A mini-monitor is an ideal solution for a small room. I look at placement issues in depth. As in the large room excellent performance requires a speaker with optimal ANSI/CEA-2034 curve sets. I show the Revel M106 achieves the same performance, matching in-room response curve of the large room. This is a result of it excellent measured performance in the ANSI/CEA-2034 curve set. Lower frequency room modes extend in-room low end response enough that full frequency performance without a subwoofer is possible. Placement of a speaker in small room is critical to getting excellent performance. Listening tests confirm that the engineering of the M106 has been done in service of the music.
A mini-monitor is an ideal solution for a small room. I am going to spend significant time in this section discussing optimal placement solutions for the Revel M106. Much of this applies for placing any speaker in a small room. When properly placed, the Revel M106 measures and sounds like it does in my large room.
You may want to place your music system in a fully enclosed room to prevent disturbing the rest of the family. I have been looking at floor plans of new construction and they tend to have large open floor plans on the first floor. I have not observed enclosed rooms larger than the master bedroom. The largest enclosed rooms in most new construction are other bedrooms. Some rooms are too small to be workable which would be around (11ft * 13ft *8ft). It is rare to find a bedroom or first floor study with the dimensions of (14ft * 16ft * 8ft). One floor plan I looked at traded one garage space for a “media room” which was only 11ft wide but long, as you will see below, 11ft is just about the limit for excellent sound quality.
If you live in an old house keep the folks from Public Broadcasting System far, far away. They tend to rip down walls and turn nice, isolated, sound rooms into public spaces.
I have two enclosed spaces where I took measurements. Small Room 1 is (12ft * 16ft * 8ft) and small Room 2 is closer in shape to a square (13ft * 15ft* *8ft). The dimensions look similar but the placement of the speaker is different. Small Room 1 has the speakers on the short 12ft wall and Small Room 2 places the speakers on the long 15ft wall.
Depending on the layout of the room, you may not have a choice of walls. A door is going to make one wall unusable and if the door is in a corner, two walls may become un-usable. If you are setting up the system in a room that was specifically designed as bedroom, you may have another door to the bathroom that may take out another wall. Closets can also make a wall unusable. These are the factors that drove the speaker placement on the longer wall in Small 2
The key to getting excellent sound out of most speakers, including the M106, is keeping the speaker to listening distance to at least 9ft. If you are lucky and have a short wall, this is what you want to use. When the speaker is on the short wall, you have plenty of room to keep the speaker from the front wall and the seat from the back wall.
In most cases, I find moving closer to a single speaker and playing a mono source results in a subjective impression of the high end being a little hotter. The sound also appears to be localizable around the tweeter. The M106 does not appear to be special with respect to exhibiting these phenomena. All speakers I had access to that were equal or larger than the M106 exhibited it. If you do this test, you have to compensate for level differences. The level will increase about 3dB as you move for 9ft to 6ft.
After I did this test, which was not blind, and thus not reliable, I made quasi-anechoic measurements of the M106, with the seat moved from 5ft back to 9ft in 1ft increment. Above 1.5kHz, the limit of a quasi-anechoic measurement at 9ft, the curves were offset in level but showed almost no change in frequency response at all.
What was I hearing and was it real? My only guess is that as I moved away from the speaker the room reflections represented a larger portion of what is being presented to the ear.
I was worried Kevin Voecks would push back on my observation that I needed 9ft for the speaker to sound good. As it turns out, Revel actually has a small room where they demonstrate speakers. The tweeter to listening seat distance Revel uses was 10.5 feet.
The Revel room is 17ft x 13ft x 8ft. The speakers are placed on the short wall. Kevin Voecks gave me the speaker and listener placement used in the room for the M106. This room is slightly larger than my Small Room 1, so I am going to use the Revel room dimensions as an example of how to get excellent sound quality out of the M106 in a small room. The complete dimensions showing the placement of the speakers and chair is shown below. The dimensions are in inches.
The M106, like most speaker, wants space from the front and side walls. Kevin wrote to me that the “typical location in the small Revel room is 49″ tweeter from the front wall, 43″ tweeter from side wall”. Obviously, you do not need to use these exact numbers. The Revel dimensions are a useful starting point for other speakers in this sized room on the short wall.
Moving the speaker away for wall moves the frequency of the boundary cancellation. This occurs at low frequencies where the radiation from the speaker is an omni-directional pattern. Bounces also occur at the floor and ceiling but you cannot do much about that.
A good way to predict the frequency response changes from the wall bounces is an Excel spreadsheet written by Thomas Barefoot of Barefoot Sound. This appears to have been developed in 2003 and it is not on the Barefoot Sound website now. It is called 2D Wall Bounce Calculator. I have contacted Barefoot Sound to see if it can be restored to the website. The Barefoot spreadsheet shows a frequency response plot (who knew you could get Excel to do this) that changes dependent on the distances you enter for speaker placement.
While electronic room correction always helps below 300Hz, it was not required with this placement. Room correction can help flatten frequencies that are enhanced by the wall bounce. It can also help to fill low Q nulls that are less than 5dB. Deep, narrow, nulls that cover a small frequency (high Q), however, cannot be removed using electronic room correction. The calculator showed a low Q dip of about 6dB using the Revel setup in the Revel room. I am not showing the graph because room modes are also at play and will change shape of the complete response of the room below 300Hz. The issues of wall bounce and room modes apply to any speaker and this is not a specific issue to the M106.
I repeated the room diagram to show the distance of listening chair to the back wall 29 inches in the Revel room. Note the distance from the tweeter to the front wall is larger in the Revel setup than the seat to rear wall.
The room diagram shown above was created with the following website.
Once the distances are entered, the program will return the wall reflections above 300Hz and show where room treatment should be placed. No more mirror method with two people to find these.
You can see the rear wall bounce is closest to the listener. At 29 inches, back room treatment is not required; however, moving the seat further back will require the rear wall to be treated. Room treatment can become expensive if you try to blend it into a room used by the family. One nice thing about having a private listening room is you can use treatment panels in the “raw” state.
Another reason you do not want to be sitting very close to the back wall (or the speaker too close to the front or side walls) is this is where the room modes tend to combine to produce very large frequency response variations below 300Hz.
Reducing the distance from the speaker to the chair (not less than 9 feet as discussed above) will get you more distance from the seat to the rear wall as will reducing the spacing from the front wall. It is very hard to figure out where to place things exactly without a low cost measurement system. This is true of any speaker in any room and not something specific to the M106. A free measurements system that works with a USB mic is REW found
A $75 USB measurement mic, often used with REW, is found at the site.
For $100 you can get the microphone with a more accurate calibration file but you may have to edit the file for REW to accept it.
I would recommend the AcoustiSoft RplusD measurement system, which I use in this review; it is not expensive, but it does not have USB mic support.
The diagram below shows the dimensions entered into Barefoot wall bounce calculator. I am just using it as a drawing tool for the diagram below. It is not necessary to enter all the distances. Once some dimensions are entered, others are calculated by the program. Note this diagram is not to scale.
The speaker to speaker spacing is 6ft the Revel room with the speakers on the short wall placed well away from the sidewalls. The angle to the listener is 16 degrees (32 degrees between the speakers). The optimal value is almost twice, but this is impossible to achieve with the 13 foot wall and 10 foot speaker to chair distance. To get the ideal 30 degrees angle would require the short wall to be 16ft to maintain the 43 inch distance to the side wall distance and a 9ft distance from the speaker to listening seat. We have moved out of the realm of a small room.
The reduction did not appear to change the stereo effect compared to my large room. I note hotel rooms used at Hi Fi shows are typically at 12ft – 13ft the speakers are placed on the short wall. Since they need to leave space for the attendees to get a close look at the products, the first set of seats will be further away than 10ft. Even if you get the ideal center seat, it is very likely the angle is in the 15 degree range or less. Despite this, Hi Fi reports in magazines and online sources are filed with raves about imaging in some of these rooms.
I used (16ft * 12.5ft * 8ft) Small Room 1 to match the placement of the Revel room as closely as possible. I ended up with the tweeter 45 inches from sidewalls and tweeter 41 inches from the front wall. The front of the speaker to seat distance was 10ft. The distance from the seat to the back wall was 29 inches.
The room I have designated as my mini Hi Fi room is Room 2 (12.5ft * 15ft *8ft). The long wall is the only workable side of the room. The long front wall allows the space between the speakers to be increased, which also increases the angle between the speakers to the seat. In my case, 20 degrees, but the price you pay for placing the speakers on the long wall is high. The short wall is of this room is only 12.5ft. That is a loss of more than 4ft from the Revel room. I ended up with the distance from the tweeter to side walls of 51 inches and the tweeter to front wall of 32 inches. The front of the speaker to seat distance was 9ft.
Even after reducing the speaker to front wall spacing and using the speaker to chair distance to minimum 9ft, the chair was almost at the back wall. I only had 1ft between the seat and rear wall. Room correction below 300Hz is mandatory for good results with this placement. In addition I placed 4 inches of acoustic foam to decrease rear wall reflections given the short distance from seat to the rear wall.
Below are the in-room spatially averaged response measurements of the M106 in my two small rooms and a repeat of the large room. I am showing the response above 300Hz. Below 300Hz the room dominates the response.
I do not use Small Room 1 for music reproduction. Other objects in the room prevented setting up two speakers, but I could get one speaker to match how the Revel room is set up without obstructions. This allowed me to make some in-room measurements. To get a spatially average of two speaker placements I reversed the chair and speaker positions.
In Small Room 2, the presence of another pair of speakers prevented me from finding a suitable second spot for the M106. The averages for Small Room 2 are for only one speaker placement as a result.
Note how remarkably close the response curves are despite the fact that Small Room 1 was filled with furniture. Small Room 2, with the speaker on the long wall, does not show degraded response. My in-room plots, for vastly different rooms, show you can be confident that the speaker will sound as the designer intended if it is carefully placed.
That the M106 can achieve these close matches is a result of the careful engineering as documented in the ANSI/CEA-2034 anechoic performance discussed in the large room section of the review.
The low frequency response of the M106 in the small rooms was more extended that the large room. The -6dB frequency was around 35Hz. This is a result of room modes below 50Hz that boost the response. In the large room, without any room modes at these lower frequencies to help, the -6dB frequency was at 48Hz.
Do not get too excited about the 13Hz low end extension. These are small signal numbers. The speaker cannot produce significant SPL at frequencies well below its anechoic cutoff frequency. A speaker with a 5.25 inch woofer, in a small box, will not have an anechoic response that is extended enough to excite these modes so the small in-room low end roll-off could significantly higher.
Moving from the large room to the small room does not changes the amount of power required for the speaker to produce a given SPL level in the frequency range dominated by the speaker and not the room (above approximately 200Hz – 300Hz). The distance from the speaker to the listener is 9ft and is the same in both the large and small rooms, so the acoustical attenuation is the same. It is the distance that determines the loss from the near field output level to the level at the listening chair.
The lack of modes below 50Hz in the large room, which cost about 10Hz at the bottom end, might explain the misconception that large rooms need more power than a small room at a constant speaker to seat distance. Adding a subwoofer transfers the job of producing bass in the area below the room mode support in the large room. The subwoofer does have to work harder without the room modes. Tom Nousaine wrote a number of articles on these issues for Stereo Review. Unfortunately with his passing his website with the articles was taken down. The website can still be found using the Internet Archive.
The low end extension of the M106 in the small rooms is adequate so a subwoofer is not required if you keep the listening levels under control. The extension is now low enough to match the bottom end of most musical instruments. You can hear it with piano, double bass, bassoon etc. Unfortunately percussion instruments are also the ones that produce the peak level on many CDs (0dBFS) and the M106 cannot reproduce them with full impact if they are below the 50Hz-60Hz range.
Part 1 System Setup
The permanent resident of Small Room 2 is the discontinued NHT Xd active, tri-amped, DSP speaker system with integrated subwoofer ($6000 when it was discontinued in 2006. Pricing would be five figures given the recent trend of huge pricing increases for low volume speakers). The NHT Xd satellite section uses the very expensive Seas Excel W15 5.25 inch magnesium woofer. It is placed in a small box. The tweeter selected was small Seas neodymium unit allowing for a close hugging, contoured, baffle molded with BMC. Heat from the small tweeter is directed to the rear by an aluminum rod. This is in a polar opposite direction to using a waveguide. The DSP compensates for the baffle step introduced by the curvy baffle as well as driver frequency response errors . Both the subwoofer – satellite and satellite woofer –tweeter crossover at 8th order LR (Linkwitz–Riley ).
The most significant in-room measured difference between the NHT Xd and Revel M106 above 300Hz was at the woofer – tweeter crossover. The NHT transition occurred more quickly but without a waveguide, the depth of the notch caused by the woofer becoming directional was larger. The Revel M106 had a longer transition range with its 4th order crossover, but it was smoother despite the fact it had a larger 6.5 inch woofer which has more energy loss off axis. In the sit-down / stand-up test, the NHT Xd was the winner as a result of the 8th order crossover
The DSP crossover limits the frequency sent to the satellite to 110Hz dropping like a stone below that with an 8th order filter, but as explained in the subwoofer section above, the Xd satellite’s 5.25 inch woofer struggles from 110Hz – 200Hz. My review from 2006 is below but the figures have been deleted.
My standard room setup with the NHT Xd is with the speaker to seat distance of 7ft allowing more space from the seat to rear wall. Obviously both speakers were auditioned at the 9ft distance for comparison.
I hear you saying it is unfair to compare the Revel M106 to the NHT Xd, but I did it with unexpected results.
For listening tests, I removed the added low end extension of the NHT Xd subwoofer. I deployed a prototype Dirac analog in/out room correction system. I used the Dirac target curve which matched the low end response of the Revel M106 in this room. Dirac has a very flexible target curve shaping system. This allowed me to create a target curve that matched the roll-off to the Revel M106 to the NHT Xd speaker in the small room. The target curve was then flat to 300Hz which removes room effects that have nothing to do with the speaker’s performance. Above 300Hz I did not apply room correction. This is where the speaker dominates the sound. The same approach was used for comparative testing of speakers in the large room using the Anthem ARC room EQ. The curve set below is a 9pt spatial average of each speaker.
The first plot above is a Dirac frequency response measurement of the M106. The plot shows the before and after the electrical correction applied. The maximum frequency of the graph is set to 100Hz to allow the a detailed view of the low end response. Note how the target curve is adjusted so it closely follows the M106 roll-off. The -6dB frequency of the M106 in this small room is shown to be 35Hz as I had previously stated. The M106 is flat in anechoic measurements to its low end roll-off, but in the small room, with the chair near the rear wall, the room is supplying significant gain to 100Hz. I took out all the room gain for the listening tests in this small room.
The second plot shows the NHT Xd before and after correction. Below 100Hz this is all in the subwoofer range which is located at the center of the room direct against the front wall. The room effects are very different with this placement as you can see in the before correction curve. You can see how the target curve has forced the response to match the Revel post correction. The HPF target curve, with its very large attenuation factor from the uncorrected response, can cause dynamic range issues with Dirac filters. You should not try a HPF simulation with Dirac in normal use.
While a good room correction system can clean up excess energy it cannot do much for frequency nulls, which in this room, with short seat placement on the short wall, was a pair of closely spaced nulls centered at 130Hz. Dirac is a little aggressive trying to fill these but the M106 appeared unfazed by the +10dB push Dirac applied in a failed effort to fill the nulls.
I am not recommending the Dirac room correction over the Anthem ARC used in the large room tests. The Anthem ARC was tied up doing satellite – subwoofer integration duties in the large room. All current, affordable, Dirac enabled products, use sub-optimal bass management not developed by Dirac. Dirac does not have a bass management system that it sells to 3rd parties. The third party Holman bass management system, supplied with current Dirac products, except the Datasat, date from the 1980s and the Holman system does not produce a smooth transition from the main channel to subwoofer. Its application space is for movie sound effects. More details on bass management issues with the Holman system will appear in a separate article.
The well-known problem of Dirac, the inability to accurately normalize the level of the speakers post correction, could be worked around by doing the amplitude adjustment manually post Dirac. Anthem ARC typically gets levels correct. The frequency response curve below is presented to show how I removed the error in this Dirac run.
I took single shot, frequency response measurements, at the prime seat location, for each speaker. Below is an example plot. This has no averaging so the curves are not as smooth. You can see, the speaker responses keep crossing each other so the level adjustment is a process of trying to get the average level of the curves to be approximately the same. Using a level meter is not as accurate as looking at the response. I needed to remove more than 1dB from one channel to compensate for the Dirac level compensation error to get this match.
Note the curve starts at 400Hz. Room EQ is active below active to 300Hz and the transition occurs to 400Hz. I am not identifying the individual speakers since no averaging was used. You can see they have similar profiles confirmed in my averaged measurements of both speakers in Room 2 with the mic centered at the listening seat. The NHT may have the tweeter level set slightly higher after the crossover transition, although they converged at higher frequencies. This may be responsible for some of the subjective observations I make below.
Part 2 Listening Test Results
Listening tests were in mono with short repeated loops of the music tracks. I will have much more to say on why I did this in a follow up article on NRC listening test methodology and my modifications to allow a wider genre of music to be used during these tests. I could not do blind tests in the small room and I was obviously highly biased to the NHT given all the DSP technology it use (I have published papers on active and high order crossovers), the expensive drivers and the fact I purchased it.
I could get only one Revel correctly placed in small room 2 with the NHT Xd units taking up the rest of the space. In NRC listening tests, mono is preferred to detect errors in the anechoic, early reflection or power response. It also allows driver resonance to be more easily heard. This observation goes back to the early 80s at the NRC.
F.E. Toole, Subjective Measurements of Loudspeakers: A Comparison of Stereo and Mono Listening”, presented at the AES 74th convention, October 1983, Convention Paper 2023
I verified this along with many other protocols in the NRC controlled listening tests which will be in my article on subjective testing procedure developed at the NRC.
A speaker with excellent performance in mono should work optimally in stereo. I checked the ability of a pair of M106 to create a good stereophonic “spatial illusion” in the large room using two larger speakers in the $2000/pair price range for comparisons. In the large room, the low end response of all 3 speakers was matched and flattened to 300Hz using the Anthem ARC room EQ with no subwoofer. I can report the M106 spread the sound stage across the front wall with little localization at the speakers. The excellent pair matching of the stereo pair of M106 samples I was supplied likely enhanced the performance in stereo. While I do not go into detail about the listening tests in the large room, I uncovered nothing different from what is reported for the small room testing with the exception of the bass range where the larger speakers had lower distortion although not that much more extension.
Despite all the money in the Seas Excel W15 woofer in the NHT Xd, the M106, with its larger woofer and cabinet, had a clear advantage in the ability to deal with dynamics in the 110Hz – 250Hz range. This octave you cannot patch with a subwoofer. I have never been happy with the performance of NHT Xd, in the 100Hz – 250Hz range, at levels above 85dB SPL peak at 7ft distance. The M106 was happy to go above 90dB SPL at 9ft where the overall level is down 2dB. I suspect the M106 would go to 95dB SPL peak but I do not like listening at that level. You quickly lose the ability to make distinctions between the Devices Under Test.
The listening selections below are a small sample of what I used during the subjective comparisons in the small and large room. None, I discuss below would be considered optimal for listening tests at Harman facilities:
S. E. Olive, “A Method for Training of Listeners and Selecting Program Material for Listening Tests”, 97th Convention, Audio Eng. Soc., Preprint No. 3893 (1994 November).
The extension I developed will be described in the separate article, on controlled listening test, which I think overcomes the issue Harman has with a wide range of program material. Using a wide range of program material significantly slows down the time to evaluate the speaker because you need more of it to evaluate the full range performance of the speaker.
Extended evaluation time is not an issue for a reviewer but a significant one when doing statistical analysis of subjective comments, by a large number of test participants, each evaluating speakers individually in the specially designed Harman listening room. For the perspective purchaser, the time at a dealer is not enough. You should be able to borrow the speaker overnight to gain the time you need. Dealers should be willing to do this now, at least with mini-monitors, given the extended trial periods offered by the direct to consumer channel.
Preparation of the test material needs to be done in advance. I will describe this in the upcoming article.
I used a couple of test clips that would meet Harman’s criteria for evaluating speakers although not the selections they recommend. The clips I used were well known to me, in most cases for years, but with the new methods I was able to zero in on some of the most useful and revealing passages and dig deeper into why I found them useful. Once a short passage was identified, I clipped it out and repeated continuously per NRC procedures. Easy for me with Audacity freeware. In the 60s (yes it goes back that long), Dr. Toole writes in his text they hired a disc jockey to do the repeats.
None of the example test clips below would be considered by Harman to be ideal for speaker comparison. With my new method of using a real time spectrograph, they are excellent for subjective evaluation.
Samuel Baron “The Art of Fugue for String Quartet and Wind Quintet”Samuel Baron’s highly acclaimed arrangement of The Art of Fugue for String Quartet and Wind Quintet works better to my ears then the string quartet arrangements. The only, in print recording, was released as an mp3 by DG but the CD is available at the artist’s site.
If you have any interests in this, grab it while they still have it. This is demo quality sound while I expect the mp3 is dreadful. The performers for this recording are some of the finest chamber musicians as full members of the Chamber Music Society of Lincoln Center.
Contrapunctus 9 is scored for the full 9 players. Thank goodness a video is on the web to unscramble what the individual players are doing although the group on the video cannot approach this performance.
One question we ask of a speaker is can it present all 9 players playing together and can it do it with each instrument balanced. Samuel Baron helps this by keeping the instruments at different ranges.
In the last part of the movement, the bassoon is at its lowest register, down to 80Hz and stays in this range for the whole segment. The Revel has no problem, in the small room, reproducing the low register of the bassoon cleanly and evenly. The first violin, with a relatively close microphone placement, stays up high staying away for the middle voices. On the NHT, the 1st violin takes on an edge that appears to be centered near the top of its range in the 2kHz band. The Horn is centered in the 300Hz range but the harmonics go much higher. Once the horn comes in it dominates the middle register, as it tends to do in a woodwind quintets. The horn was a little more balanced with the rest of the group on the Revel although it was hard to determine were the emphasis was centered on the NHT.
The conclusion brings all the instruments in filling the spectrogram for 80Hz out to harmonics at 10kHz. It is impossible to make out everything with the horn dominating the middle voices until the horn pulls back at the final long chord. The spectrogram shows nine district lines spaced across the range of the instruments. The Revel will not let you hear each individual instrument, which does not even happen on headphones, but the neutrality of its presentation of this unique orchestration is impressive.
Elgar Sea PicturesThis is considered the definitive performance of the Elgar Sea Pictures. Where the Corals Lie has minimal orchestration below great Mezzo Janet Baker.
On the NHT, the sibilance are emphasized. These have significant energy in the 4kHz – 8kHz range. The notes Mezzo sings are centered in the 500Hz octave but strong harmonics are in the 2kHz – 4kHz range. It was in that range that the NHT Xd had significant emphasis.
I could not identify a specific coloration on the voice on the Revel M106. Of much significance is that the Revel stays clean even in the fortissimo passages at the highest notes in the piece. I was surprised when I looked at my SPL meter that I had set the volume level at peak level of almost 90dB (mono at 9ft). That is really too high relative to live, but the composure of the M106 must have encouraged me to push the level up. Listening at that level will quickly degrade your ability to hear accurately short term, so I backed off to 85dB SPL I used for the chamber works.
Getting away from the spectrograph, the Revel allowed me to enjoy the wonderful performance of the deep and moving Sea Pictures.
Smetana Piano Trio is at the top rank of piano trios as judged from this recording. Excellent piano trios are uncommon with the pianist of the group trying to simultaneously play solo gigs preventing the full time interaction between the players that occurs with a string quartet.
My test clip is 1.20 min in from the first movement of the Smetana. I present frequencies of the instruments fundamental and harmonics as shown on the real time spectrogram, but if you want to look at the score instead, it shown on YouTube synchronized to the music.
I am no musician so this really helps for score reading. The score was useful in confirming some left hand piano playing, at the beginning of the clip, around 60Hz fortissimo. This is repeated twice more and the left hand remains in the range for a few more measures. In the small room, the Revel had the extension to produce this at realistic levels. It was attenuated in the large room.
After this passage, things settle down for a beautifully cello passage written down in its middle range. The violin then takes the passage to a higher register moving to its upper range with harmonics expending out to 10 kHz. The recording places the microphone further away than the Bach. With this recording I was reminded why I purchased the NHT. I did not hear the colorations I identified in the other 3 recording I chose to highlight in this listening section. The Revel was a match, although the two speakers do not sound identical.
Randel Thompson “The Peaceable Kingdom”Randel Thompson is a favorite of amateur choirs unfortunately virtually unknown to the general public save his Alleluia. The greatness of the music can be appreciated only with finest professional performance as on the CD with the Schola Cantorum of Oxford. The a cappella ensemble is beautifully captured by Hyperion. So much happens in little over a minute in the Howl Ye section of the Peaceable Kingdom.
The writing is for 8 part chorus with rapid jumps between the two divisions. Only the most transparent recording will let you hear this. While some of vocal part is melisma, the words should otherwise be clear. The work builds quickly to a fortissimo passage that shows the dynamic capability of the speaker.
On the Revel, the tonality does not change and the different parts of the chorus are distinct. Across the whole movement, which runs from the bottom end of the bass (about 130Hz) to the top of the soprano lines with harmonics out to 7kHz, the Revel is neutral. By contrast, the NHT Xd is having trouble in the fortissimo pages across the whole choirs. Throughout the work, the octaves around 1kHz and 2kHz stick out with the NHT with emphasis on the upper voices. The Revel stays composed throughout this stormy little piece with clean diction and flat tonality and no sign of a change in tonality with the chorus in full cry.
I was not expecting such a significant difference between the two speakers with this recording or any recording. The combination of the ability to stay composed at high SPLs (90dB peak which is the level one would hear live row h in a small hall) and the tonal neutrality achieved by the M106 when faced with complex tasks like reproducing the 8 part chorus was responsible for this.
In the end, the tonality of the M106 was more neutral at 9ft, despite the fact that the NHT Xd was using more expensive drivers that were being powered directly from separate power amplifiers. These in turn where driven by DSP based, 8th order, crossovers.
The single power amp signal has to pass through a passive crossover in the M106.
Aside from the dynamic range restrictions in the NHT, I found the tonality shifts to be most prominent in the 1kHz – 4kHz bands. The Harman proprietary waveguide technology is likely the key to this. The transition between the woofer and tweeter is far less abrupt. Having the $2000 M106 passive speaker go toe to toe with the $6000 NHT Xd, that I own, and win, was not at all expected.
Nobody would question a DSP based active speaker should be able to sound better than a passive design but the NHT Xd is a 10 year old design done with significantly less DSP computational power than what is available today. One is left to wonder what a Revel M106 would sound like active, with a DSP crossover (we are talking a 2X price increase).
In the subwoofer integration section, that proceeded this section on small rooms, I explained how to implement an active DSP based subwoofer / satellite system for the M106 using the Anthem ARC room correction. Both the subwoofer and M106 are powered by different amplifiers so this is an active crossover. This is a more advanced approach than used in the 2006 NHT since separate room EQ is in the satellite and subwoofer channels. The crossover frequency was reduced from 110Hz to 80Hz. The crossover is reduced from 8th order to 4th order with the Anthem ARC, but the M106 does not need to be rolled off at the slope of a brick to protect the small woofer even with the reduced crossover frequency. The M106 / Anthem ARC / subwoofer combo in my large room took the prize with no ambiguity.
As much as I hate to say it my favorite speaker has been overthrown. The Revel M106, an Anthem ARC enabled product, and good subwoofer (use more for better results) now represents the pinnacle of what can be achieved with a mini-monitor using passive technology. In a small room, the M106 alone will get you down to 35Hz, the bottom end of most musical instruments, although not with the impact of a subwoofer based system below 70Hz.
THE REVEL M106 offers Bench Mark Measurements and Sound for a Two-way Passive Mini-Monitor
- Waveguide offers excellent integration between woofer and tweeter
- State of the art anechoic performance for a two way mini-monitor
- Closely matching in-room measurements for rooms large and small
- Smooth spectral balance in controlled listening tests
- Bass extension to 35Hz in a small room.
- Bass extension, compression and distortion characteristics that allow smooth, localization-free integration with a subwoofer, with the appropriate room EQ / bass management at an ideal 80Hz crossover.
- Tweeter level control reinstated from previous model to enhance performance when the room issues prevent the speaker from being well placed.
- A step up active version with DSP crossover, driver EQ and room EQ coupled with amps for each speaker.
Many audiophiles dream of being able to afford the sound of million dollar system shown in rooms on the 34th and 35th floor at Venetian Hotel during CES week. The money in these speakers is spent in areas that have not been shown to provide any measurable improvement. Prices have little correlation with the cost to produce these speakers. Add to this the fact many of these speakers have very poor measured results. The Consumer Electronic Association is working to unmask the pretenders with Standard ANSI/CEA-2034 specifying specific measurements to be shown to consumer.
On the 34th and 35th floor, giant arrays of woofer top and bottom are attached to some products that drive up price but it is never clear if the performance of a speaker configured this way matches what can be archived with best subwoofers and appropriate room EQ and bass management. Once you get past the giant woofers, all you see are coned midrange and tweeter connected with passive crossovers.
The simple fact is the Revel M106 exceeds the measured performance of almost all these speakers within it bandwidth limit. Subjective evaluation, using techniques to enhance perception of spectral imbalance, correlated with the measurements I presented in this review. The M106 provide bass extension to the lowest frequency of most musical instrument in a small room. Combining the M106 with excellent room EQ below 300Hz and one or more subwoofers provides near state of the art performance.
The Revel F208 large floor standing speaker does produce better anechoic and in-room curves above 1.5kHz – 8KHz with significant bass extension. It is an alternative if you have the room and the extra cash. In most cases, the F208 would not need a subwoofer. To significantly advance the objective and subjective performance beyond the Performa3, requires amplifiers directly connected to each speaker and advanced DSP processing.
You do not have to take my word for it. If you are seriously considering purchasing the small M106 at an authorized Revel dealer, they should let you take it home for an evening as a trial period.
The author would like to thank Robert Kozel for reviewing this article.