Are audio companies all involved in a huge conspiracy? - Page 34 - AVS Forum
Forum Jump: 
 
Thread Tools
Old 10-23-2012, 09:28 AM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,530
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1205
Quote:
Originally Posted by sdurani View Post

I posted the B&K graph from 1974 to demonstrate that listener preference hasn't changed. Anyone comparing it to Olive's paper from a quarter century later will notice that both curves were smooth and downward tilted. The similarity also demonstrates that listener preference was not tied to products, since a quarter century of separation meant that different products were used, even though listener preferences remained consistent.

The above post shows considerable ignorance or outright distortion of the actual context in which the cited frequency response plot was presented.

The earliest and closest source of the plot you've been posting that I can find is:

http://www.bksv.com/doc/17-197.pdf (from 1974)

(page 4)

There is also an 1974 AES paper but it is very similar.

The associated text is:

"The closer evaluation of these curves concern the following two criteria. First, the curves should be as smooth and straight as possible, indicating that all frequencies are reproduced at approximately equal level. Secondly, primary emphasis should be given to the 60 Hz to 6 kHz range.

When music is recorded under farfield conditions, it will contain a suit*able mixture of direct and reflected sound, and the curve ought to be absolutely flat in that case.
This is true for recordings, for instance, made with two B &K condenser mi*crophones in the far-field.

However, since most recordings are made as a combination of nearfield and far-field information, the curve should boost a little at low frequencies and roll off a little at high frequencies. A suitably shaped curve is shown in Fig.5.
"

In fact the abstract of the paper says it is about measurement technologies, not listener preferences

"
This application note will describe the measurement possibilities, and it will show some results of the methods and compare them with re*sults obtained by listening tests.
Five loudspeakers in three different rooms and a test team consisting of five critical listeners were used.
"

That would be 2 listening rooms and an unknown number of listeners more than Harman used, no> ;-)

If you actually read the paper, the diagram that you have been hanging your argument about the time-invariant nature of shaped room curves on has nothing to do with a representative survey of consumer preferences. It is presented in a context where the ideal was that system response curves should be "...as smooth and straight as possible".

The shaped curve was justified by generalizations about the prevalent recording technology in 1974 which was as we all should know in the pre-digital age. The prevailing recording technology has changed greatly since then, both in terms of media and methodology. Modern recordings are overwhelmingly made in the near field. A more pronounced curve would seem to be in order for use with modern recordings.

And whoops, does that not indicate that there has been a great sea change in consumer preferences since 1974?
arnyk is offline  
Sponsored Links
Advertisement
 
Old 10-23-2012, 09:40 AM
AVS Special Member
 
Randy Bessinger's Avatar
 
Join Date: Jun 2009
Posts: 1,347
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 67 Post(s)
Liked: 66
Quote:
Originally Posted by Chu Gai View Post

I reread Sean Olive's blog on this study as well as the 30 or so slides from his presentation.
  • I note that the Harman product(s) were prototypes rather than commercially available products. I can't help but think there must be prototypes for all the products tested and that not all features or capabilities in prototypes find their way into the consumer release. Kind of like alphas and betas of Windows that people get a chance to play with but then the final version has several hoped for capabilities eliminated. Perhaps they find their way into Service Packs or maybe the next version of Windows. Would a meal prepared by Maurimoto at his signature restaurant in Philadelphia taste just a bit better than the same dish prepared by the chefs tasked with cooking there?
  • In Sean's slides he looks at the amplitude responses for the various products tested. The Audyssey if memory serves me correctly resulted in somewhat worse bass correction than than using nothing. I don't know enough or anything for that matter WRT that particular product but can anyone who does shed some light on whether that product ought to have performed that way? If it shouldn't have, that suggests that a variety of things could have been amiss. Makes me wonder if the Audyssey folks had ever commented on that study.
  • Certainly, as Craig has pointed out, if the goal was to shed light on the things that can be done to an in-room frequency response curve to improve the user experience, they could simply have used their prototype and examined various scenarios. But as Olive explicitly states, the goal was to evaluate existing room correction products including a prototype that was not commercially available at that time. Dare I say, had the prototype fared worse, there likely would have been another prototype.
Chu excellent post. It was surprising how quickly the competitors units were unofficially revealed. Certainly, I remember it causing a stir. Call me a skeptic, but I do believe there is a bit of marketing that goes on here...
Randy Bessinger is online now  
Old 10-23-2012, 10:29 AM
AVS Addicted Member
 
Chu Gai's Avatar
 
Join Date: Sep 2002
Location: NYC area
Posts: 15,067
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 632
Wasn't it Ben Franklin who said three people can keep a secret so long as two of them are dead? Iuf that study had been by someone in college doing a research paper maybe in part to fulfill a degree requirement I'd think nothing of it. Look, the findings are great. Anyone who has ever done critical evaluations of speakers can get behind those that roll off like a brick at 70-80 Hz sound thinner than full range ones. Anyone who bought a sub like that JBLwhose solo FR went down to around 20 Hz and found that they were missing a couple of those octaves woul have said what the fvck? If they had even a modicum of placement freedom that sub would be somewhere else. Maybe even near field. I seriously doubt they would be looking at a room correction device to bump them a dozen or so dbs. While the overall conclusions make sense, it didn't take long to determine the players and one and one of them wasn't a production model.

"I've found that when you want to know the truth about someone that someone is probably the last person you should ask." - Gregory House
Chu Gai is offline  
Old 10-23-2012, 10:46 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,490
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 1069 Post(s)
Liked: 558
Quote:
Originally Posted by Chu Gai View Post

Look, the findings are great. Anyone who has ever done critical evaluations of speakers can get behind those that roll off like a brick at 70-80 Hz sound thinner than full range ones. Anyone who bought a sub like that JBLwhose solo FR went down to around 20 Hz and found that they were missing a couple of those octaves woul have said what the fvck? If they had even a modicum of placement freedom that sub would be somewhere else. Maybe even near field. I seriously doubt they would be looking at a room correction device to bump them a dozen or so dbs. While the overall conclusions make sense, it didn't take long to determine the players and one and one of them wasn't a production model.
Chu, may I ask if you had read this reply I wrote: http://www.avsforum.com/t/1425262/are-audio-companies-all-involved-in-a-huge-conspiracy/930#post_22513621?

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is online now  
Old 10-23-2012, 10:49 AM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 20,211
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 1679 Post(s)
Liked: 1131
Quote:
Originally Posted by arnyk View Post

In fact the abstract of the paper says it is about measurement technologies, not listener preferences
You have to read past the abstract to find out what gave the curve its shape: "The person listening was asked to choose which of the two loudspeakers he preferred to listen to at any given time."
Quote:
Originally Posted by arnyk View Post

And whoops, does that not indicate that there has been a great sea change in consumer preferences since 1974?
Woops, no "great sea change" in preference demonstrated, especially when the curves in Olive's test and B&K's test are both downward tilted and smooth, displaying similarity despite 35 years of separation and completely different equipment used.

Sanjay
sdurani is offline  
Old 10-23-2012, 11:12 AM
AVS Addicted Member
 
Chu Gai's Avatar
 
Join Date: Sep 2002
Location: NYC area
Posts: 15,067
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 632
Quote:
Originally Posted by amirm View Post

Chu, may I ask if you had read this reply I wrote: http://www.avsforum.com/t/1425262/are-audio-companies-all-involved-in-a-huge-conspiracy/930#post_22513621?
yes. Can I get the twitter version?

"I've found that when you want to know the truth about someone that someone is probably the last person you should ask." - Gregory House
Chu Gai is offline  
Old 10-23-2012, 11:28 AM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,530
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1205
Quote:
Originally Posted by sdurani View Post

"Woops, no "great sea change" in preference demonstrated, especially when the curves in Olive's test and B&K's test are both downward tilted and smooth, displaying similarity despite 35 years of separation and completely different equipment used.

Yes, let's put those allegedly similar curves into the same post:

We have the Harman curves:



Notice the drop from ca. 95 dB @ 20 Hz to ca. 85 dB (10 dB drop) in the most preferred curve (red).

We have the B&K curve:



Notice the drop from ca. 82 dB @ 20 Hz to ca. 77 dB (5 dB drop)

IOW 5 dB = 10 dB in Durani world! ;-)
arnyk is offline  
Old 10-23-2012, 12:49 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,490
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 1069 Post(s)
Liked: 558
Quote:
Originally Posted by Chu Gai View Post

yes. Can I get the twitter version?
Sure. There were serious problems in the mid-frequencies and lack of smoothness, none of which gets remedied with more bass boost, or better curve. The problems here garnered scathing feedback from listeners in blind testing:

viewer?pid=explorer&srcid=0B97zTRsdcJTfY2U4ODhiZmUtNDEyNC00ZDcyLWEzZTAtMGJiODQ1ZTUxMGQ4&docid=6acfbb12ff81eb2c6a886a651d8b41c8%7C17ef4ac2e0c6e985ad557a5d306c89c9&a=bi&pagenumber=18&w=800

You seem to be saying it was all predictable and had something to do with subwoofers. Do you think your comments fairly apply to all the findings in the research paper?

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is online now  
Old 10-23-2012, 12:49 PM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 20,211
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 1679 Post(s)
Liked: 1131
Quote:
Originally Posted by arnyk View Post

Notice the drop from ca. 95 dB @ 20 Hz to ca. 85 dB (10 dB drop) in the most preferred curve (red).

Notice the drop from ca. 82 dB @ 20 Hz to ca. 77 dB (5 dB drop)
Notice they both drop, just like RC3 (Lyngdorf), which is what's key to listener preference, not the exact number of dB. The last line from Olive's presentation, "Flat in-room response is not the optimal target response", makes no mention of specific number of dB. As a demonstration, notice that RC3 didn't have the exact same (down to the decibel) slope as RC1, but it was still preferred to no room correction by having a downward slope (maintaining room gain).

Since Olive didn't specify the number of dB, the more telling graph is the compainion to the one you posted, which shows how those downward sloped responses were perceived by listeners. Despite having different amounts of measured tilt, the most preferred responses had at least one thing in common: they were all perceived as relatively flat, at least compared to the lesser preferred responses.

viewer?pid=explorer&srcid=0B97zTRsdcJTfY2U4ODhiZmUtNDEyNC00ZDcyLWEzZTAtMGJiODQ1ZTUxMGQ4&docid=6acfbb12ff81eb2c6a886a651d8b41c8%7C365633405897b87dcd7bea289c819844&a=bi&pagenumber=25&w=796

Again, if you have any evidence to refute Olive's findings that a downward sloped and smooth response is preferred, then please post it. And, just as a reminder, "smooth" is not "flat"; the former has to do with irregularities on the curve (peaks & dips) while the latter has to do with orientation of the curve (laying horizontal vs angled down).

Sanjay
sdurani is offline  
Old 10-23-2012, 12:59 PM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,530
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1205
Quote:
Originally Posted by sdurani View Post

Notice they both drop, just like RC3 (Lyngdorf), which is what's key to listener preference, not the exact number of dB.

The above as presented would be an unsupported assertion, AKA Opinion Stated As Fact (OSAF).
Quote:
The last line from Olive's presentation, "Flat in-room response is not the optimal target response", makes no mention of specific number of dB

So what?

Are you saying that none of the curves in the paper have any meaning at all unless they are mentioned in the last line of Olive's presentation? ;-)

Trust me, not even my kids or wife would let me get away with BS of this kind. I take it that yours are far less discriminating than mine? ;-)
arnyk is offline  
Old 10-23-2012, 02:01 PM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 20,211
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 1679 Post(s)
Liked: 1131
Quote:
Originally Posted by arnyk View Post

The above as presented would be an unsupported assertion, AKA Opinion Stated As Fact (OSAF).
Except I've supported my opinion with evidence, showing that preference tests yield downward sloping curves, even when the dates of those tests are 35 years apart and they use completely different hardware (speakers, rooms, electronics). The B&K, Lyngdorf and Harman curves all exhibit a downward tilt and were all derived from testing listener preference.

By comparison, you have provided nothing (zero) to back up your opinion that Olive's findings are dated. Even your previous attempt, claiming the B&K paper was not testing listener preference, was easily refuted by posting one line from that paper. You're projecting your OSAF problem onto me.
Quote:
Originally Posted by arnyk View Post

Are you saying that none of the curves in the paper have any meaning at all unless they are mentioned in the last line of Olive's presentation?
I'm pointing out yet another one of your strawmen, where you attempted to refute Olive's findings about preference by bringing up the number of dB in the tilt, even though Olive never mentions the number of dB that the preferred curves were tilted.
Quote:
Originally Posted by arnyk View Post

Trust me, not even my kids or wife would let me get away with BS of this kind.
Explains why you save it for the forum.

Sanjay
sdurani is offline  
Old 10-23-2012, 02:03 PM
AVS Addicted Member
 
Chu Gai's Avatar
 
Join Date: Sep 2002
Location: NYC area
Posts: 15,067
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 632
Quote:
Originally Posted by amirm View Post

Sure. There were serious problems in the mid-frequencies and lack of smoothness, none of which gets remedied with more bass boost, or better curve. The problems here garnered scathing feedback from listeners in blind testing:
viewer?pid=explorer&srcid=0B97zTRsdcJTfY2U4ODhiZmUtNDEyNC00ZDcyLWEzZTAtMGJiODQ1ZTUxMGQ4&docid=6acfbb12ff81eb2c6a886a651d8b41c8%7C17ef4ac2e0c6e985ad557a5d306c89c9&a=bi&pagenumber=18&w=800
You seem to be saying it was all predictable and had something to do with subwoofers. Do you think your comments fairly apply to all the findings in the research paper?
I agree with your general assessment. I trust you understand if not neccessarily agree with my taking issue wrt using a non production unit.

I don't know of if it was all predictable. Maybe it was intentional putting the sub in that location in order to test the capabilities of the various products. However you want to look at it, I would not recommend boosting the low end of a sub by such a significant amount for concern of blowing it up.

And with the utmost sincerity, thanks for the succinct response.smile.gif

"I've found that when you want to know the truth about someone that someone is probably the last person you should ask." - Gregory House
Chu Gai is offline  
Old 10-23-2012, 02:16 PM
AVS Special Member
 
J_Palmer_Cass's Avatar
 
Join Date: May 2005
Posts: 6,495
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 294 Post(s)
Liked: 107
Quote:
Originally Posted by arnyk View Post

Quote:
Originally Posted by sdurani View Post

"Woops, no "great sea change" in preference demonstrated, especially when the curves in Olive's test and B&K's test are both downward tilted and smooth, displaying similarity despite 35 years of separation and completely different equipment used.

Yes, let's put those allegedly similar curves into the same post:

We have the Harman curves:



Notice the drop from ca. 95 dB @ 20 Hz to ca. 85 dB (10 dB drop) in the most preferred curve (red).

We have the B&K curve:



Notice the drop from ca. 82 dB @ 20 Hz to ca. 77 dB (5 dB drop)

IOW 5 dB = 10 dB in Durani world! ;-)



Here is the Synthesis target curve for frequencies below 1 kHz:







Here is the Synthesis full range target curve:


J_Palmer_Cass is offline  
Old 10-23-2012, 02:31 PM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 20,211
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 1679 Post(s)
Liked: 1131
Quote:
Originally Posted by rnrgagne View Post

That is a common response to accurate or "flat" bass, it's not exclusive to Audyssey. I remember similar responses to infinite baffle subs which tend to be pretty flat form the get go, and we're talking mutliple driver subs capable moving a ton of air.
True, it's not exclusive to Audyssey; in fact I've seen plenty of folks, from DIY hobbiests to industry pros, aim for a goal of flat bass in the name of accuracy (I'm never sure what that term means to different people). The resulting sound might be a better match for the response of the rooms were source material was mixed/mastered, at least compared to the sound you'd get by dialing in bass response based on your preference.
Quote:
Originally Posted by rnrgagne View Post

More often than not I believe it's as you suggest, that room gain seems normal, but just as frequently bloat is considered "powerful". In both cases the listener is likely "accustomed" to the sound.
Right, which is one reason why a target curve based on what we've become accustomed to might be preferred to a target curve based on established reference standards (THX dubbing stages).

Sanjay
sdurani is offline  
Old 10-23-2012, 02:34 PM
AVS Club Gold
 
craig john's Avatar
 
Join Date: Oct 2003
Location: Lancaster, PA
Posts: 10,428
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 161 Post(s)
Liked: 392


This graph totally explains this whole problem. Look at the un-EQ'd response, (which is the black dashed trace.) It rolls off from the peak dramatically at about 50 Hz. There must be a cancellation in this range. Audyssey, which is the curve just below the un-EQ'd response, has knocked down the peak, and then has a very similar roll-off. This is because Audyssey looks for the -3 dB point of the sub and does not correct below the -3 dB point. It would need to BOOST below the -3 dB point to provide correction, and boosting below the -3 dB point can be detrimental to subwoofers. Therefore, it is doing exactly what it is supposed to do.

The question is, should the actual -3 dB point be different than what Audyssey measured. This is certainly possible and may have been correctable with some different measurement locations. We don't know anything about how Audyssey was implemented in this system, so it is certainly plausible that this was a bad implementation of Audyssey. They may have been measuring close to a wall or they may have only used 2 or 3 measurements. We just don't know.

In addition, in the Audyssey curve, there is a dip through the range of the crossover, (the crossover frequency was 80 Hz.) Gee, that looks familiar...



I am certain that dip can be corrected with a slight adjustment of the subwoofer distance setting. I have done this in NUMEROUS system I have set up with Audyssey. Audyssey was said to sound the most "thin" and the least "full" in this comparison. Correcting that dip would have made it sound much fuller and less thin.

However, also take a look at the top tracing, the "most preferred" curve. It has a rising response all the way to 20 Hz. I would be very curious to know how they accomplished that, when this was the subwoofer:



The speakers couldn't do it. They're ported with a port tune in the mid 30's. Besides, they're crossed out at 80 Hz. So they had to do that with boost in the subwoofer. I can find ZERO information on the web on the HB5000 subwoofer. I presume it's a sealed design, although that hump in the response centered around 36 Hz leads me to believe there may be a port in there. If it's a ported sub, you can't boost it below port tune. The driver is unloaded and it will distort like crazy. It can even be damaged. Even if it were sealed, they would need significant boost to add that much output in the lowest frequencies. It looks like it's down about 6 dB at 20 Hz. An LT circuit would do it if the sub is sealed, but you need enough amplifier headroom and driver excursion to allow it to work. If the roll-off is caused by the room, you can't overcome that with EQ or boost

In addition, the unEQ'd curve has a narrow null at 200 Hz. None of the other curves had much impact on that null, probably because you can't EQ out a null. Yet in the "most preferred" trace that null is somehow magically gone. How'd dey do dat?

The point I've been making all along is that Audyssey needs optimization. It's clear to me that Audyssey was NOT OPTIMIZED in this implementation. IT is also blatantly obvious that the JBL system was totally optimized. The result of comparing these two products in these two implementations is totally and completely predictable... and worthless.

Craig

Lombardi said it:
Perfection is not attainable, but if we chase perfection we can catch excellence."

My System

craig john is offline  
Old 10-23-2012, 02:53 PM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,530
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1205
Quote:
Originally Posted by sdurani View Post

Except I've supported my opinion with evidence, showing that preference tests yield downward sloping curves, even when the dates of those tests are 35 years apart and they use completely different hardware (speakers, rooms, electronics)

Wow. your story changed yet again!

You just said "Woops, no "great sea change" in preference demonstrated,", and now you're excusing the fact that the preference curve actually did change by a whopping 2:1 over those 35 years.

The only point of commonality that you are now able to find is a downward slope. Of course this is contradicted by other evidence such as this. apparently from none other than Harman:



Except of course its not a downward sloping curve but instead its a ca.7 dB peak centered at about 30 Hz.

So where do you find this 7 dB peak in the B&K curve? ;-)

I've reproduced the B&K curve here for your convenience while you try to find that 7 dB peak...

arnyk is offline  
Old 10-23-2012, 08:00 PM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 20,211
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 1679 Post(s)
Liked: 1131
Quote:
Originally Posted by arnyk View Post

The only point of commonality that you are now able to find is a downward slope.
That's what I've been saying all along; as recently as earlier today:
Quote:
Originally Posted by sdurani View Post

The B&K, Lyngdorf and Harman curves all exhibit a downward tilt and were all derived from testing listener preference.
Your new tactic is to confirm what I said?

Sanjay
sdurani is offline  
Old 10-23-2012, 10:49 PM
AVS Club Gold
 
craig john's Avatar
 
Join Date: Oct 2003
Location: Lancaster, PA
Posts: 10,428
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 161 Post(s)
Liked: 392
Quote:
Originally Posted by craig john View Post

I believe the Audyssey device they tested was the stand-alone product, right???. They never licensed that to anyone, AFAIK. They sold it directly to CI's, which is the same business model as Harman's. So, at the time, Harman must have viewed it as a competitor, which is why they would have had a motivation to discredit it.
Craig.
Quote:
Originally Posted by sdurani View Post

By that logic, Lyngdorf never licensed their stand-alone box to McIntosh, Trinnov never licensed their studio EQ unit to Sherwood, Anthem never stuffed their D2 pre-pro into any of their receivers, TacT isn't currently trying to squeeze their stand-alone box into Emotiva's upcoming pre-pro.

Wait... What ???? I am totally confused.gif What does Lyngdorf's, Trinnov's TacT's or Anthem's licensing of their "consumer-based" product have to do with this discussion??? I was pointing out that Audyssey never licensed their stand-alone, Custom Installer, (CI), based product to any other entity than to Audyssey-trained CI's. Their CI product was exclusive to their CI division. If those other companies license their technologies to some other "consumer" companies, that has no bearing on their CI based market.

At the time the Olive study was done, Audyssey had a whole division that was focused on, and exclusive to, the Custom Installer, (CI), market. The products they sold were not even available directly to consumers. The CI's that sold those products had to get specific Audyssey training. The Audyssey Sound Equalizer was sold exclusively through that division. That market continued with their "Pro Kit" which was only marketed to CI's. How is that NOT a direct competitor to the RC system Harman was developing? The Harman system was to be distributed through CI's trained by Harman. That appears to be the exact same business model, is it not??? The price point was different by a factor of about 6x, but the business model was the same. Obviously Audyssey's business model has changed SINCE THEN, and it is now primarily based on a "licensing" model. Still that doesn't change what was happening back when this "study" was conducted.
Quote:
Originally Posted by sdurani View Post

Olive compared room correction technologies, not the containers they arrived in. You'll notice in his AES paper a complete lack of comments about the hardware; no mention of build quality, location of microphone jacks, A/D converters, ease of use, length of supplied mic cable, cosmetics, etc.
By the time the Sound Equalizer came out, MultEQXT had already been in Denon receivers for a couple of years. The main difference was that the version of XT in the stand-alone box had 8x the resolution for subwoofers and 2x the resolution for main channels compared to XT in receivers, making it the best version of Audyssey room correction to use for comparison purposes at the time. Harman's room correction (ARCOS) wouldn't show up in Synthesis systems until the year after the Audyssey SE was discontinued, so one stand-alone unit wasn't competing with the other, either in time line or price range (not even remotely close, going by audiophilesavant's post).
Audyssey's business model since then has been in licensing, which Harman doesn't do with any of their proprietary technologies, so they're not competing there.
The first consumer unit planned to have ARCOS was the next Lex processor, which was slated to cost twice as much as the most expensive Audyssey equiped pre-pro (Denon A1), so again one was not going to compete in the other's price range.
The next model planned to have ARCOS was rumored to be Mark Levinson's next pre-pro. You really think that boutique piece is intended to be a competitor to any Audyssey-equipped gear?
At the time of the comparison and now, ARCOS isn't competition for Audyssey, neither in time line nor price range nor licensing.
You keep saying that Audyssey's CI-based product was in a lower price range than the Harman product, and therefore it wasn't a competitor. However, if someone wants to sell a more expensive, but similar product, to a technically-savvy market like the CI market, they need to convince those potential customers that the money spent buys some benefit. What better way to do that than to conduct an in-house (pseudo) "scientific" study that discredits the lower-priced competitor as "no better than no EQ"? Then present it at an AES convention as if it's some revolutionary new knowledge. How does this get past "peer review"?

Sanjay if the *real* purpose of this study was to elucidate the "most preferred" target curve, the correct, "scientific" way to do that would be to use ONE RC system. and then change the target curve within that system. The "wrong" and most unfair way to do it would be to bring in a couple of competitors systems, set them up badly, and then compare them for preference to your own system... which has been tweaked and optimized to be as advantageous as possible. Of course, the latter method is absolutely assured to get you the results your Marketing Department wants... even though those results have no scientific validity whatsoever.

Craig

Lombardi said it:
Perfection is not attainable, but if we chase perfection we can catch excellence."

My System

craig john is offline  
Old 10-23-2012, 11:17 PM
AVS Addicted Member
 
Chu Gai's Avatar
 
Join Date: Sep 2002
Location: NYC area
Posts: 15,067
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 632
If the Audyssey product in that test was only available to custom installers, how did Harman get it?

"I've found that when you want to know the truth about someone that someone is probably the last person you should ask." - Gregory House
Chu Gai is offline  
Old 10-24-2012, 12:05 AM
AVS Special Member
 
J_Palmer_Cass's Avatar
 
Join Date: May 2005
Posts: 6,495
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 294 Post(s)
Liked: 107
Quote:
Originally Posted by craig john View Post



This graph totally explains this whole problem. Look at the un-EQ'd response, (which is the black dashed trace.) It rolls off from the peak dramatically at about 50 Hz. There must be a cancellation in this range. Audyssey, which is the curve just below the un-EQ'd response, has knocked down the peak, and then has a very similar roll-off. This is because Audyssey looks for the -3 dB point of the sub and does not correct below the -3 dB point. It would need to BOOST below the -3 dB point to provide correction, and boosting below the -3 dB point can be detrimental to subwoofers. Therefore, it is doing exactly what it is supposed to do.

The question is, should the actual -3 dB point be different than what Audyssey measured. This is certainly possible and may have been correctable with some different measurement locations. We don't know anything about how Audyssey was implemented in this system, so it is certainly plausible that this was a bad implementation of Audyssey. They may have been measuring close to a wall or they may have only used 2 or 3 measurements. We just don't know.

In addition, in the Audyssey curve, there is a dip through the range of the crossover, (the crossover frequency was 80 Hz.) Gee, that looks familiar...



I am certain that dip can be corrected with a slight adjustment of the subwoofer distance setting. I have done this in NUMEROUS system I have set up with Audyssey. Audyssey was said to sound the most "thin" and the least "full" in this comparison. Correcting that dip would have made it sound much fuller and less thin.

However, also take a look at the top tracing, the "most preferred" curve. It has a rising response all the way to 20 Hz. I would be very curious to know how they accomplished that, when this was the subwoofer:



The speakers couldn't do it. They're ported with a port tune in the mid 30's. Besides, they're crossed out at 80 Hz. So they had to do that with boost in the subwoofer. I can find ZERO information on the web on the HB5000 subwoofer. I presume it's a sealed design, although that hump in the response centered around 36 Hz leads me to believe there may be a port in there. If it's a ported sub, you can't boost it below port tune. The driver is unloaded and it will distort like crazy. It can even be damaged. Even if it were sealed, they would need significant boost to add that much output in the lowest frequencies. It looks like it's down about 6 dB at 20 Hz. An LT circuit would do it if the sub is sealed, but you need enough amplifier headroom and driver excursion to allow it to work. If the roll-off is caused by the room, you can't overcome that with EQ or boost

In addition, the unEQ'd curve has a narrow null at 200 Hz. None of the other curves had much impact on that null, probably because you can't EQ out a null. Yet in the "most preferred" trace that null is somehow magically gone. How'd dey do dat?

The point I've been making all along is that Audyssey needs optimization. It's clear to me that Audyssey was NOT OPTIMIZED in this implementation. IT is also blatantly obvious that the JBL system was totally optimized. The result of comparing these two products in these two implementations is totally and completely predictable... and worthless.

Craig



Why is it clear to you "that Audyssey was NOT OPTIMIZED in this implementation"? Maybe it is just you having your blinders on. Was a single seat optimized or 6 seats? Some people use a Room EQ system to EQ more than one seat.








as compared with



J_Palmer_Cass is offline  
Old 10-24-2012, 12:43 AM
AVS Special Member
 
J_Palmer_Cass's Avatar
 
Join Date: May 2005
Posts: 6,495
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 294 Post(s)
Liked: 107
JBL HB-25000 specifications from the web. Should be easy to EQ flat to 20 Hz.


J_Palmer_Cass is offline  
Old 10-24-2012, 01:19 AM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 20,211
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 1679 Post(s)
Liked: 1131
Quote:
Originally Posted by craig john View Post

The biggest problem with the Olive study is that guys like Amirm will try to use it as "proof" that Audyssey is a bad system and "worse than no EQ." He's already done so in this thread multiple times. Your attempts to legitimize this "study", unfortunately, do the same, in spite of your disclaimers:
He is one of the reasons I posted the disclaimer. The other reason was to make my views clear for those who feel the findings of this comparison have to do with hardware rather than preferred frequency response. All I can do is state my position explicitly: "I don't use Olive's 2009 comparison to comment on current gear". Despite me posting that, if you're still objecting to my posts because of something being done by Amir, not me, then I can't help that.

If I'm legitimizing the Olive comparison, it is because I consider the findings legitimate. For all the objections in this thread, no one has posted anything to refute Olive's findings that downward sloping and smooth response is preferred.
Quote:
Originally Posted by craig john View Post

I have tried to show that there is ample opportunity to mess up an Audyssey calibration without intentionally sabotaging it.
If that is true, shouldn't a comparison of automated room correction technologies (not the boxes they come in) reflect the fact that some of them are not as automated as others?
Quote:
Originally Posted by craig john View Post

If you don't do anything to optimize the Audyssey result, if you just accept the "out of the box" result, then it is absolutely possible to get a bad result.
Why is that Olive's problem and not something that Audyssey should have addressed? Accepting the 'out of the box' result didn't keep the Lyngdorf unit from being preferred to no EQ.
Quote:
Originally Posted by craig john View Post

We can be certain that Harman would have no motivation to optimize the Audyssey result. OTOH, we can also be certain they they made every effort to ensure their own product was set up optimally.
They had an approach that gave higher priority to fixing the most audible problems, typically in the lower frequencies, which I detailed in my reply to Markus. Audyssey would end up validating that approach a few years later when they released XT32. The technique works; just that Harman's prototype was doing it at the time and Audyssey wasn't. This wasn't something that needed to be "set up optimally", simply let the algorithm do what it was programmed to.
Quote:
Originally Posted by craig john View Post

Bottom line, this "study" doesn't PROVE or DISPROVE anything about Audyssey.
Exactly! So why the consternation? If you told CalMan that you used their video calibration software in a preference-based comparison, and that it didn't score well because viewers preferred an image that was bluer, brighter and more contrasty (like what they do to TVs to stand out in show rooms), they would have no reason to get defensive. After all, their goal is reference, not preference. Likewise, Audussey has always promoted their goal as reference rather than preference. So why would a preference-based comparison have any bearing on Audyssey?
Quote:
Originally Posted by craig john View Post

Whether you want to paint Audyssey and Harman as non-competitors, it is abundantly clear that they were comparing their own product to what they perceived as a "competitor's" product.
I didn't simply "paint Audyssey and Harman as non-competitors" but instead showed why they were non-competitors, detailing marketplace time line (products didn't overlap), price range (not even close) and business model (licensing vs no licensing). Look over those reasons, and if you find anything that's not factual, let me know.

Sanjay
sdurani is offline  
Old 10-24-2012, 06:03 AM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,530
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1205
Quote:
Originally Posted by Chu Gai View Post

If the Audyssey product in that test was only available to custom installers, how did Harman get it?

I have to admit that the answer to this question is so obvious to me that I'm stuck trying to figure out what its second meaning is! ;-)

On the worst day of its life, Harman just calls a few of their custom installer clients and finds one that i also represents Audessey. On a good day, they just call up Audessey and make a deal.
arnyk is offline  
Old 10-24-2012, 06:07 AM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,530
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1205
Quote:
Originally Posted by J_Palmer_Cass View Post

JBL HB-25000 specifications from the web. Should be easy to EQ flat to 20 Hz.

IME there's nothing on that spec sheet that gives a reliable answer as to whether or not the device has enough undistorted output at 20 Hz to be worth trying to equalize that low. The missing hardware spec is woofer driver Xmax, while the missing performance spec is maximum SPL output at 20 Hz at say, 10% THD. As it stands, it is a pig in a poke.

My gut feeling is that it would take at least 2 of them to be marginally useful in a room the size and with all of the absorbtive materials that were probably in the room that the test was run in.
arnyk is offline  
Old 10-24-2012, 06:11 AM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,530
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1205
Quote:
Originally Posted by sdurani View Post

H So why would a preference-based comparison have any bearing on Audyssey? I didn't simply "paint Audyssey and Harman as non-competitors" but instead showed why they were non-competitors, detailing marketplace time line (products didn't overlap), price range (not even close) and business model (licensing vs no licensing). Look over those reasons, and if you find anything that's not factual, let me know.

Now, you're pretty well echoing the arguments I made and you dismissed a few days and pages back. ;-)

Glad to see you come to your senses. Looking forward to see how you dismiss this post, as well.
arnyk is offline  
Old 10-24-2012, 07:53 AM
AVS Special Member
 
J_Palmer_Cass's Avatar
 
Join Date: May 2005
Posts: 6,495
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 294 Post(s)
Liked: 107
Quote:
Originally Posted by arnyk View Post

Quote:
Originally Posted by J_Palmer_Cass View Post

JBL HB-2500 specifications from the web. Should be easy to EQ flat to 20 Hz.

IME there's nothing on that spec sheet that gives a reliable answer as to whether or not the device has enough undistorted output at 20 Hz to be worth trying to equalize that low. The missing hardware spec is woofer driver Xmax, while the missing performance spec is maximum SPL output at 20 Hz at say, 10% THD. As it stands, it is a pig in a poke.

My gut feeling is that it would take at least 2 of them to be marginally useful in a room the size and with all of the absorbtive materials that were probably in the room that the test was run in.




That was the subwoofer that was used in the Harman test, and one must assume that the people who ran the test knew what they were doing. The two main speakers that the subwoofers had to blend with are probably the limiting factor.

The JBL HB-25400 did measure flat to 20 Hz in the test using the Harman EQ. Top two average curves show Harman single seat and multi-seat EQ on both charts.

The Audyssey EQ was pretty flat over 6 seats, but did not have a slope like the Harman EQ. Change the Audyssey slope a bit, and the Audyssey EQ would probably move up in the Harman preferred rating scale.






J_Palmer_Cass is offline  
Old 10-24-2012, 08:55 AM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,530
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1205
Quote:
Originally Posted by J_Palmer_Cass View Post

That was the subwoofer that was used in the Harman test, and one must assume that the people who ran the test knew what they were doing.

One must assume nothing! ;-)

My main point is that we don't know how competent these speakers were @ 20 Hz. Some negative comments have been made, and if I thought I saw evidence that would overcome those comments, I would point it out. AFAIK they are mystery meat. Reasonably competent, yes but exactly what does that mean?
Quote:
The two main speakers that the subwoofers had to blend with are probably the limiting factor.

I'm not buying that. Besides the tests were done in mono, and to me that means one main speaker and one subwoofer. Others have granted this opinion - mono means one each woofer and main. I think that two subs might have been what most of us would call a better match to the room, but that isn't how the test apparently was run.
Quote:
The JBL HB-25400 did measure flat to 20 Hz in the test using the Harman EQ. Top two average curves show Harman single seat and multi-seat EQ on both charts.

It is a fact of life that in a suitably quiet room, you can run your FR tests at such a low SPL that they really don't reflect on the competence of the subwoofer for general use.

AFAIK the SPL's shown in Olive's paper were artificially displaced in the vertical direction so that the FR curves didn't fall on top of each other. IOW, they don't reflect real world SPLs.

We're back to a choice of targets. IME a little excess bass in the lowest octave or generally two elicits favorable reactions from most audiophiles, within reason. It looks like the Harman auto-eq box was skewed in that direction. The listener preferences may have been more along the lines of crank 'er up than more accurate reproduction.
arnyk is offline  
Old 10-24-2012, 09:11 AM
AVS Special Member
 
rnrgagne's Avatar
 
Join Date: Jan 2003
Location: Canada
Posts: 6,635
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 30 Post(s)
Liked: 53
Quote:
Originally Posted by Chu Gai View Post

If the Audyssey product in that test was only available to custom installers, how did Harman get it?

Well it was a pain, but if one wanted to go through some hoops it was attainable. Trained CI's only was more like guidelines than a hard rule, I was able get my hands on one up here. Still I think the point being made is it wasn't a mainstream consumer product, until later in its' lifespan.

I hadn't read arnyk's post prior to posting this and I agree with his assumption.
Randy Bessinger likes this.
rnrgagne is offline  
Old 10-24-2012, 09:21 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,490
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 1069 Post(s)
Liked: 558
Quote:
Originally Posted by J_Palmer_Cass View Post

The JBL HB-25400 did measure flat to 20 Hz in the test using the Harman EQ. Top two average curves show Harman single seat and multi-seat EQ on both charts.
Indeed, not only do we have measurements that confirm that, we also have listening test results. Here is the graph showing preferences again:

viewer?pid=explorer&srcid=0B97zTRsdcJTfY2U4ODhiZmUtNDEyNC00ZDcyLWEzZTAtMGJiODQ1ZTUxMGQ4&docid=6acfbb12ff81eb2c6a886a651d8b41c8%7C17ef4ac2e0c6e985ad557a5d306c89c9&a=bi&pagenumber=18&w=800

The low scale is 33 Hz. Clearly if the sub was beating its brains out there, the top performing solution, RC1, would not garner near ideal scores of zero. So both the measurements and double blind tests confirm the same thing. Is the suggestion that we should throw out both and substitute our our guesses and predisposition that the results "can't be?" Should this be the way we judge other double blind results and measurements?

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is online now  
Old 10-24-2012, 09:22 AM
AVS Addicted Member
 
Chu Gai's Avatar
 
Join Date: Sep 2002
Location: NYC area
Posts: 15,067
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 632
Quote:
Originally Posted by rnrgagne View Post

Well it was a pain, but if one wanted to go through some hoops it was attainable. Trained CI's only was more like guidelines than a hard rule, I was able get my hands on one up here. Still I think the point being made is it wasn't a mainstream consumer product, until later in its' lifespan.
I hadn't read arnyk's post prior to posting this and I agree with his assumption.
I'm just wondering if the Audyssey product was configured by trained installers for the purposes of the test.

"I've found that when you want to know the truth about someone that someone is probably the last person you should ask." - Gregory House
Chu Gai is offline  
 
Thread Tools


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off