For people interested in room correction: recently, we did some controlled listening tests and objective measurements on different room correction products, and presented the results at the recent 127th AES convention in NYC.
I've summarized the results in my recent blog posting entitled, "The Subjective and Objective Evaluation of Room Correction Products". Make sure you look at the AES presentation that accompanies the blog posting.
Cheers
Sean Olive
Audio Musings
Full Version: Listening Tests on Room Correction Products
Thank you, I'm interested. Will read.
Thanks for the test and detailed report. I wish the devices were identified because I was just about to buy an Onkyo receiver (has Audyssey RC) but I don't want it if it's one of the units with equal or worse performance to no EQ at all.
Excellent post, Sean. I'm afraid I neither had the time to catch this session at the convention, nor got the chance to actually meet you - every time I saw you, it looked like you were talking to somebody more important than me 
Would the observation that a flat in-room response is not preferred be possibly explained by the "circle of confusion", or does it represent any more of a fundamental truth about ideal reproduction? That is, if the circle of confusion is reduced, does this non-flat preference go away?
Not really related to this specific paper, but: Could the observation that listeners are far more sensitive to response defects with mono playback imply that increasing the number of reproduction channels is a valid way of improving sound quality, orthogonal to improving room response itself?
Would the observation that a flat in-room response is not preferred be possibly explained by the "circle of confusion", or does it represent any more of a fundamental truth about ideal reproduction? That is, if the circle of confusion is reduced, does this non-flat preference go away?
Not really related to this specific paper, but: Could the observation that listeners are far more sensitive to response defects with mono playback imply that increasing the number of reproduction channels is a valid way of improving sound quality, orthogonal to improving room response itself?
[url="http://seanolive.blogspot.com"]
Thanks Axon. Next time you see me at AES in conversation with someone, make sure you interrupt me since no one is "more important" to me than a Hydrogenaudio subscriber.
It seems plausible that the preferred non-flat target function could be explained by the circle of confusion, if recordings are monitored through a similar target function. If tomorrow studios started equalizing their control room monitors to produce a flat in-room response, an interesting question is whether they would add more bass to their mixes to compensate for the missing room gain? Of course, if they mix over Yamaha NS10s (which have no bass) they might already be doing that.
I don't think increasing the number of loudspeakers will necessarily compensate for poor spectral quality. Remember that as soon as you have a vocal or talking head panned hard center, you are essentially back to a single mono speaker situation. It's best to just buy 7 good loudspeakers. They don't cost any more than the bad ones, and in many cases, they actually cost much less.
Cheers
Sean Olive
QUOTE (Axon @ Nov 2 2009, 14:19) 
Excellent post, Sean. I'm afraid I neither had the time to catch this session at the convention, nor got the chance to actually meet you - every time I saw you, it looked like you were talking to somebody more important than me
Would the observation that a flat in-room response is not preferred be possibly explained by the "circle of confusion", or does it represent any more of a fundamental truth about ideal reproduction? That is, if the circle of confusion is reduced, does this non-flat preference go away?
Not really related to this specific paper, but: Could the observation that listeners are far more sensitive to response defects with mono playback imply that increasing the number of reproduction channels is a valid way of improving sound quality, orthogonal to improving room response itself?
Would the observation that a flat in-room response is not preferred be possibly explained by the "circle of confusion", or does it represent any more of a fundamental truth about ideal reproduction? That is, if the circle of confusion is reduced, does this non-flat preference go away?
Not really related to this specific paper, but: Could the observation that listeners are far more sensitive to response defects with mono playback imply that increasing the number of reproduction channels is a valid way of improving sound quality, orthogonal to improving room response itself?
Thanks Axon. Next time you see me at AES in conversation with someone, make sure you interrupt me since no one is "more important" to me than a Hydrogenaudio subscriber.
It seems plausible that the preferred non-flat target function could be explained by the circle of confusion, if recordings are monitored through a similar target function. If tomorrow studios started equalizing their control room monitors to produce a flat in-room response, an interesting question is whether they would add more bass to their mixes to compensate for the missing room gain? Of course, if they mix over Yamaha NS10s (which have no bass) they might already be doing that.
I don't think increasing the number of loudspeakers will necessarily compensate for poor spectral quality. Remember that as soon as you have a vocal or talking head panned hard center, you are essentially back to a single mono speaker situation. It's best to just buy 7 good loudspeakers. They don't cost any more than the bad ones, and in many cases, they actually cost much less.
Cheers
Sean Olive
To Sean Olive: You have a great blog. It's refreshing to observe audio people thinking and experimenting, rather than listening to "Fantasy Audiophiles" with their "I know what I hear and couldn't be wrong" methodology!
You wrote: "I've summarized the results in my recent blog posting entitled, The Subjective and Objective Evaluation of Room Correction Products"
My Two Cents on that AES slide show (I assume the slides were the presentation):
(1) Sound absorbers are the only "room correction products". The devices you compared can't correct a whole room, which becomes obvious when you compared results at one listening position with the average of six listening positions. Any electronic device that can do a good job equalizing one listening position will be compromised when set using averaged measurements from multiple seats. A real listener sits in one seat, not in six seats!
(2) I believe you may have started out with unusually good room acoustics, where smaller and less audible corrections were needed (compared with a more typical 20' x 15' x 8' room with no acoustic treatments)
(3) What data leads to your conclusion that 'Flat in-room response in not the preferred target' ? Did any of the listeners actually hear flat in-room response? If no one heard a flat frequency response, how can you be sure they wouldn't like it? ... I think most audiophiles would like flat bass and midrange frequency response measured at the listening position -- that's a theoretical ideal. Where in your data do listeners specifically say they do not like a flat bass and mid-range frequency response?
Based on my own experiences with equalizers since 1971, most listeners want some degree of treble roll off -- flat treble frequency response measured at the listening position sounds too bright unless the listener has hearing damage. In addition, at low and moderate volumes some bass boost under 50Hz. may be desired. In plain English: If a flat frequency response from 20Hz. to 20kHz. WAS auditioned, I believe most listeners would have reported the sound was too bright. And at modest SPL's, many listeners might have preferred louder deep bass. The preferred treble roll off would probably vary among listeners, and among listening rooms. I don't know if an average of these treble roll-off preferences, or one "house curve" for a specific listening room, would satify most listeners (because individual high frequency hearing ability, especially among males over 40 years-old, varies so much).
Once again I want to thank you (and Floyd Toole, and Sigmund Linkwitz, and others) for sharing so much audio research and knowledge for free.
You wrote: "I've summarized the results in my recent blog posting entitled, The Subjective and Objective Evaluation of Room Correction Products"
My Two Cents on that AES slide show (I assume the slides were the presentation):
(1) Sound absorbers are the only "room correction products". The devices you compared can't correct a whole room, which becomes obvious when you compared results at one listening position with the average of six listening positions. Any electronic device that can do a good job equalizing one listening position will be compromised when set using averaged measurements from multiple seats. A real listener sits in one seat, not in six seats!
(2) I believe you may have started out with unusually good room acoustics, where smaller and less audible corrections were needed (compared with a more typical 20' x 15' x 8' room with no acoustic treatments)
(3) What data leads to your conclusion that 'Flat in-room response in not the preferred target' ? Did any of the listeners actually hear flat in-room response? If no one heard a flat frequency response, how can you be sure they wouldn't like it? ... I think most audiophiles would like flat bass and midrange frequency response measured at the listening position -- that's a theoretical ideal. Where in your data do listeners specifically say they do not like a flat bass and mid-range frequency response?
Based on my own experiences with equalizers since 1971, most listeners want some degree of treble roll off -- flat treble frequency response measured at the listening position sounds too bright unless the listener has hearing damage. In addition, at low and moderate volumes some bass boost under 50Hz. may be desired. In plain English: If a flat frequency response from 20Hz. to 20kHz. WAS auditioned, I believe most listeners would have reported the sound was too bright. And at modest SPL's, many listeners might have preferred louder deep bass. The preferred treble roll off would probably vary among listeners, and among listening rooms. I don't know if an average of these treble roll-off preferences, or one "house curve" for a specific listening room, would satify most listeners (because individual high frequency hearing ability, especially among males over 40 years-old, varies so much).
Once again I want to thank you (and Floyd Toole, and Sigmund Linkwitz, and others) for sharing so much audio research and knowledge for free.
QUOTE (solive @ Nov 1 2009, 21:38)
For people interested in room correction: recently, we did some controlled listening tests and objective measurements on different room correction products, and presented the results at the recent 127th AES convention in NYC.
I've summarized the results in my recent blog posting
I've summarized the results in my recent blog posting
I am in agreement with your reference in the blog to "so called" room correction, as it is anything but the room being "corrected". The correction is taking place at the acoustic source(s) (loudspeakers) which results in a change in the soundfield created by their interaction with the room. The two cannot be separated.
Perhaps "listener room/loudspeaker response correction" might be more appropriate?
Interesting choice of music tracks and loudspeakers. Perhaps selected by marketing dept so that the Harman processor will be aimed at the largest group of perspective customers?
It would still be interesting (maybe not from a financial/marketing perspective) to see how a speaker system with the polar/power response like the Gradient Revolution or even a JBL Control 30 fared under the same circumstances.
Or even outside the Harman room, in a more typical, untreated living room with furnishings.
cheers,
AJ
QUOTE (Richard Greene @ Nov 3 2009, 13:14)
(1) Sound absorbers are the only "room correction products".
Once again I want to thank you (and Floyd Toole, and Sigmund Linkwitz, and others) for sharing so much audio research and knowledge for free.
Once again I want to thank you (and Floyd Toole, and Sigmund Linkwitz, and others) for sharing so much audio research and knowledge for free.
(1) I would refer to those a "speaker deficiency correction products", due mainly to (poor) polar response and/or positioning.
p.s.sst Hey RBNG, it Siegfried Linkwitz. Sigmund is the guy needed over at AA
QUOTE (ajinfla @ Nov 4 2009, 01:16)
QUOTE (solive @ Nov 1 2009, 21:38)
For people interested in room correction: recently, we did some controlled listening tests and objective measurements on different room correction products, and presented the results at the recent 127th AES convention in NYC.
I've summarized the results in my recent blog posting
I've summarized the results in my recent blog posting
I am in agreement with your reference in the blog to "so called" room correction, as it is anything but the room being "corrected". The correction is taking place at the acoustic source(s) (loudspeakers) which results in a change in the soundfield created by their interaction with the room. The two cannot be separated.
Perhaps "listener room/loudspeaker response correction" might be more appropriate?
Interesting choice of music tracks and loudspeakers. Perhaps selected by marketing dept so that the Harman processor will be aimed at the largest group of perspective customers?
It would still be interesting (maybe not from a financial/marketing perspective) to see how a speaker system with the polar/power response like the Gradient Revolution or even a JBL Control 30 fared under the same circumstances.
Or even outside the Harman room, in a more typical, untreated living room with furnishings.
cheers,
AJ
Let me get this straight, If I make a change (correction) in my *room* that affects the acoustic propperties of *the room* I have actually not made a change (correction) in the room but I have made a change (correction) in the speaker? "anything but the room being "corrected". The correction is taking place at the acoustic source(s) (loudspeakers)"
I'm pretty confident that when I make a change (correction)in my room I am not making any change (correction) in my speakers or any other speaker anywhere in the world that I may some day put in my room.
As a footnote I want to point out why I state "change (correction)." any addition of acoustic treament in the most literal sense is simply a change in the acoustic proppeties of that room. They are a "correction" or not based on whether or not the change is a desireable one or not. So room corrections are simply a subset of room changes. when arguing against what appears to be a semantical argument i don't want to be guilty of the same.
Fascinating stuff, Sean,and I regret even more now not being able to attend more days of AES in NYC. But these were 'high end' products (even the Audyssey was the standalone unit version), and the Harman schemes are still in development. I'd love to see the same sort of comparison include the digital room correction schemes found in mass market AVRs -- i.e., Denon/Marantz/Onkyo's Audyssey flavors, Pioneer's MCACC, Yamaha's YPAO. These are what most consumers are going to encounter if they encounter DRC at all.
QUOTE (Richard Greene @ Nov 3 2009, 09:14)
To Sean Olive: You have a great blog. It's refreshing to observe audio people thinking and experimenting, rather than listening to "Fantasy Audiophiles" with their "I know what I hear and couldn't be wrong" methodology!
You wrote: "I've summarized the results in my recent blog posting entitled, The Subjective and Objective Evaluation of Room Correction Products"
My Two Cents on that AES slide show (I assume the slides were the presentation):
(1) Sound absorbers are the only "room correction products". The devices you compared can't correct a whole room, which becomes obvious when you compared results at one listening position with the average of six listening positions. Any electronic device that can do a good job equalizing one listening position will be compromised when set using averaged measurements from multiple seats. A real listener sits in one seat, not in six seats!
(2) I believe you may have started out with unusually good room acoustics, where smaller and less audible corrections were needed (compared with a more typical 20' x 15' x 8' room with no acoustic treatments)
(3) What data leads to your conclusion that 'Flat in-room response in not the preferred target' ? Did any of the listeners actually hear flat in-room response? If no one heard a flat frequency response, how can you be sure they wouldn't like it? ... I think most audiophiles would like flat bass and midrange frequency response measured at the listening position -- that's a theoretical ideal. Where in your data do listeners specifically say they do not like a flat bass and mid-range frequency response?
Based on my own experiences with equalizers since 1971, most listeners want some degree of treble roll off -- flat treble frequency response measured at the listening position sounds too bright unless the listener has hearing damage. In addition, at low and moderate volumes some bass boost under 50Hz. may be desired. In plain English: If a flat frequency response from 20Hz. to 20kHz. WAS auditioned, I believe most listeners would have reported the sound was too bright. And at modest SPL's, many listeners might have preferred louder deep bass. The preferred treble roll off would probably vary among listeners, and among listening rooms. I don't know if an average of these treble roll-off preferences, or one "house curve" for a specific listening room, would satify most listeners (because individual high frequency hearing ability, especially among males over 40 years-old, varies so much).
Once again I want to thank you (and Floyd Toole, and Sigmund Linkwitz, and others) for sharing so much audio research and knowledge for free.
You wrote: "I've summarized the results in my recent blog posting entitled, The Subjective and Objective Evaluation of Room Correction Products"
My Two Cents on that AES slide show (I assume the slides were the presentation):
(1) Sound absorbers are the only "room correction products". The devices you compared can't correct a whole room, which becomes obvious when you compared results at one listening position with the average of six listening positions. Any electronic device that can do a good job equalizing one listening position will be compromised when set using averaged measurements from multiple seats. A real listener sits in one seat, not in six seats!
(2) I believe you may have started out with unusually good room acoustics, where smaller and less audible corrections were needed (compared with a more typical 20' x 15' x 8' room with no acoustic treatments)
(3) What data leads to your conclusion that 'Flat in-room response in not the preferred target' ? Did any of the listeners actually hear flat in-room response? If no one heard a flat frequency response, how can you be sure they wouldn't like it? ... I think most audiophiles would like flat bass and midrange frequency response measured at the listening position -- that's a theoretical ideal. Where in your data do listeners specifically say they do not like a flat bass and mid-range frequency response?
Based on my own experiences with equalizers since 1971, most listeners want some degree of treble roll off -- flat treble frequency response measured at the listening position sounds too bright unless the listener has hearing damage. In addition, at low and moderate volumes some bass boost under 50Hz. may be desired. In plain English: If a flat frequency response from 20Hz. to 20kHz. WAS auditioned, I believe most listeners would have reported the sound was too bright. And at modest SPL's, many listeners might have preferred louder deep bass. The preferred treble roll off would probably vary among listeners, and among listening rooms. I don't know if an average of these treble roll-off preferences, or one "house curve" for a specific listening room, would satify most listeners (because individual high frequency hearing ability, especially among males over 40 years-old, varies so much).
Once again I want to thank you (and Floyd Toole, and Sigmund Linkwitz, and others) for sharing so much audio research and knowledge for free.
Hi Richard,
Thanks for your comments and feedback.
(1) I didn't invent the term "room correction" - the industry did - and I agree that it's an unfortunate term - probably invented in a marketing department - that is a misnomer, and causes a lot of confusion. Nonetheless I used the term because it defines a product category that we tested. That's why I wrote "so-called" room correction.
I agree with you entirely on the points you make here.
(2) One of the best things about our room is the lack of stiffness of the inner walls: we used a single layer of 1/2-inch drywall on 3 5/8-inch metal studs, which provides good damping of room modes. Our standard playback system in this room uses 4 subwoofers in the corners to cancel odd-order modes (we call this Sound Field Management), and there is very little "room correction" needed below 100 Hz. In the room correction study, we only had 1 subwoofer so there was more to correct. We have another listening room in Germany with 2-layers of drywall and had to purchase some RPG bass absorbers because the multiple subs and room correction could still not produce sufficient low seat-seat variance that we were after.
(3) When I say the listeners didn't prefer a "flat in-room response" I refer to the steady-state measurements spatially averaged around the primary listening seat. This measurement contains a summation of the direct and reflected sounds radiated by the loudspeaker. If you compare the in-room responses of the room corrections (slide 24) the more preferred room corrections (e.g. RC1) had slopes that were tilted down compared to say RC5 and RC6 which had an in-room response closer to flat. We can argue that RC5 wasn't perfectly flat, but the trend is certainly apparent. Also, we have tested RC1 (ours) with different target slopes that included a truly flat version, and it sounded too bright and thin.
If we measured ONLY the direct sound of the loudspeaker at the listening position (or anechoically after EQ) it would indeed be closer to flat, than the in-room target response indicates in slide 24: remember that the in-room target response integrates both the direct and reflected sounds and looks more like the anechoic sound power response of the loudspeaker (with some LF room gain) than its on-axis response.
I think our experimental results tend to agree with your real-world experiences of loudspeakers in rooms.
Cheers
Sean Olive
Director of Acoustic Research, R&D Group
Harman International
QUOTE (krabapple @ Nov 4 2009, 09:31)
Fascinating stuff, Sean,and I regret even more now not being able to attend more days of AES in NYC. But these were 'high end' products (even the Audyssey was the standalone unit version), and the Harman schemes are still in development. I'd love to see the same sort of comparison include the digital room correction schemes found in mass market AVRs -- i.e., Denon/Marantz/Onkyo's Audyssey flavors, Pioneer's MCACC, Yamaha's YPAO. These are what most consumers are going to encounter if they encounter DRC at all.
I'm working on it. Be patient
QUOTE (solive @ Nov 5 2009, 11:50)
QUOTE (krabapple @ Nov 4 2009, 09:31)
I'd love to see the same sort of comparison include the digital room correction schemes found in mass market AVRs -- i.e., Denon/Marantz/Onkyo's Audyssey flavors, Pioneer's MCACC, Yamaha's YPAO. These are what most consumers are going to encounter if they encounter DRC at all.
I'm working on it. Be patient
That would be Great! Of course, you'll include EzSet/EQ and/or EzSet/EQ II? Please update us on that process here.
Off topic:
BTW, Is it possible for you to briefly go over some improvements of EzSet/EQ II as implemented in the AVR 7550HD?
Thanks!
QUOTE (solive @ Nov 1 2009, 19:38)
For people interested in room correction: recently, we did some controlled listening tests and objective measurements on different room correction products, and presented the results at the recent 127th AES convention in NYC.
I've summarized the results in my recent blog posting entitled, "The Subjective and Objective Evaluation of Room Correction Products". Make sure you look at the AES presentation that accompanies the blog posting.
Cheers
Sean Olive
Audio Musings
I've summarized the results in my recent blog posting entitled, "The Subjective and Objective Evaluation of Room Correction Products". Make sure you look at the AES presentation that accompanies the blog posting.
Cheers
Sean Olive
Audio Musings
Very nicely done, Sean.
I've tried equalization for a few years now, with active equalization (eq at line level) and honestly it's always sounded to me like someone put some cloth on some keys of a piano and not others. I always end up going to no eq.
based on my observations of "audiophiles" as they evaluate a system, i've always seen that they tend to brighten perceptibly when listening to a system with slightly exaggerated bass with very low extension and reduced room mode effect; and also when listening to a very clear midrange with slightly rolled off highs.
Much of this can be obtained with some simple active crossovers and a subwoofer or two (preferably two), without the need for active eq, which in my opinion tends to color more than it solves.
thanks again for your good work.
Just Joe
QUOTE (analog scott @ Nov 4 2009, 12:20)
Let me get this straight, If I make a change (correction) in my *room* that affects the acoustic propperties of *the room* I have actually not made a change (correction) in the room but I have made a change (correction) in the speaker? "anything but the room being "corrected". The correction is taking place at the acoustic source(s) (loudspeakers)"
The changes that are being described are applied in the electronic domain, not the acoustical domain. The room is in the acoustical domain, not the electronic domain. Since the electronics may or may not be in the room without changing their effects on the performance of the complete audio system, it is reasonable to exclude changes to the electronics from what we call changes to the room.
QUOTE (Arnold B. Krueger @ Nov 10 2009, 05:38)
QUOTE (analog scott @ Nov 4 2009, 12:20)
Let me get this straight, If I make a change (correction) in my *room* that affects the acoustic propperties of *the room* I have actually not made a change (correction) in the room but I have made a change (correction) in the speaker? "anything but the room being "corrected". The correction is taking place at the acoustic source(s) (loudspeakers)"
The changes that are being described are applied in the electronic domain, not the acoustical domain. The room is in the acoustical domain, not the electronic domain. Since the electronics may or may not be in the room without changing their effects on the performance of the complete audio system, it is reasonable to exclude changes to the electronics from what we call changes to the room.
Thanks Arny. For some reason I thought acoustic treatments were included in the tests. My mistake.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.