I'm under the impression all generalizations are worthless but make good sales pitches?
Not everybody wants a flat response, some (like me) find it boring and lifeless... but each to his own of course.
But anyway, whatever you're after, a combination of measurements, simulations, listening and try and error is the best way to get what you want. So surely measure and simulate, but don't get blinded by the numbers only. It's your ears that give the final judgement. But the numbers and graphs mostly show the way to get there and how close you are.
The authors conclude:
"It is stressed that none of these experiments thus far has indicated a present requirement for phase linearity in loudspeakers for the reproduction of music and speech."
Honestly, Earl, on this forum we expect a certain basic honesty from those who post. Just posting a blast with no back-up or qualification and just based on your self-assumed authority to make pronouncements is a slippery slope.
B.
Last edited:
That is not what the paper said. It said that phase was clearly audible in certain situations and that they had 99% confidence that it could be heard in certain musical passages over loudspeakers. This is precisely what I said as well. That phase linearity was not a major effect but could be perceived. Dr. Lipshitz tests were not also conclusive in that they used only certain kinds of phase modifications that were easy to do (they had limited equipment.) More recent tests, like mine, show that group delay, i.e. nonlinear phase, has an audibility that increases with SPL. The Lipschitz tests were done at fairly low SPL levels and as such may not have reached the audibility level. At higher SPLs these effects will always be heard.
Professor Moore showed this same thing in his AES paper on the audibility of group delay.
So please do not make accusations that are clearly false - "we expect a certain basic honesty from those who post"
Excuse me but this statement "no back-up or qualification and just based on your self-assumed authority" is clearly rude since no one here has more qualifications and authority over this subject than I do. Keep insulting me and I'll blacklist you.
Professor Moore showed this same thing in his AES paper on the audibility of group delay.
So please do not make accusations that are clearly false - "we expect a certain basic honesty from those who post"
Excuse me but this statement "no back-up or qualification and just based on your self-assumed authority" is clearly rude since no one here has more qualifications and authority over this subject than I do. Keep insulting me and I'll blacklist you.
Last edited:
Not everybody wants a flat response, some (like me) find it boring and lifeless... but each to his own of course.
I don't think that anyone really likes flat. Harman has concluded that people prefer a slight HF rolloff, especially for CD sources, and a slight bass boost. The idea that room modes constitute a "bass boost" is misguided. If the room is tamed with absorption and multiple subs then bass boost is required since in a small room this will result in a very fast sound decay. A long sound decay in LFs, like we would get in a large room or auditorium, boosts bass perception through the increased time factor. The perception of a signal level increases with signal duration, at least over fairly brief periods. It increase as the signal plays,but then levels off. The small room with short decay times will not have as high a perceived level as the longer decay time room will. This is dominantly a LF effect in a small room since at HFs the ear processes the signal faster than the decay - not so at LFs.
Stan and his friends (from the nearby University of Waterloo (Ontario)) also showed you could tell polarity from its reversal although no one seems to draw your conclusions from that discovery*. Not sure about today, but Stan was a big-time enthusiast for Quad ESLs.
There is always some kind of weird manipulation of a sensory input you can cook up that makes it stand out in some way in a lab. Likewise, it is extremely challenging (as Stan has also shown) to do a perfectly matched A-B comparison even if nominally "blind".
Curiously, you can show that no difference can be heard in your A-B testing; but when people can ID differences from any particular manipulation there could be some subtle mis-match in A or B leading to that. So where Stan says "99%" can hear the difference (but BTW have no preference for one over the other), there may be subtle cues that differentiate them apart, but are not inherent in examining phase changes.
Whatever the physics that might appear to the eye looking at charts relating "group delay" to "phase linearity", they sure ain't the same thing as far as having a driver cone (or ESL diaphragm) start moving in response to a Dirac pulse as compared to the undetectable scrambling of phases in my room. One you might be able to hear under certain test circumstances and one you can't.
B.
* it is explained in terms of "rectification" in the ear
Last edited:
I mostly agree with your post, but high Q correction is not as bad as people say, if the listening point is fixed, and if one wants to hear true to the source reproduction.
I have never read this before. I thought tilting response is mostly due to the circle of confusion, and CD or vinyl does not matter.
I think the only time high Q correction may be ok is if it's correcting a high Q anomaly in the speaker.
Actually I don't totally ban room eq, but I don't see it necessary always. Small rooms with concrete walls often need that. My small HT room has a terrible suckout at spot, but that cannot be eq'd in the first place...
I do high Q dip below 200Hz for a few dB, and it subjectively has a positive effect. The real problem is, it is easy to overdo EQ trying to make the frequency response look flat at one mic position.
Yeah, we can't fix dip anyway. High Q boost is always no no.
The trouble with this test is that almost any change is initially viewed as positive. This is an inherent built in bias that people have. In tests, people prefered when the light level was raised, preferred it again, and then prefered it when it was lowered. They just like the change.
I did this exact same test with a loudspeaker crossover change. Change - better - change again - better, change back to the original - better still. Its just one big circle.
It takes a lot of work to determine what subjective perceptions are real and what aren't. Simple test rarely work. Harman does this very well. Very few others do.
Re: Downward tilt of measured in-room response from Floyd Toole:
A preference for a downward tilting steady-state room curve is the result of two things:
1. beginning with my very first double-blind listening tests in the late 1960s, through the detailed tests in my 1985-86 JAES papers, continuing to this date, the highest rated loudspeakers have had the smoothest, flattest on-axis anechoic response. This is the direct sound.
2. The normal forward-firing configurations of drivers inevitably start out as omnidirectional at low frequencies, becoming progressively more directional at higher frequencies. The rising bass energy yields a steady-state room curve with a downward tilt.
Ref: NORMS AND STANDARDS FOR DISCOURSE ON ASR | Page 7 | Audio Science Review (ASR) Forum
Sean Olive also wrote about this when comparing room correction products: Audio Musings by Sean Olive: The Subjective and Objective Evaluation of Room Correction Products and PDF presentation: The Subjective and Objective Evaluation of Room Correction Products.pdf - Google Drive
The money shot is on slide 24 and a very interesting phenomena about how we perceive in-room frequency response on slide 25. The tilted in-room response is perceived by our ears/brain as flat or neutral. Exploring this a bit more, I wrote a section on this called, "The Science of Preferred Frequency Responses for Headphones and Loudspeakers" and peer reviewed by Sean Olive.
Wrt room eq, from Floyd Toole, "it is now widely accepted that in-room measurements and EQ are beneficial at low frequencies, and that reasonably high frequency resolution is necessary to address room modes." Why don't all speaker manufacturers design for flat on-axis and smooth off-axis? | Page 17 | Audio Science Review (ASR) Forum
I use room eq below 500 Hz given my sig. I also use constant directivity wavegides, which require constant directivity horn eq to compensate.
My experiments show that room eq is good at multiple mic locations, including time alignment of drivers and not just at one mic position. Finally, broad band tilting tone control to adjust to a similar in-room target response as described in the articles above.
A preference for a downward tilting steady-state room curve is the result of two things:
1. beginning with my very first double-blind listening tests in the late 1960s, through the detailed tests in my 1985-86 JAES papers, continuing to this date, the highest rated loudspeakers have had the smoothest, flattest on-axis anechoic response. This is the direct sound.
2. The normal forward-firing configurations of drivers inevitably start out as omnidirectional at low frequencies, becoming progressively more directional at higher frequencies. The rising bass energy yields a steady-state room curve with a downward tilt.
Ref: NORMS AND STANDARDS FOR DISCOURSE ON ASR | Page 7 | Audio Science Review (ASR) Forum
Sean Olive also wrote about this when comparing room correction products: Audio Musings by Sean Olive: The Subjective and Objective Evaluation of Room Correction Products and PDF presentation: The Subjective and Objective Evaluation of Room Correction Products.pdf - Google Drive
The money shot is on slide 24 and a very interesting phenomena about how we perceive in-room frequency response on slide 25. The tilted in-room response is perceived by our ears/brain as flat or neutral. Exploring this a bit more, I wrote a section on this called, "The Science of Preferred Frequency Responses for Headphones and Loudspeakers" and peer reviewed by Sean Olive.
Wrt room eq, from Floyd Toole, "it is now widely accepted that in-room measurements and EQ are beneficial at low frequencies, and that reasonably high frequency resolution is necessary to address room modes." Why don't all speaker manufacturers design for flat on-axis and smooth off-axis? | Page 17 | Audio Science Review (ASR) Forum
I use room eq below 500 Hz given my sig. I also use constant directivity wavegides, which require constant directivity horn eq to compensate.
My experiments show that room eq is good at multiple mic locations, including time alignment of drivers and not just at one mic position. Finally, broad band tilting tone control to adjust to a similar in-room target response as described in the articles above.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.