’s acoustics are having on the performance of your speakers, and then applies frequency equalization and possibly additional temporal compensation.
- It can optimize the sub crossover region with each satellite pair.
- Some systems support two (or more) independent sub channels and provide several different configurations for their setup and equalization.
The basic adjustments are very straight forward and easy to understand. The advanced processing that compensates for the rooms various frequency and temporal (time) issues is much more complex. Remember that no matter how good the digital room correction system is, it cannot compensate for all room problems. That is one of the reasons for trying to provide some room acoustic treatment prior to using these systems in order to make their job easier and more accurate. To understand why lets take a very brief look at what happens in a room at different frequencies (see here for a more detailed introduction to sound and room acoustics):
- 1Hz to approximately 300Hz
- At these low frequencies, particularly below 120Hz, standing waves are the biggest problem. This phenomina creates fixed maxima and minima of low frequency sounds within your room. It is not generally possible to provide enough absorption to prevent these so equalization is required to try to compensate for them. Re-arrangement of speakers, subs and adding more subs will also often help if thats practical. While compensation for excessive (an antinode) or boomy bass is possible, compensation for lack of bass (a node) is difficult to achieve without moving the MLP and/or adding more subs.
- 300Hz to 20,000Hz.
- Placing enough absorption and/or diffusion to control these frequencies is generally reasonably practical. Unfortunately in many rooms its shared use prevents the optimal, or any, acoustic treatment. This can result in levels of reflections with long durations and sound paths that can interfere with the final sound quality. Frequency and temporal equalization in this frequency range needs to be very carefully applied. There are a number of well known researchers who feel that our ears/brains are generally well enough equipped to sort things out in this region, if the room acoustics are anything close to reasonable, and that applying correction in this range can result in various unpleasant side effects such as a hard mid-range or a dull and lifeless sound. Remember that an omni-directional single point measurement microphone doesn’t ‘hear’ sound like your pair of ears. (Trinnov use a microphone array at each measurement point in order to try to mitigate this problem by taking a “3D” audio field measurement for each speaker.)
Trinnov 3D Measurement Microphone
SMS-1/REW, Denon Audyssey XT32 & Audyssey Pro Kit Microphones
Despite all the possible pit-falls that a room correction system can fall into, if you just do what the equalization software tells you to do they generally do a reasonable job at improving what you hear, some do an outstanding job but some………? Well lets see! If you really want to wring the very last bit of acoustic performance out of your room you can spend a significant amount of time, effort and money on learning how to use some of the more sophisticated room equalization systems software and hardware functionality.
In Use
My practical experience is mostly limited to the grand daddy of EQ systems Audyssey, MultEQ XT and MultEQ XT32 for both the multi-channel and sub only equalizers. I have extensive experience with the Velodyne SMS-1 and REW software, have dabbled with Dirac and miniDSP and used numerous analog and digital multi-band equalizers using just my ears and more recently REW.
Audyssey
EQ2 the basic solution doesn’t deal with bass frequencies and has a very limited number of filters for mid and high frequencies; I do not recommend it at all. MultEQ doubles the number of filters and adds some basic bass correction but again it is far too limited and I cannot recommend its use. MultEQ XT provides eight times the filter resolution of MultEQ and is at the bottom of where a room correction system becomes useful. I have had extensive experience with this system with and without Audyseys Pro-Installer kit and I find that its mid-range is too hard and its ability to effectively equalize anything more than two identical subs less than stellar. MultEQ XT32 is used in many top of the line receivers, pre-amps and processors. It supports 256 times the filter resolution of MultEQ XT and handles two subs (SubEQ HT) slightly better than MultEQ XT. It too supports the use of the Pro-Installers kit to edit the target results to better suit your room environment. Both versions of XT support eight measurement positions with the kit allowing for up to thirty-two. The Pro kit requires a lifetime license to be purchased for each receiver/processor. Its software (MultEQ Pro), besides recommending the optimal crossover frequency for each pair of satellite speakers, also allows them to be adjusted by the user and provides an additional routine that then optimises the crossover region between each satellite pair and the sub which is executed prior to the final results being uploaded to the receiver/processor.
Audyssey can generate filters in the range 10Hz to 24KHz, but it does not generate filters once it measures more than a continuing 3dB dip in room/speaker response (F3), nor does it generate any filters below the crossover frequency for the satellite speakers. This is notable as it prevents damaging subs and satellite speakers by applying excessive boost. However it can result in other restrictive issues, see my room equalization white paper here. As Audyssey does not calculate any satellite speaker filters below the selected satellite speaker crossover frequencies you should not drop your crossover frequency after the data is saved or you will have parts of the satellites speakers response with no equalization. You can however raise your crossover frequency if you feel it necessary with minimal impact. If you do not use a Pro Kit to make crossover changes you will loose its ability to optimize the sub crossover region for each satellite as the receiver software does not provide this function.
NOTE: All of the following graphs are from my speaker system and custom designed AV room.
MultEQ XT. 2 Ported Subs, FL and FR Frequency Responses. Before and After EQ.
A notable audible improvement but a long way from achieving what MULTEQ XT32 can do – see below.
After upgrading to four subs; two sealed and two ported these were the sub EQ results.
4 Subs MultEQ XT32/Sub. EQ HT. Frequency Response
4 Subs MultEQ XT32/Sub EQ HT. Room Decay
Note the inability of Audyssey to optimally equalize below 15Hz and between 30Hz and 70Hz plus the less than stellar room decays. This equalized bass sounded colored and muddy in nature.
MultEQ XT32 Final Speaker Frequency Responses. The Four Subs include: SMS1+REW+Rane AD22D – see below.
Audyssey’s algorithms (like many other programs) are predicated on taking many readings at different positions in order to obtain the best OVERALL room response so that everybody gets a reasonable seat at the movies. It does not optimize one seat like the MLP. The sound is therefore not really optimal anywhere. Measuring your rooms response at one point after Audyssey has completed its task will therefore often show room responses, particularly LF, that do not agree with the softwares predicted graphs. Audyssey also claims to be able to improve (reduce) undesirable room reflections (temporal responses). I have never found this to be the case.
MultEQ XT32 satellite speaker settings always seemed to be very accurate for all frequencies and the EQ generally provided a flat response as predicted, with good crossover integration to the subs. Unfortunately no matter how hard I try I have always found the corrected mid-range, in particular vocals, to be a little ‘hard’ and no amount of curve editing or applying the MultEQ Pro built-in BBC ‘dip’ curves corrects this. As Audyssey correction can be bypassed for just the front left and right hand speakers I have the option to use a pair of analog Rane ME30 1/3rd octave equalizers to optimize them for stereo music listening, leaving Audyssey to deal with multi-channel sound and movies, which it does very well. Stereo imaging, depth of image and dynamics have always been excellent after running MultEQ XT32.
Audyssey Sub EQ
For those who just wanted to equalize their sub(s) Audyssey used to sell a two channel stand alone sub equalizer that is basically MultEQ XT32 and SubEQ HT software and requires the Pro Kit (no license) to be used. It does a reasonable job of setting level, timing and frequency equalization of two identical subs placed at the same or different room locations. Being two totally independent channels it is more flexible as to how the subs are to be combined and equalized compared to a receivers/processors dual sub EQ provided by MultEQ XT32/SubEQ HT. It supports a user friendly software interface and a well designed level setting routine. Like the basic software it can also detect speaker phase inversions. However I found this not to be very accurate.
The MultEQ Pro Kit software doesn’t allow any editing of the final response nor does it apply any crossover processing. It just recommends a crossover frequency that can be used in the associated receiver/processor. It does allow you to ‘audition’ the equalized sub results prior to committing them to memory and overwriting the current sub EQ.
The equalization results below are for my two pairs of SVS mixed subs.
4 Subs – Best Equalized Frequency Response
4 Subs – Modal/Room Decay
Again note Audyssey’s inability to optimize the frequency response below 15Hz and the mediocre room decay pattern. With two identical subs the same issue was not encountered and the room rolled of at 15Hz as expected. However, the room decays were still not optimal when compared to what REW and the SMS-1 achieved.
While Sub EQ handles two identical subs quite well, I have never found that any Audyssey sub room correction provides optimal frequency and decay correction for mixed types of subs (I use two sealed and two ported). At no time did I find that Audyssey optimized low frequency ringing and decay, it had difficulty in detecting subs that were phase inverted and while it often displayed a flat frequency response, when measured it was often far from the truth.
The resulting bass was generally colored and muddy in nature.
Velodyne SMS-1 + REW + Rane AD22D
Although the included automatic EQ process offered by the SMS-1 does a reasonable job for single and two equidistant identical sub arrangements, it is not better than Audyssey. Measuring the SMS-1 results with REW generally showed a similar response to that displayed by its software but with only marginal improvements in a rooms modal decay times. Also just like Audyssey it does not handle dissimilar sub types well.
The Rane AD22D (a two channel delay processor) is used to initially correctly impulse time the two pairs of subs to the MLP prior to equalization. Final combined sub and satellite speaker timing and crossovers were achieved using the receivers MultEQ XT32 with the MultEQ Pro kit and REW.
The following responses were obtained after my four subs were impulse (time) aligned using REW and the Rane AD22D (an essential requirement). They look quite acceptable at first glance but a deeper analysis shows that the rooms decays were not improved by an appreciable amount. It is NOT just about getting a flat frequency response. The rooms time response must also be optimized.
SMS-1 Room Response Measurement – No EQ
SMS-1 Response after Auto EQ and A Little Tweaking
REW has the ability to analyze a frequency response and to determine its component resonant frequencies and nodes & antinodes and create the exact opposite in response so that it not only flatness the frequency response but also significantly reduces the accompanying modal decay times. REW allows you to continually re-run its software while you adjust the various parameters until you achieve the closet target curve to what you are looking for. Allowing you to see exactly its impact on the rooms frequency response and all the associated room resonance decay times. It will then create the frequencies, cut/boost levels and Q required to program into the selected equalizer. I have always found this combination to create a room response which is virtually identical to that predicted and measured by the REW software. Note: a frequency programming error of 1Hz to the predicted REW frequencies can have a very measurable effect, particularly on modal/room decay times. At these low frequencies an accuracy of better than 0.5Hz would be optimal but is not usually available.
Here are the results for just my four subs.
Combined 4 Sub Response, Impulse (Time) Aligned, No EQ
Sub response After REW Data is Loaded into SMS-1
Room Mode Decays, Subs Impulse (Time) Aligned, No EQ
Room Mode Decays After REW Data is Loaded into SMS-1
The SMS-1 provides very good equalization and a stellar improvement in low frequency decay and room resonance levels when used with the REW software. The long decay at approximately 15Hz is a structural resonance that cannot be further treated without rebuilding my room!
I finally opted to use the Velodyne SMS-1/REW and Rane AD22D for all initial sub EQ. After which the receivers MultEQ XT32 does very little to the EQ result achieved when it equalizes the satellite speakers and sub channel again, see below.
Purple SMS-1 + Rane AD22D only. Blue +Denon AVP-A1 No Audyssey EQ. Red +Denon AVP-A1 With MultEQ XT32
You can clearly see that the internal AC coupling and internal processing of my Denon AVP-A1 is rolling off the LF below 20Hz. Also note how minimally MultEQ XT32 has impacted the SMS-1 equalized result, (as would be expected). With my crossover of 40Hz or even 60Hz or 80Hz the XT32 improvements would be inaudible. Furthermore the EQ filters that Audyssey applied slightly increased the room decays that had been achieved using just the SMS-1 and REW. Those slight increases in decay times are far more audible than the minimal improvements in flatness and extension. Such a pity you cannot exclude Audyssey EQ on a per channel basis.
The resulting bass was tight, clean and very well defined.
1/3rd Octave Equalizers + REW
Stand alone equalizers have to be placed in the line level feeds to the power amplifiers or active speakers. This is quite doable with pre-amplifiers and processors and all sub feeds. However, it is often not possible for an AV receiver unless it has a line level loop-through allowing the equalizer to be placed within the feed to the receivers internal power amplifiers.
I use the Rane ME30B for either two or three channels (L/C/R front) of analog equalization. In my studio days I mainly used White 4000 series equalizers for control room equalization which were mostly set up by ear. Why you ask by ear? Back in those days programs like REW and real time analyses did not exist, and anything that could do a similar job was terribly expensive and quite basic. So a pair (actually several pairs) of ‘educated’ ears were generally just as good.
Rane ME30B 31 Band Analog Graphic Equalizer
REW or a similar program is used to measure the selected speakers response at the MLP running real time sweeps while the equalizer(s), are manually adjusted to minimize any significant peaks or troughs in the frequency response to create a flattish response or a specific room curve as shown below. Care must be taken as this is an omni-directional point measurement and some of the response anomalies will disappear if the microphone is moved even a few inches. Be especially careful trying to boost frequency troughs and move the mic around a little to see if it is just a point issue or a general issue.
My Optional Analog EQ Curves – Blue FC, Green FL, Yellow FR – All Four Subs Active, Crossover of 40Hz.
NOTE: These equalizers have a fixed high pass filter at 20Hz that has a significant group delay that is not compatible with my front speakers and subs. Even with the equalizer at unity gain with no boost/cut, and using a 40Hz crossover it causes the bass response to rise rapidly from 80Hz to 20Hz coloring the bass. Significant LF cut was required to produce the above response that fortunately did not adversely effect the rooms decays by any significant degree. As these L/C/R equalizers are engaged for material that has little/no content below 20Hz on these channels, this is not an issue. I will shortly post a more detailed review on the use and performance of these equalizers.
Dirac + miniDSP
My experience with these products is limited to watching others use them and listening to their final results. The software is very similar in operation and use to Audyssey but with much more extensive options and information. It is therefore more complicated to use for the average enthusiast. The biggest improvements that I found over Audyssey was that the software was a lot more flexible as was setting and adjusting target curves. Clearly their algorithms, filter and acoustical analysis implementations are very different to that used by Audyssey as I found the mid-range and bass equalization to be excellent. Given the opportunity I will use Dirac or ARC enabled hardware over Audyssey when I come to replace my Denon AVP receiver.
My Rooms Final Frequency Response
Here are the final results for my system using Audyssey MultEQ XT32 and the SMS-1+REW+Rane AD22D to equalize my 4 subs. Crossovers are all set to 40Hz except the front heights which are set to 60Hz.
(HF response rise after 12KHz is a sampling error)
If you are looking to ‘do it yourself’ then take a serious look at:
