Listening Tests on Room Correction Products

[tonmeister86]

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" see http://seanolive.blogspot.com/2009/11/subjective-and-objective-evaluation-of.html

 

Make sure you look at the AES presentation that accompanies the blog posting.

 

 

[TimDH]

Thank you for this very interesting post. I was wondering if you can comment on whether all of the room correction products claim to be phase aligning across frequency, or whether some just boost/low the ampitude response without trying to improve the phase response?

 

[kana813]

How can you test Room Correction Products without including the industry leader TacT Audio?

 

Did you use the same measurement mic on all the products?

 

 

 

[tonmeister86]

Since we used a subwoofer and satellite speaker at different distances from the seats, all of the room corrections attempted to make the time arrivals the same by adding a time delay to the satellite. This makes it easier to do proper summation of the subwoofer and satellite speaker.

 

Some of the room correction products we tested performed equalization using IIR filters, whereas another product we tested used FIR filters, which are linear phase. The differences are mostly academic when it comes to equalizing loudspeakers in rooms.

 

Most loudspeakers are minimum phase over much of their operating range so correcting the amplitude response will generally improve the phase (the time and frequency domains are mathematically related according to the Fourier theorem). Moreover, humans are relatively insensitive to monaural phase differences, and can tolerate rather large phase distortions and group delays before they become a problem. This is particularly true when listening to continuous music in rooms where the room reflections scramble the phase of sounds produced by the loudspeaker. If we were sensitive to these phase differences, listening to loudspeakers in rooms would drive us crazy since moving your head a few inches causes a huge phase shift. Also, if you look at the scientific literature on loudspeaker tests (Toole, Klippel,etc) you will find that the frequency response (and directivity) is by far, the best predictor of sound quality. Phase response is a poor predictor of sound quality.

 

 

[tonmeister86]

I didn't realize TacT Audio was the industry leader in room correction. Is that based on total units of sales or some other criteria?

 

For the calibrations of each product we used the microphone supplied by the room correction manufacturer.

 

For the objective measurements shown in the results, we used the same 6 calibrated microphones located at the listening seats: their positions were held constant while we measured each room correction device.

 

Cheers

Sean

http://seanolive.blogspot.com

 

[kana813]

If you don't know about TacT Audio(mulitple TAS Golden Ear Awards)verses the models you tested, you obviously don't know much about the industry.

 

Why didn't you included any PC based room correction systems like http://www.acourate.com/ in your testing?

 

Did you check the calibration and on/off axis response of the supplied microphones?

 

Unless you used the same high quality calibrated measurement microphone for all units under test, your measurements are meaningless.

 

[The Computer Audiophile]

Kana - What's going on here? Sean let's everyone know about listening tests and measurements that were conducted and presented at the AES convention and you are jumping all over him because he didn't do what you think he should have done?

 

I think you missed the point of his comments about TacT. Suggesting Sean doesn't know much about the industry and his measurements might be meaningless are preposterous. This is the kind of language that drives good people away from forums and deprives everyone of much needed insight.

 

Please show some respect to people trying to do good work and advance our wonderful hobby.

 

[Thomas J]

Chris, Sean's presentation was conducted under the banner of Harman International's R&D Group for whom Sean is an employee. I think we can assume he's plugged into the industry. Perhaps Sean can speak to this point.

 

[Audio_ELF]

kana asked... "How can you test Room Correction Products without including the industry leader TacT Audio?"

Isn't the (tested) Lyngdorf unit essentially the same as the TacT?

 

Eloise

 

[TimDH]

Sean,

 

I read through your interesting presentation and your response to my question. Thank you for great information and please don't take one reader's off-base comments as your reward for a service to the rest of us.

 

[tonmeister86]

I know TacT very well including the original Danish founder who worked at Harman. We just didn't include their product in this study: don't take it personally or as an offense to the company. There were several other models we didn't test as well. Our selection was based on what was available at the time, and what we perceived as being our main competitors in this particular product category, which didn't include PC-based correction systems (actually, all but one of these products uses a PC for the measurement and filter calculation process).

 

The microphone(s) supplied with each product were calibrated, and did not deviate significantly from the calibrated ones we used for our measurements. If you look at the measurements we did after room calibration (slide 24 in the link below) you see the room corrections differed most below 100 Hz and around 2 kHz, but were otherwise very similar. That suggests that the microphones supplied with the products were operating as expected.

http://docs.google.com/fileview?id=0B97zTRsdcJTfY2U4ODhiZmUtNDEyNC00ZDcyLWEzZTAtMGJiODQ1ZTUxMGQ4&hl=en

 

[tonmeister86]

Thanks TimDH. Don't worry, I am not that easily offended. If I were, I wouldn't have chosen a career in the field of audio.

 

[Audio_ELF]

Sean ... are you able to reveal which product was which in your final "results" table?

 

Eloise

 

[Nicholas.Bedworth]

Exactly... in a field where strong opinions are sometimes inversely related to knowledge and experience, one has to maintain a constructive and educational attitude. Things work out better that way

 

 

[kana813]

Chris-

 

Do you understand how these room correction units work?

 

Please read the owners manuals.

 

Basic RC operation: take a measurement and apply a target curve or "voicing filters" and calculate a correction curve.

 

None of the units tested were using the same measurements, target curves or "voicing filters," as Lyngdorf Audio calls them.

 

The Anthem Statement D1, the Audyssey Room Equalizer, the Lyngdorf DPA1 don’t allow users to modify or import target curves or "voicing filters." God only knows how the Harman prototype products work.

 

So, the listeners in this test didn't pick which room correction system sounded the best, they picked which target curve or "voicing filter," sounded the best.

 

For this test to have any meaning, all of the units should have been tested with the same measurement and target curve or "voicing filter."

 

Just because the findings were presented at the AES convention, doesn't mean they're worth anything.

 

FYI- the TacT RCS gives users hundreds of target curves to select from and the ability to modify/design and import their own target curves. It also allows for variable room correction (I only use RC from 500 hz down) and has more HP/LP filter selections.

 

The best of the bunch, the Lyngdorf Audio RoomPerfect™ system only offers seven fixed "voicing filters" and six filter slopes, so no Eloise, it's not the same as the TacT RCS.

 

Based on the test system block diagram, all of these units were operated in their analog mode. Why would a computer audiophile want to add an extra AD/DA unit to their playback system?

 

For a computer audiophile, digital room correction should take place

before their DAC...less conversions, boxes and cables. That's why I

asked our friend from the Harman International's R&D Group, why they didn't include any PC based room correction systems in their test.

 

IMO, Harman International's R&D Group isn't "advancing our wonderful hobby," they're just trying to justify the need for a new product.

 

Aloha,

 

Dan

 

 

 

 

 

[Nicholas.Bedworth]

That a large-scale manufacturer was willing to share the internal workings of their consumer testing procedures is truly remarkable, IMO. Reaching out to the audiophile community, to help educate us as well as get our feedback, is something that few companies of HK's size do. Let's make friends with these guys!

 

My only criticism of their research is that the differentiating qualities, to me, seemed vague, compared to the kind of subjective listening schemas that we're used to in the high-end audio press. But if you think about it, for the subjects tested, even these general criteria may have been challenging.

 

We can certainly help HK with more specific evaluation criteria, not that they don't know this stuff already, but we can give them a feel for priorities in our marketplace.

 

And there's always the next time.

 

Instead of clobbering them, how about all of us asking HK if they would do the same testing with audiophiles, audio journalists off the record, and so forth? HK surely is interested in getting knowledgeable feedback, and they make products all across the board in terms of quality, price-point, etc.

 

So, I have no problem with their methodology, keeping in mind that they're selling to the broader audio market, and at least they're trying out a variety of alternatives to see what mainstream consumers really like, what they can differentiate, etc.

 

And that the HK participated in our forum in the first place is definitely a very good thing; I suspect that the feedback will be given a great deal of consideration by their product people.

 

Just remember, we're way out there in terms of being a tiny niche segment of the audio world, which in turn is a tiny speck compared to the TV world. The world doesn't revolve around us, but if we can give HK constructive feedback, that's the way to get our inputs woven into their product planning.

 

 

 

[Nicholas.Bedworth]

To get a better context within which to understand room EQ, please read the materials at http://www.aes.org/sections/pnw/ppt.htm

 

Room EQ is way more subtle than just evening out with the measureable amplitude and phase data. That's an essentially mechanical approach, which is only part of the answer. You also have to figure out which of the measurements is meaningful, psychoacoustically, and then rank them according to various priorities and criteria. And what you can do in electronics, let along in software, is way beyond what can be done passively. Sure, it's always good to eliminate major issues in the room to start with, but you can only take tube traps and panels so far. They're crude, mechanical interventions, and of course software, and EX, usually operate in transform domain spaces where all kinds of counter-intuitive things are the norm.

 

My initial experimentation with software-based room EQ is really encouraging. As you mentioned, doing the EQ before the DAC has some advantages, namely eliminating another layer of A/D, D/A, cabling, complexity, etc. And from what I can tell, the room EQ in software may be similar in sophistication to what you find in some stand-alone units.

 

Sending the DAC bit-perfect data may well give better results from the DAC stage, but I always found that the extra A/D/D/A layer caused a dulling of the sound. This was a few years ago, and probably current gear is better in this respect.

 

Anyway, with a garden-variety dual-core laptop, the EQ eats perhaps 3-5% CPU. Big deal, especially when today's Costco PC has 4 cores, hyperthreaded to 8, with 8 GB of RAM. There's enormous processing power available on today's clients.

 

Even on a "casual listeinng" system in my office, the free room EQ in Win 7 and Vista makes a remarkable improvement. Price is right, too: Zero.

 

[Nicholas.Bedworth]

Another reader pointed me to the interesting and provocative presentations by James Johnston, as in JJ. He did the Vista/Win 7 audio EQ software, and is someone Sean probably knows, at least by reputation. He's presently Chief Scientist with DTS, a company that knows a bit about audio

 

Here's the link again for everyone's benefit. There are several interesting PowerPoints here that may be helpful. http://www.aes.org/sections/pnw/ppt.htm

 

JJ deals in an integrated fashion with the engineering and psychoacoustical issues. Basically, there are many measureable defects one could correct, but in some cases, it may be better to leave them alone. In his approach, what's targeted are the factors that people hear, and he attempts to assign priorities to them as well.

 

It's an interesting perspective.

 

[Audio_ELF]

Kana said... "FYI- the TacT RCS gives users hundreds of target curves to select from and the ability to modify/design and import their own target curves. It also allows for variable room correction (I only use RC from 500 hz down) and has more HP/LP filter selections.

 

The best of the bunch, the Lyngdorf Audio RoomPerfect™ system only offers seven fixed "voicing filters" and six filter slopes, so no Eloise, it's not the same as the TacT RCS."

 

By "importing their own target curves" ... is this what you mean?

http://www.lyngdorf.com/index.php?option=com_content&task=view&id=163&Itemid=98

 

As for knowing how room correction works - this document -- http://www.lyngdorf.com/downloads/product_descriptions/roomperfect_productdescription_english.pdf -- off Lyngdorf's website may be of use to some people.

 

Eloise

 

PS. it was a genuine question as to if the TacT and Lyngdorf were essentially the same as there were the same company for a long time. I do agree that avoiding an additional AD and DA stage would be preferable.

 

[tonmeister86]

Hi Nicholas,

Thanks for your comments, support, and feedback. We do invite journalists, audio dealers, students, etc to our labs to participate in tests from time to time. Last Friday I ran 84 music students visiting from Japan through a series of tests to determine if their sound quality preferences are different from their American peers. For me, these tests provide an opportunity to determine whether the responses of our trained listeners are too far off base from how consumers respond to sound quality differences. Some of the differences have been discussed here:

http://seanolive.blogspot.com/2008/12/part-2-differences-in-performances-of.html

 

I am interested in your comments and suggestions about how we might improve our listening test methodology, particularly the scales or questions we ask. In the room correction listening test, we limited the listener responses to preference, spectral balance and comments. To ask more, would have made the task quite difficult for the listeners, and possibly adversely affected their performance. We didn't ask listeners to rate non-linear distortion and spatial attributes because I don't believe they were significant factors in these particular tests. The RC devices were not operated outside their range of linear operation, so there were no audible measurable differences in distortion. Any spatial effects were related to spectral differences among the RC devices since the tests were done with a single loudspeaker/subwoofer.

 

 

[BobH]

So, firstly a big thank you to Sean Olive for the post - nice one Sir!

 

@Eloise - I think the answer to your question is that no, they are not really the same beast. I think the distinction is that the Lyndorf will allow you, with the use of their new tool, to select from a number of pre-defined presets, whereas the Tact will allow you to make your own.

 

On my unit, once the room measurements have been taken, the software shows you a graph of the response plot and you can drag that line about any which way you want and save out the result. So I can have the unit recreate any curve I like. So I think the distinction really lies in the approach, the Lyngdorf holds your hand and the Tact lets you get on with it and do what you want!

 

@Sean - you are absolutely correct in your careful choice of questions! It isn't necessary for the listeners to have an understanding of the physics of room correction but simply to report on whether or not they liked what you did! I, on the other hand, would love to know all of your dirty little secrets! My Tact is recently acquired and any little hints and tips you may have for making sure I'm getting a good working baseline of measurements would be very gratefully received!

 

Keep it coming!

 

[Nyal Mellor]

I'm working on a set of blog posts ("articles" - my word not his - Editor) (still in research mode) about how to manually set up room correction on a Mac. This will be about hand dialling (sic) EQ, rather than relying on an automated room correction system. [content removed - Editor]

 

The articles are going to cover:

 

- some acoustics (above the transition frequency vs below)

- typical acoustical problems in a listening environment

- what problems can't be fixed with correction

- before correction (measurements and listening impressions)

- an example of how to collect the measurement required

- an example of how to set up the software required within Mac environment

- after correction (measurements and listening impressions)

 

[The Computer Audiophile]

Hi Nyal - Since you are selling similar services to people your post is a bit disingenuous. Your one other post also appears to be an end-around way of promoting your services as well. Please do not post about any services you may offer unless asked a direct question about one of your products. I will remove the questionable comments from you post.

 

[Nyal Mellor]

Hi Chris sorry if I broke some rules, I couldn't find posting guidelines for manufacturers / those of us offering services (e.g. is it ok to include a link to your company or not). I looked at some people's postings and saw they had links embedded so took this to be ok. I notice Steve from Empirical often mentions his products in his postings.

 

My posting was intended in good faith, I am working on some 'articles', which will have educational value to the computeraudiophile community. Maybe I should separate my blog from my company at a different website location? Can I link to my blog like the OP of this thread? Or include my website in the signature?

 

Sorry I am still trying to figure out the etiquette (just set up my business in the last month)!

 

What would you recommend? I don't know if this is applicable but the audiocircle forum seems to have reasonably clear guidelines - see http://www.audiocircle.com/index.php?topic=41871.0

 

[kana813]

Bob H-

 

I really doubt Sean is going to give any hints/tips for your Tact 2.0S.

 

If you want better measurements/improved performance, I suggest you join these forums:

 

http://tech.groups.yahoo.com/group/TacTAudioUsersGroup/?yguid=332166746

 

http://tech.groups.yahoo.com/group/TheRealTacTHackers/?yguid=332166746

 

Aloha,

 

Dan

TheRealTacTHackers, Moderator