Room correction affecting loudspeaker choice?

[AcuDefTechGuy]

It's all about equilibrium. You probably don't want to spend $5K on speakers and $500 on AVR. Maybe $5K (MSRP) on speakers and $1K-$2K (MSRP) on AVR.

As for room treatment, just do natural like curtains, drapes, carpet, rugs, sofas, pillow, etc. IMO, no need to spend another $500 on room treatment if you already do it naturally.

 

[TLS Guy]

TLS Guy said: ↑
Good speakers get the best out of poor rooms

Bad speakers magnify room problems.

This is actually easy to understand. Room dimensional problems lead to addition and subtraction of frequencies in the bass especially the former. So sounds are prolonged. This is a form of resonance and rooms have Qs.

Your next problem is that you are focused on frequency response on axis alone as guide to a good speaker. This is one of only many attributes a speaker must have to be any good. Speakers are more often than not resonant to varying degrees. High Q speakers are the most resonant, low Q designs around 0.5 are essentially non resonant and have good impulse responses and do not "ring".

It is easy to see how a high Q speakers in a high Q room will maximize each others shortcomings. Whereas a low Q speaker in a resonant room will minimize that room problem.

In addition the off axis performance of a speaker is also if the highest importance in this matter. Those speakers with broad and smooth off axis response, will minimize problems caused by irregular frequency dependent reflections.

And yes, a couple of good speakers are far better then more poor ones.

A good stereo pair should have a wide and deep involving sound stage. They will do a far better job of throwing a convincing sound stage than a bunch of inferior ones.

 

[Swerd]

Gen. Audiologist

I think you might be overestimating the effect of room correction software.

See this quote (I've added the bold to emphasize text):
"Audyssey MultEQ: A technology that allows consumers and professionals to fix the acoustical problems in rooms that arise from the interaction of sound from the loudspeakers with the surfaces in the room. MultEQ uses acoustical measurements in the time domain taken by a microphone around the listening area and combines this information to evaluate the acoustical problems that cause audible distortions in the frequency response. MultEQ then creates a room equalization filter for each speaker and subwoofer in the system to correct these problems. Audyssey EQ is a direct extension of MultEQ for products that come with attached loudspeakers such as televisions and home theater in a box systems."

It cannot fix acoustical problems that are inherent to the speakers, and as such, it cannot equalize different speakers so that a poor sounding speaker sounds similar to a good sounding one.

Floyd Toole has directly addressed this question. His answer is buried in this somewhat long, but worthwhile, video. If anyone remembers what point in the video he speaks about room correction software, please tell us.

EDIT: see slides 33-37 (not all slides show their number), during minutes 35-40 in the video. But if you haven't watched the entire video, it may be better to start at the beginning, just to understand his definitions for the terms he uses.

 

[PENG]

It cannot fix acoustical problems that are inherent to the speakers, and as such, it cannot equalize different speakers so that a poor sounding speaker sounds similar to a good sounding one.

Floyd Toole has directly addressed this question. His answer is buried in this somewhat long, but worthwhile, video. If anyone remembers what point in the video he speaks about room correction software, please tell us.

EDIT: see slides 33-37 (not all slides show their number), during minutes 35-40 in the video. But if you haven't watched the entire video, it may be better to start at the beginning, just to understand his definitions for the terms he uses.

I can agree to that notion to a point. Room EQ systems cannot fix a lot of issues inherent to the speakers such as harmonic distortions, crossover or drive related issues, impulse response, etc., but I do believe they can improve certain frequency and time domains related issues, but again only to a point. Any such improvements could well be not too significant.

It seemed to me Dr. Floyd Toole referred to equalizer that works only on the frequency domain, in that case I can see such equalizer could in fact mess up the only good thing going for that particular speaker. Even if that is the case though, if messing up the on axis response would end up with improved response at the listening sweet spot position, I will take that every time. So I do take issue with his logic, but of course it is entirely possible I misunderstood his point.

I read enough about Audyssey to know that the algorithms they use are much more complicated and the system definitely works in the time domain, involving serious mathematical analysis, such as Fourier, fuzzy logic etc. They don't just measure the on axis response and ignore the effects of the reflected sound waves. I have no idea how exactly Audyssey, Anthem ARC, Dirac work but I am quite sure they are more sophisticated then what Dr. Floyd Toole was alluded to in those couple slides. Again, at the end of the day, IMO the most important thing is to have the best possible sound when listening from my sitting position, and that typically means off axis. For clarity, I repeat, I am willing to accept poor on axis performance if it means great performance from where I sit.

I would love to see those PhD's debate on this very topic.

 

[Swerd]

I read enough about Audyssey to know that the algorithms they use are much more complicated and the system definitely works in the time domain, involving serious mathematical analysis, such as Fourier, fuzzy logic etc...

I would love to see those PhD's debate on this very topic.

The only PhD I know of on this subject is Toole. Few, if any, would argue with him.

Dennis Murphy is a PhD, but in economics not audio, and TLS Guy is a physician, not a real scientist . But both have extensive knowledge and experience in this subject. And they both would nod their heads in agreement with Toole, that a speaker with good on- and off-axis frequency response will sound better than a speaker that lacks it. No EQ, even if it is done with the latest & smartest software, has been able to substitute for that.

Keep in mind that automated EQ, such as Audyssey, is a work in progress. There has been much recent progress with it, and there is likely to be more in the future. So stay tuned…

 

[Floyd Toole]

Hello everybody,
I can tell from this discussion that some additional facts would be useful. The McGill University YouTube lecture that Swerd showed a link to is a very good start, if you can invest an hour and 13 minutes of your time some of your questions might be answered. The room EQs I refer to include any and all products. The assumption that an omnidirectional microphone is a substitute for two ears and a brain is grossly simplistic. It cannot provide the answers we need - except at low frequencies where the measurements and EQ are both useful. If you have the time and desire to learn what the state of the art is today, I just published a 30 page paper in the Journal of the Audio Engineering Society. It is "open access" - free to all. Just go to www.aes.org, click on "publications", then "open access", type in "Toole", and my paper will appear and can be downloaded at no cost. It is a lot of information, some of it applying to movie sound, but if you watch movies at home even that is relevant to what you are hearing. A lot of the science we need exists, but we need more to fill in some gaps.

One thing that becomes very clear is that if one is absolutely dedicated to great sound quality, the weak links include the recordings and sound tracks themselves - nothing at all is adequately standardized. So sound quality, sound stage and imaging are all up for grabs. On the sound quality side, it is hard to avoid the notion that old fashioned bass and treble controls are still needed - in 2015! A lot of people think that recordings are somehow pristine and that all problems are at the playback end. Not so. The most "perfect, neutral" loudspeaker in the world will not sound good with all recordings. And some of the common room EQ procedures have a significant risk of degrading the performance of a good loudspeaker. This is not a good situation.

The topic of "time domain" keeps popping up as if it is being ignored by frequency-response measurements. Loudspeaker transducers - woofers, midranges and tweeters - are "minimum-phase" devices. This means that the time domain performance is predictable from the amplitude (frequency) response. If the anechoic frequency response is smooth and flat, there are no resonances, no ringing, no colorations. This is why flat and smooth has always been the target performance for all electronics, and now we have come to the realization that it applies to loudspeakers too. Good loudspeakers yield good sound and good looking room curves. But, unless you know a lot about the loudspeakers, the reverse may not be true. See the last curves shown in the lecture and in the paper.

The industry needs guidance so that audiophiles can relax and just enjoy the music and movies.

 

[Swerd]

Thanks for replying. I hope that general_audiologist sees this.

 

[AcuDefTechGuy]

Yeah, IMO, hoping that EQ could fix any speaker and situation is a good fantasy, but fantasy nonetheless.

 

[PENG]

The only PhD I know of on this subject is Toole. Few, if any, would argue with him.

I thought you know of Dr. C. Kyriakakis of Audyssey, co founder of of Audyssey Laboratories. I am sure there are others knowledgeable in this field though may not be considered as accomplished as Dr. Toole.

I hope you don't think I would even dream of arguing with Dr. Toole. If you did, please re-read my post because I would dare..

that a speaker with good on- and off-axis frequency response will sound better than a speaker that lacks it. No EQ, even if it is done with the latest & smartest software, has been able to substitute for that.

I fully agree, as I can understand the basic ideas behind. Again, I hope nothing I said in my post would indicate otherwise. I did make this point in post#24: "but I do believe they can improve certain frequency and time domains related issues, but again only to a point. Any such improvements could well be not too significant." Perhaps you missed a couple of key words in that sentence.

The only other part of your post that I have slight disagreement is the following:

"It cannot fix acoustical problems that are inherent to the speakers, and as such, it cannot equalize different speakers so that a poor sounding speaker sounds similar to a good sounding one."

I think it can, if done right, fix certain problems inherent to the speakers to limited extent but I fully agree that it cannot make it sound similar to a good sounding one. Case in point, if I remember right at one point Dr. Toole did say EQ could get rid of a bump/resonance, and I believe he was only citing it as an example and perhaps meant to be a theoretical one.

 

[PENG]

Yeah, IMO, hoping that EQ could fix any speaker and situation is a good fantasy, but fantasy nonetheless.

I doubt that will ever happen if "speaker" mean any and all speaker.

 

[PENG]

It cannot provide the answers we need - except at low frequencies where the measurements and EQ are both useful.

Dr. Toole, thanks for your help here but I thought at one point in the presentation, around 1:02:10 you mentioned that an equalizer (yes you stated some conditions) could get rid of "the bump". So I took it that in theory, if done right, EQ could in fact be useful in resonance related issues, however limited in extent that may be and perhaps such EQ does not exist in the real world yet. I did feel I might have misunderstood the point you were making though.

A lot of people think that recordings are somehow pristine and that all problems are at the playback end. Not so. The most "perfect, neutral" loudspeaker in the world will not sound good with all recordings.

Wow, I always felt the quality of the recordings is paramount, ranked top, to the point I have started a thread in AH to collect inputs from members on what they consider high quality recordings and have compiled a list of about 115 such recordings in CD, BR, SACD and Vinyl albums so far.

Thanks for confirming that my instinct is right.

And some of the common room EQ procedures have a significant risk of degrading the performance of a good loudspeaker. This is not a good situation.

I don't doubt that at all. I ran REW to plot a bunch of curves each time after I ran Audyssey.

The topic of "time domain" keeps popping up as if it is being ignored by frequency-response measurements. Loudspeaker transducers - woofers, midranges and tweeters - are "minimum-phase" devices. This means that the time domain performance is predictable from the amplitude (frequency) response. If the anechoic frequency response is smooth and flat, there are no resonances, no ringing, no colorations. This is why flat and smooth has always been the target performance for all electronics, and now we have come to the realization that it applies to loudspeakers too. Good loudspeakers yield good sound and good looking room curves. But, unless you know a lot about the loudspeakers, the reverse may not be true. See the last curves shown in the lecture and in the paper.

It has been many years since I studied Fourier transforms but I do remember its application in converting between the frequency and time domain. I was just wondering if some Room EQ software could at least theoretically do something to improve certain less than good speakers in room performance. I think I understand your point about the limitation of using a single mic, but as you know some of those software do rely on placing the mic in multiple locations in order to collect more data. Audyssey, for example claimed they did not do that for averaging, but for obtaining more information for their algorithm, but then again I could have misunderstood what they were actually saying.

I did watch the whole video presentation and saw the last curves. Many thanks again.

 

[TLS Guy]

Hello everybody,

The topic of "time domain" keeps popping up as if it is being ignored by frequency-response measurements. Loudspeaker transducers - woofers, midranges and tweeters - are "minimum-phase" devices. This means that the time domain performance is predictable from the amplitude (frequency) response. If the anechoic frequency response is smooth and flat, there are no resonances, no ringing, no colorations. This is why flat and smooth has always been the target performance for all electronics, and now we have come to the realization that it applies to loudspeakers too. Good loudspeakers yield good sound and good looking room curves. But, unless you know a lot about the loudspeakers, the reverse may not be true. See the last curves shown in the lecture and in the paper.

The industry needs guidance so that audiophiles can relax and just enjoy the music and movies.

Many thanks for joining this discussion! I think all of us who have worked with loudspeakers have come to realize that frequency response aberrations are symptoms of underlying problems, and usually serious ones.

I really don't think you can correct a speaker with Eq, especially at a distance in a listening room.

I do know that good loudspeakers are actually the least fussy about everything and poor speakers exacerbate flaws from everywhere.

However the better the source the better a good speaker really shines.

I have a VPN tunnel to the UK, and have been enjoying the BBC Proms on iPlayer both audio only and video. Their HD stream is amazing this year. Quite the best reproduction I have ever encountered. All instruments caught perfectly and beautifully balanced. The RAH organ showing strong output below 20 Hz. I have never heard such realistic strings. The brass blazes and the dynamic range is enormous. It really is like being there.

I don't know what the BBC have been up to, but I suspect they have been timing the mics and may be using techniques to limit mic bleed.

Anyhow it has been the most realistic reproduction I have heard, and it comes over the NET.

Progress is being made now at an astonishing rate.

 

[Floyd Toole]

A point of clarification. If you go back and read my post I said "If the anechoic frequency response is smooth and flat . . ." I did not mean the "room curve". The two are separated by the room which introduces massive non-minimum-phase phenomena - a.k.a reflections. Once the sound has left the loudspeaker one has limited ability to interrogate the performance of the loudspeaker using in-room measurements - no matter how clever the mathematics. Since what we hear is dominated by the loudspeaker (above about 300 Hz) this means that starting with a good loudspeaker is an enormous advantage. This, as I point out in the McGill lecture, is why loudspeakers with dedicated amps and DSP have a distinct advantage - assuming the engineer has access to the necessary design tools. This is all explained in turgid detail in the AES paper.

 

[PENG]

A point of clarification. If you go back and read my post I said "If the anechoic frequency response is smooth and flat . . ." I did not mean the "room curve". The two are separated by the room which introduces massive non-minimum-phase phenomena - a.k.a reflections. Once the sound has left the loudspeaker one has limited ability to interrogate the performance of the loudspeaker using in-room measurements - no matter how clever the mathematics. Since what we hear is dominated by the loudspeaker (above about 300 Hz) this means that starting with a good loudspeaker is an enormous advantage. This, as I point out in the McGill lecture, is why loudspeakers with dedicated amps and DSP have a distinct advantage - assuming the engineer has access to the necessary design tools. This is all explained in turgid detail in the AES paper.

Yes you did, and I did not miss that part. I was only wondering if any such REQ system could improve things slightly, I mean very slightly such as minimizing a bump or two. Since you mentioned "limited ability", I would assume you meant a good room EQ system could in fact have be some "ability", though limited, or highly limited in the >300 Hz range. I realize all of us should try to buy the best speakers we can afford and not rely on any EQ systems. I guess being an EE I tend to get hung up on the theoretical side, while realizing in practice it may be of very little value.

I have downloaded your 30 page paper and intend to read it all this weekend. Thanks gain for the free education.

 

[Floyd Toole]

Peng - I have a PhD in Electrical Engineering, so I'm not immune to your "theoretical" instincts. I got to where I am because very early in my career , while enjoying my audio hobby, my ears were telling me something different from some of the measured data. I like finding answers Hence, I learned what was necessary to do the serious psychoacoustic research, and had an employer (the National Research Council of Canada) that had the facilities and allowed me the privilege to do the necessary research - I was a research scientist paid by the Canadian government. There was no connection with any brand or product - the goal was knowledge. By a remarkable circumstance, when I joined Harman 25 years later I was allowed to create a genuine research group and to publish all of its findings - public information, not corporate secrets. After my retirement in 2007, it has continued, under the guidance of Dr. Sean Olive, which gives me great satisfaction. We don't yet have all of the answers, but we have a lot of them, and they are in the public domain. The result, I would like to think, is better sound for more people.

 

[PENG]

Dr. Toole, I am very happy to learn that you are not just any PhD but a PhD in EE from IC, the best university of science and technology in the U.K. It reminds me of my university days when I often visited that campus on weekends to see my friends who were doing their PhD there, in EE too actually.

 

[herbu]

general_audiologist,
All this technical talk notwithstanding, here's what really happened...

One day, Bubba was reared back in his La-Z-Boy, drinking beer and listening to his stereo and it occurred to him... "Unless my speakers are exactly the same distance from me, their sound won't get to me at the same time. I know how fast sound travels, so if I measure how much closer one speaker is than the other, and can delay the signal appropriately to that speaker, I could make the sound hit my ears at the same time".

After a couple more beers, he figured he could do the same thing to "match" all speakers in a surround system. Then he ran out of beer, and cracked open the moonshine. After a couple swallers, he realized he could also measure the volume of each speaker, at his chair, and match their volume so the closer speakers didn't sound louder. (He didn't realize this could be useful in smoothing peaks and valleys caused by room geography and furniture, but heck, it's only Bubba.) Then he passed out and slept. But while he was sleeping, he dreamed about checking phases too.

So that's the real story behind Audyssey. You can see there is nothing there about making a bad speaker sound good. It's only about trying to mitigate or "balance" effects from the environment that exist between the speaker and your ear(s).

Dr. Toole,
You sound like a pretty good guy. Thanks for trying to help. But I gotta say, you sure use a lot of big words.

 

[general_audiologist]

It's been a couple of busy days, so I haven't had the time to come here. I see there are a lot of interesting replies, thank you all for contributing. I have to admit I'm overly excited to see Dr.Toole have replied to my thread.

If you have the time and desire to learn what the state of the art is today, I just published a 30 page paper in the Journal of the Audio Engineering Society. It is "open access" - free to all. Just go to www.aes.org, click on "publications", then "open access", type in "Toole", and my paper will appear and can be downloaded at no cost.

Thank you, I will definitely check this out! Me and two of my former classmates turned to your 2012 "Sound Reproduction: The Acoustics and Psycoacoustics of Loudspeakers and Rooms" during our work with our bachelor assignment; "What level of objective speech intelligibility is achieved by different loudspeaker configurations, for sound field amplification systems used in a classroom setting?". For anyone interested we used speech transmission index measurements at seating positions to compare the different configurations.

I wish to reply to more of the contributions that have been made to this thread, but I'll have to save that for later during the weekend.

 

[DD66000]

Well, two loudspeakers make for a very poor surround sound experience Let's face it, you will always be able to afford better loudspeakers going for two instead of seven, the question here is what is a "lesser one". At some point you have to accept that in order to have five, you will have to go for a lesser model than you could have if you only needed two.

Please remember that a the "lesser option" might be a pair of loudspeakers retailing for 800 dollars, and the better option is a 1400 dollar pair. The good old "bang-for-the-buck" ratio declines as we move up the price ladder, so if you take room correction into account, will that ratio remain unaltered, or does it grow or decline even more?

With proper placement, quality L/R mains can and will give you an enveloping sound. Not quite as good as 5, 6, 7 but close.
And if one wants to have more/less the same high quality all around, then the obvious solution is to buy your speakers as money comes available...not buying all at once, thus settling for lower quality.

To get the most of room correction...I use Dirac...one has to start with quality speakers...properly setup L/R mains...meaning getting the best possible stereo soundstage for that room. Then moving on to the center/surround/sub placement. Then apply RC.

 

[PENG]

I know next to nothing about psychoacoustics but so far I don't get the impression that Dr. Toole has said anything that contradicts my original belief, but I am really happy and thankful that he participated regardless. I will try to remember at least the following, quoted from his posts# 26 and 33.

"Once the sound has left the loudspeaker one has limited ability to interrogate the performance of the loudspeaker using in-room measurements - no matter how clever the mathematics. Since what we hear is dominated by the loudspeaker (above about 300 Hz) this means that starting with a good loudspeaker is an enormous advantage. This, as I point out in the McGill lecture, is why loudspeakers with dedicated amps and DSP have a distinct advantage - assuming the engineer has access to the necessary design tools. This is all explained in turgid detail in the AES paper."

"The assumption that an omnidirectional microphone is a substitute for two ears and a brain is grossly simplistic. It cannot provide the answers we need - except at low frequencies where the measurements and EQ are both useful. If you have the time and desire to learn what the state of the art is today, I just published a 30 page paper in the Journal of the Audio Engineering Society. It is "open access" - free to all. "

I also like what Swerd said "Keep in mind that automated EQ, such as Audyssey, is a work in progress. There has been much recent progress with it, and there is likely to be more in the future. So stay tuned…"

Hopefully some day in the future room EQ systems can and will do more to help than to worsen (as Dr. Toole mentioned about that risk) what we have now.

If the OP has not read the very interesting interview linked below, I suggest he should read them.

http://www.audioholics.com/room-acoustics/audyssey-room-eq-interview

http://www.audioholics.com/room-acoustics/anthem-arc-room-eq-interview