April 2015 Toole video on sound reproduction

[Hi-FiGuy]

We had a good discussion thread when it came out:
http://www.whatsbestforum.com/showt...gued-by-JBL-M2-speaker-system-Opinions-please

We permanently upgraded our theater at work with the M2s. They are the most dynamic yet delicate sounding speakers I have heard. They have the toughness and dynamics/power of JBL professional speakers and the nuanced sound of the Revel. Their looks however is that of JBL so it makes a negative impression on audiophiles. We have them behind our video screen to solve that .

I was there the day you got those in. You only had two sitting in the front of the theater and the reps were there. I managed to get about 30 minutes by myself with them. They were not calibrated at all. but I still enjoyed my time.

Have to get in there this summer, and watch a concert flick.

A fantastic music/video demo DVD is Pat Methany The Way Up Live, check it out.

 

[dallasjustice]

I'm not aware of any DSP software which claims to correct off axis. Can you be a little more specific about which software you are talking about and how it doesn't do what it claims to do?

His comments are rather confusing since his books advocate using DSP to deal with low frequency minimum phase room bass issues.

Actually, I'm pretty sure Toole's knowledge of DSP is very up-to-date -- and his history of it is excellent, too. That said, while I haven't spoken to him directly about it (he mentioned this exact thing a week or so ago to me), I'm pretty sure what he's talking about is exactly what others have pointed out -- that it can fix the end response, but it can't fix the radiating pattern of the speaker. So using DSP in that way -- which is what happens often with these correction devices -- puts a patch in that doesn't fix what's happening up front.

Doug Schneider
www.SoundStage.com

 

[amirm]

I'm not aware of any DSP software which claims to correct off axis.

But that is precisely what they attempt to do. They measure the sum total response at the microphone and apply correction. They have no data on how the speaker measures by itself without the room. The exception is Harman ARCOS where it comes with a library of speaker measurements for Harman products. All others will attempt to correct speaker+room response which by definition includes off-axis.

His comments are rather confusing since his books advocate using DSP to deal with low frequency minimum phase room bass issues.

This is indeed one of the areas one can get confused about. He indeed fully advocates DSP use for low frequencies where off-axis response is not an issue. It is above transition frequencies of a few hundred hertz where these comments apply.

 

[dallasjustice]

All of the DSP software I have used make good use of windowing. The time window will determine how much so called off axis will be dealt with. For most DSP software the windowing is the biggest difference in the final result. I just think its a little trite to lump all DSP software into this category. That is, not all DSP attempts to correct off axis in this way.

But that is precisely what they attempt to do. They measure the sum total response at the microphone and apply correction. They have no data on how the speaker measures by itself without the room. The exception is Harman ARCOS where it comes with a library of speaker measurements for Harman products. All others will attempt to correct speaker+room response which by definition includes off-axis.


This is indeed one of the areas one can get confused about. He indeed fully advocates DSP use for low frequencies where off-axis response is not an issue. It is above transition frequencies of a few hundred hertz where these comments apply.

 

[das]

The time window will determine how much so called off axis will be dealt with.

As far as I know, that's impossible -- the best ones are taking multiple measurements, but it's of the combined responses when it grabs them. They can't deal with on- and off-axis responses separately -- the things combine into one.

I read Amirm's comments on the other parts and as far as I'm aware, he's fully correct in what he's saying.

Doug Schneider
www.soundstage.com

 

[Phelonious Ponk]

But that is precisely what they attempt to do. They measure the sum total response at the microphone and apply correction. They have no data on how the speaker measures by itself without the room. The exception is Harman ARCOS where it comes with a library of speaker measurements for Harman products. All others will attempt to correct speaker+room response which by definition includes off-axis.


This is indeed one of the areas one can get confused about. He indeed fully advocates DSP use for low frequencies where off-axis response is not an issue. It is above transition frequencies of a few hundred hertz where these comments apply.

And Harman includes DSP for bass response in their better active monitors.

Tim

 

[dallasjustice]

Off axis gets to the mic later than on axis. Of course, there's no perfect solution. But in the case of speakers/room, perfection is not required; only an improvement. This isn't difficult to achieve in any room. I've used both multi point measurement and single point. The one I am using now is single point. However, even the multi point ones use windowing and averaging to remove enough late arrival sound as the frequency goes higher.

As far as I know, that's impossible -- the best ones are taking multiple measurements, but it's of the combined responses when it grabs them. They can't deal with on- and off-axis responses separately -- the things combine into one.

I read Amirm's comments on the other parts and as far as I'm aware, he's fully correct in what he's saying.

Doug Schneider
www.soundstage.com

 

[dallasjustice]

Amir,
Is Harman planning a consumer version of the M2?
Michael.

 

[NorthStar]

Could be a Revel F12. Not sure of the MSRP, but it is close to that. A model that has been out a few years.

It is not the Revel Concerta F12. ...Different specs and different measurements and different price ($1,500/pair MSRP).

Dr.Floyd's has long been known as a speaker guru within Canada. And I'm pretty certain I know the $1.8k speaker he refers. While he was working at Canada's National Research Council, he helped design a very *accurate* monitor speaker for CBC studios. The company helping to develop & manufacture this speaker was Audio Products International (well prior to being purchased by Klipsch) who's main brand were Energy & Mirage. The R&D cost supposedly went into the millions, as the NCR facilities were not cheap.

The price ($1,800) and the freq. curve are telling clues, this the exact cost of this speaker when introduced, with identical specs.

It is not the Energy 22 Reference Connoisseur and/or Energy Reference 22 loudspeaker; different price (on the first), different measurements.
It's a 1987-88 circa speaker, and was one of the very best @ that price and above, much more above. ...Very good guess, and I am not 100% certain that it is not. But I looked @ the measurements, same type of measurements showed in that book by doctor Floyd E. Toole, and they differ (not the exact same).

If indeed it is this speaker, and I'm pretty certain it is, it would cost considerably more today to manufacturer. It used expensive proprietary drivers, a well designed xover, and a solid well designed (in & out) cabinet. It also invited specific mods, which raised it's performance considerably. IIRC, UHF Magazine still use a pair in there HT based reference system.

The Energy Reference 22 Connoisseur was $2,600/pair back in 1988. The less expensive Energy Reference 22 was indeed in that neighborhood of $1,800/pair.
But again, the measurements are not the exact same. ...But who knows for certain if doctor Toole was not referring to that second one....

? I also looked @ several other loudspeakers; all @ $1,800/pair and with same measurements (full range). ...From way back then to today, and all the JBL and Revel and Harman Kardon loudspeakers, plus all the big Canadian names; Energy, Mirage, Image, PSB, PMC, Paradigm, Newform Research, DCM Time Frame, Axiom (Axiom M2 by the way, but a smaller 2-way monitor speaker @ $462/pair Canadian), also the Aperion Audio Verus Grand ($1,800/pair), the Atlantic Technology AT-2 ($1,800/pair), the NHT Classic Four ($1,800/pair), the Revel Performa F206 ($1,800 each), some Snell Acoustics loudspeakers,
and the PSB Stratus Gold. ...That last one; that could be it. ...I'll look more into it, before Amir can find out.

* Energy Veritas v1.8 ? ...No, but great speaker.
** Paradigm Reference Studio 100 v2 ? ...No, but great speaker too.

By the way, I searched for over more than two hours (yesterday). ...And more with today (this morning).

Probably a pair of Genelec's

The Genelec pro studio active monitors (some models) have the best speaker measurements ever, along with some Magico loudspeaker models.
And a certain Genelec active monitor model is in that range of approximately $1,800/pair. Genelec has some "reference" measurements that all other speaker's manufacturers aspire to.

Some active ATC loudspeakers are also very impressive in their performance/measurement ratio.

<<<>>> Active loudspeakers are the future today; I agree with doctor Toole. ...With integrated DSP room correction and equalization. ...In my own opinion.
...Like some Meridian loudspeakers. ...McLaren loudspeakers?

? Genelec active pro monitors are the best. ...IMO

___________

Cool thread, cool video provided by John our OP. ...Rewarding back to my sources, in time when I was younger and auditioning all the best great Canadian loudspeakers.
I had few of them in different places I was renting back in the 80s and early 90s. ...Mirage M1 among them. ...They never made the M2. ...But they did make the M3.

 

[Whatmore]

But that is precisely what they attempt to do. They measure the sum total response at the microphone and apply correction. They have no data on how the speaker measures by itself without the room. The exception is Harman ARCOS where it comes with a library of speaker measurements for Harman products. All others will attempt to correct speaker+room response which by definition includes off-axis.

So this is the bit I don't understand.
We listen to the sum total of room plus speaker so isn't that what we should correct?

 

[TBone]

It is not the Energy 22 Reference Connoisseur and/or Energy Reference 22 loudspeaker; different price (on the first), different measurements. It's a 1987-88 circa speaker, and was one of the very best @ that price and above, much more above. ...Very good guess, and I am not 100% certain that it is not. But I looked @ the measurements, same type of measurements showed in that book by doctor Floyd E. Toole, and they differ (not the exact same).

Well, I have these speakers (cons & pro's) and the accompanying specs of the pro's look very similar. The only difference was the upper extension which went to 30khz, but I'm not certain +/- 3db. This from the Energy site, concerning the design of the 22's ...

It all started with a pioneering psycho-acoustic research study hosted by the Canadian National Research Council (NRC) that ran from 1977 to 1986. This joint study between the NRC and select Canadian speaker manufacturers sought to determine what measurable factors people equate with a good-sounding loudspeaker. As it turned out, flat frequency response, wide dispersion and low distortion consistently scored high during listening trials.

I've also read that Gerard Rejkind (UHF), who knows Floyd's work well, and unless I'm mistaken ... he claimed these were in fact the speakers Floyd & API developed for CBC studio's - later be made available to the public *specifically* for $1.8k. Below is a pic of its proprietary tweeter I was working on not long ago ...


My Cons are heavily modded (the xover parts/design/slopes were kept, the connections within the xover, and all the internal wiring were modded. They sound spectacular driven by mono Classe DR3's.

 

[Kal Rubinson]

But that is precisely what they attempt to do. They measure the sum total response at the microphone and apply correction. They have no data on how the speaker measures by itself without the room. The exception is Harman ARCOS where it comes with a library of speaker measurements for Harman products. All others will attempt to correct speaker+room response which by definition includes off-axis.

DEQX also measures the speaker response with near-field placement and gating. It also will measure power response with mic placement at the MLP(s).

 

[TBone]

Floyd / NCR / Axiom link ...




http://www.axiomaudio.com/NRC

 

[Bruce B]

Just spent the last hour listening. Awesome talk. Saw the Salon 2 a few times.

 

[Nyal Mellor]

Having had a few discussions with Floyd I am pretty sure he is just trying to get across the point that the main thing that is important above the 'room mode dominated region of the frequency response' (~250Hz) is the speaker. And that you can't correct for speaker off axis issues such as suckouts around the crossover point between two drivers with DSP, which is correct.

So, focus your efforts on speakers first, not room correction. The research shows that speakers that sound best have smoothly varying off axis response that resembles the direct sound except maybe with some narrowing of directivity at the highest frequencies.

ARCOS is a very smart system, because it only applies gentle corrections above the room mode dominated region, and those corrections are based on anechoic measurements of the (Harman group only) speakers.

Having said that there are some very smart people working on 'full frequency range' room correction and the best software (Dirac Live / Trinnov) can offer a small but worthwhile improvement above the room mode dominated region, even in a well treated room.

My final thoughts are that whilst I think the focus on directivity is very important there are other extremely important things that contribute to end sound quality. Things that are not really mentioned or discussed much in the Sound Reproduction book or in the talks. Directivity is only one of a bunch of things that are important to how a speaker sounds. That's a lesson I learnt from ATC.

I wrote a bunch of articles on speaker off axis response last year, covering the directivity characteristics of different speakers and how that influences the acoustic treatment of rooms. http://www.acousticfrontiers.com/category/blog/acoustics/speaker-directivity/

 

[das]

Off axis gets to the mic later than on axis.

Actually, no. What you're referring to is, say, a short tone burst and then gating the response to try to filter the reflections out -- the "quasi anechoic" systems that many try to use for speaker measurements (better just to use a real chamber).

That said, what's measured by these correction systems is the on- and off-axis responses COMBINED. Nothing's being gated. You're also getting floor bounce, etc. So if you're flat on axis and have a mess off axis, that flat + mess will combine at the listening chair (or wherever you have the measurement microphone). That's why you can't use these systems to correct for poor off-axis behavior.

Doug Schneider
www.soundstage.com

 

[Whatmore]

Actually, no. What you're referring to is, say, a short tone burst and then gating the response to try to filter the reflections out -- the "quasi anechoic" systems that many try to use for speaker measurements (better just to use a real chamber).

That said, what's measured by these correction systems is the on- and off-axis responses COMBINED. Nothing's being gated. You're also getting floor bounce, etc. So if you're flat on axis and have a mess off axis, that flat + mess will combine at the listening chair (or wherever you have the measurement microphone). That's why you can't use these systems to correct for poor off-axis behavior.

Doug Schneider
www.soundstage.com

But if you correct *everything* that arrives at the listening position (including the off axis "mess") why do you care about the off axis performance. It's all there at the listening position.

 

[das]

But if you correct *everything* that arrives at the listening position (including the off axis "mess") why do you care about the off axis performance. It's all there at the listening position.

Precisely because what arrives at the listening position governs the tonal balance of the speaker(s) that you hear -- it's not just the direct response, but the direct combined with the first reflections and the reverberant responses. So unless you listen in an anechoic chamber, you want the on- and off-axis responses to match well.

Doug Schneider
www.soundstage.com

 

[Whatmore]

Precisely because what arrives at the listening position governs the tonal balance of the speaker(s) that you hear -- it's not just the direct response, but the direct combined with the first reflections and the reverberant responses. So unless you listen in an anechoic chamber, you want the on- and off-axis responses to match well.

Doug Schneider
www.soundstage.com

I still don't get it. Surely if you correct what actually arrives at the listening position, you are by definition, correcting the on and off- axis reponses.

I'll ask a different way. Will a speaker+room corrected at the listening position sound better (at the listening position) if the on and off -axis responses match well? If so, why? (assuming there are 'welll behaved" corrections - eg no massive suck-outs etc)

 

[dallasjustice]

In a simplistic way, you are correct that measurement tools used to prepare DSP filters are actually measuring a combined result. The ear/brain hears a combined result as well. It's a huge oversimplification to say that the listener listens "through" the room. This is where psychoacoustics come into play. For example, a software called Acourate uses what Dr. Uli Brueggemann calls Frequency Dependent Windowing. The FDW varies with the frequency in way designed to mimick how the ear/brain combines sound in a real speaker/room system. Other softwares use similar techniques to filter the measurement.

I only explain this to you because you said it's "impossible" to correct off-axis. I agree with that but the argument you make is an oversimplification which ignores how advanced DSP softwares work. The guys that design these softwares (eg. DIRAC live, Acourate and Audiolense) are very much aware of the shortcomings for the over-correction results early attempts at so called "room correction" yielded. I personally don't like that name and I use the term DSP instead mainly because I think the speaker/room come first and the DSP is the icing on the cake; nothing more.

Btw, I think the way ARCOS does correction above Schroeder is very smart and probably better in some ways than other methods since the speaker's off axis is known. But even ARCOS offers full frequency correction according to a curve set to the user's preference. The curve does correct both the speaker and the room combined above Schroeder.

Actually, no. What you're referring to is, say, a short tone burst and then gating the response to try to filter the reflections out -- the "quasi anechoic" systems that many try to use for speaker measurements (better just to use a real chamber).

That said, what's measured by these correction systems is the on- and off-axis responses COMBINED. Nothing's being gated. You're also getting floor bounce, etc. So if you're flat on axis and have a mess off axis, that flat + mess will combine at the listening chair (or wherever you have the measurement microphone). That's why you can't use these systems to correct for poor off-axis behavior.

Doug Schneider
www.soundstage.com