It wasn't quite 15 years ago that I read everything that DDF posted on the old "Bass List." Even among the amazing numbers of Industry Giants on that forum, Dave's thoughts were, IMHO, especially worth reading.
Dave, I'm glad to see you're posting here and I for one appreciate that you're taking time to offer your opinions.
Best Regards,
TerryO
Stop being ridiculous and read what I wrote--if you don't understand it, just ask. We are all adults here.🙄 If you read, you'll see you are wrong on this statement. It's not that I don't understand the complexity of the issues(not even saying I do), you just don't see or refuse to admit the fundamental flaw in your reasoning--or even the possibility of it. I learned about the HRTF quite some time ago right here on this board from Markus/Dr. Geddes. The fact is, I don't see what it has to do with loudspeaker design when there is someone is at the helm of the recording process striving to make a recording sound good to themselves and the end user. See Dave S, Feyz, Dr. Toole, Dr. Geddes, or Dr. Olive for solid and sensible evidence of my claim. I am attacking your rationale, not you. Please understand that before you reply. If you can provide evidence or a rationale that takes into account common recording practices, I'd accept and buy into your theory. I understand accepting data can be a tough thing to do when it doesn't support what you intend to prove, but it's necessary and I'm willing to do what is so.
Dave, that's good thinking! I've never even tried to seriously playback my binaural recordings through speakers. In the near field it could be made fairly close I would think so long as you can hold still--every set of speakers I've measured near field varies a lot! Maybe that would be less of an issue with a head in there d/t a local reduction in crosstalk/comb filtering?😕 I think that's what you were talking about. Thoughts here would be appreciated as well. Sorry for the OT, but it's kind of relevant b/c of SL/DDF's assertions. Anyway, who was that William of Ockham guy and what was he talking about?
Dan
Last edited:
Two speakers can not in theory or in practice establish the auditory sensory experience of anything that was not originally those two speakers (and I am not sure they can do that either). Nor can headphones unless you are talking about people with "standardized" pinnas and whose neck has been temporarily paralyzed in order to improve the fidelity.
My point is not just to emphasize this obvious obstacle or to complain about endless hokey math based on assumptions so reductionist as to render the analysis nonsense, but the question is where do we go from here to produce good sound in our music rooms.
Last edited:
You'll get no argument from me there Ben. In fact it seems you're support my standardization point to some degree.
To me anything beyond a smooth on axis with what we've come to know as 'constant directivity' seems like a difficult proposition, where the 'constant directivity' is fairly simple in comparison. The only people I know who don't believe it had their minds made up prior to trying it. Which takes us back to the last sentence of the first paragraph in my previous post. I just can't see how getting further from the conditions of the original recording improve your chances of good sound.
Where to go from here is a good question. That would be for several great minds including the recording end of the question, but I think some type of SS is the probably the answer. Way out of my league for sure, but some sort of standardization would have to be important.
Dan
To me anything beyond a smooth on axis with what we've come to know as 'constant directivity' seems like a difficult proposition, where the 'constant directivity' is fairly simple in comparison. The only people I know who don't believe it had their minds made up prior to trying it. Which takes us back to the last sentence of the first paragraph in my previous post. I just can't see how getting further from the conditions of the original recording improve your chances of good sound.
Where to go from here is a good question. That would be for several great minds including the recording end of the question, but I think some type of SS is the probably the answer. Way out of my league for sure, but some sort of standardization would have to be important.
Dan
Building into speaker sound correction for head-effects follows from the reductionist math. But it results in perceptions which are "anomalous" (that's psych-speak but take it as a kind of euphemism). You don't want listeners sitting there and having to deal with anomalous perceptions - so better to not do it.
It is anomalous in the same sense that 3D movie images can fly out of the screen and over your head... and make some people vomit. The binocular cue is in conflict with the other visual cues which provide their testimony that the image is on or behind the screen while the binocular cue is sometimes (during show-off moments) saying otherwise.
Likewise when you build a "second head" into the sound - which confuses the real head.
It is anomalous in the same sense that 3D movie images can fly out of the screen and over your head... and make some people vomit. The binocular cue is in conflict with the other visual cues which provide their testimony that the image is on or behind the screen while the binocular cue is sometimes (during show-off moments) saying otherwise.
Likewise when you build a "second head" into the sound - which confuses the real head.
I don't know anything about efforts at standardization of rooms or recording so this may say nothing new.
Seems to me tonight, we need to start with conceptual control of three sound streams in addressing, as a system, recording and playback.
1. Direct-to-sweet-spot sound. Kind of like headphone sound.
2. Sound that reaches the listener during the 10-millisecond or so blur zone. We should all agree it is a bad thing and set up speakers and rooms (including recording studios) to minimize it because it can't otherwise be controlled.
3. Later reverberant sound which adds crucial virtues extolled by Toole. At home, that is handled as is a kind "Richness adjustment knob," finally in the hands of the listener; speakers, placement, room acoustics, EQ, etc. are the means of adjustment.
#3 relates to Toole/richness. #1 (which should be sort of flat) and #3, when mixed in the listener's room are partly the basis to the listener's decisions about flatness.
So how does the recording engineer deal with #3?
Seems to me tonight, we need to start with conceptual control of three sound streams in addressing, as a system, recording and playback.
1. Direct-to-sweet-spot sound. Kind of like headphone sound.
2. Sound that reaches the listener during the 10-millisecond or so blur zone. We should all agree it is a bad thing and set up speakers and rooms (including recording studios) to minimize it because it can't otherwise be controlled.
3. Later reverberant sound which adds crucial virtues extolled by Toole. At home, that is handled as is a kind "Richness adjustment knob," finally in the hands of the listener; speakers, placement, room acoustics, EQ, etc. are the means of adjustment.
#3 relates to Toole/richness. #1 (which should be sort of flat) and #3, when mixed in the listener's room are partly the basis to the listener's decisions about flatness.
So how does the recording engineer deal with #3?
Part of the discussion seems to be about changing our desired response based on HRTF issues. I'm not sure that is called for.
A little background for those that aren't up on all the terminology: HRTF is usually measured by placing small probe microphones at the subjects ear canal. Place the subject in an anechoic chamber and fire a reference speaker at them. Rotate 10 degrees and repeat. You will see a few effects, first, the head is a blocking object of a certain size so there is a natural diffraction around it. Sounds on the opposite side will diffract around, more so for low frequencies and less so for upper frequencies. Secondly, the outer ear is a reflector and waveguide of sorts, so there will be minor directional effects due to that. Frontal sounds will be slightly different relative to rear arriving sounds. The pina (end of the ear canal flap) plays a small role in this too. This front to back difference isn't very strong and our ability to discriminate between sounds in these directions seems to be more related to small subconscious head turnings than response effects.
The end result is a very messy frequency response, especially if the ear canal is included with a large peak (typically) at 3kHz.
Similar curves can be taken in a diffuse field, such as a reverberation chamber or, equivalently, you can average the curves of measurements in every direction. The Killian curve is a typical diffuse field measurement to the inner ear location. Note that these curves can only be averages because every ear and head shape and size is likely to be different.
Now, these curves look very non flat with great variation with angle, but are ever present. They represent the response of our everyday experience of sound. They are there whether we are listening to real music or music reproduced over speakers. As such there is no reason to worry about them or compensate for them.
One slight exception comes from the creation or virtual sound sources. If a speaker 30 degrees left and another 30 degrees right are given the same signal in the hopes of creating a well centered virtual source, then we may have the illusion of a centered source with the frequency response of a pair of 30 degree sources (in a more complicated way we also have slight comb filtering at each ear with each ear hearing a near source and farther source).
DDF showed some plots of this earlier. The response at 30 degrees isn't the issue so much as the difference between 0 and 30 degrees. By his calculations it was a mild peak of about 2.5 dB at 3000 Hz.
Should we compensate for that? The difficulty is that sources meant to be perceived from the left or right wouldn't need compensation. A source at the left speaker will be perceived with the same HRTF that a true source at that location would have. So compensation would make frontal centered sources more accurate and left or right channel sources less so.
As much as the HRTF varies with angle, I'm not sure I ever notice frequency response varying as I turn my head. Maybe it is another thing that we live with daily and have gotten used to. It is natural so we accept it. Beyond this, I'm not sure I want to get upset abut a 2.5dB peak that will be present for some reproduction angles but not others.
David S.
A little background for those that aren't up on all the terminology: HRTF is usually measured by placing small probe microphones at the subjects ear canal. Place the subject in an anechoic chamber and fire a reference speaker at them. Rotate 10 degrees and repeat. You will see a few effects, first, the head is a blocking object of a certain size so there is a natural diffraction around it. Sounds on the opposite side will diffract around, more so for low frequencies and less so for upper frequencies. Secondly, the outer ear is a reflector and waveguide of sorts, so there will be minor directional effects due to that. Frontal sounds will be slightly different relative to rear arriving sounds. The pina (end of the ear canal flap) plays a small role in this too. This front to back difference isn't very strong and our ability to discriminate between sounds in these directions seems to be more related to small subconscious head turnings than response effects.
The end result is a very messy frequency response, especially if the ear canal is included with a large peak (typically) at 3kHz.
Similar curves can be taken in a diffuse field, such as a reverberation chamber or, equivalently, you can average the curves of measurements in every direction. The Killian curve is a typical diffuse field measurement to the inner ear location. Note that these curves can only be averages because every ear and head shape and size is likely to be different.
Now, these curves look very non flat with great variation with angle, but are ever present. They represent the response of our everyday experience of sound. They are there whether we are listening to real music or music reproduced over speakers. As such there is no reason to worry about them or compensate for them.
One slight exception comes from the creation or virtual sound sources. If a speaker 30 degrees left and another 30 degrees right are given the same signal in the hopes of creating a well centered virtual source, then we may have the illusion of a centered source with the frequency response of a pair of 30 degree sources (in a more complicated way we also have slight comb filtering at each ear with each ear hearing a near source and farther source).
DDF showed some plots of this earlier. The response at 30 degrees isn't the issue so much as the difference between 0 and 30 degrees. By his calculations it was a mild peak of about 2.5 dB at 3000 Hz.
Should we compensate for that? The difficulty is that sources meant to be perceived from the left or right wouldn't need compensation. A source at the left speaker will be perceived with the same HRTF that a true source at that location would have. So compensation would make frontal centered sources more accurate and left or right channel sources less so.
As much as the HRTF varies with angle, I'm not sure I ever notice frequency response varying as I turn my head. Maybe it is another thing that we live with daily and have gotten used to. It is natural so we accept it. Beyond this, I'm not sure I want to get upset abut a 2.5dB peak that will be present for some reproduction angles but not others.
David S.
Last edited:
Yes, of course we don't notice it, as you say, in exactly the same sense that we don't "notice" that the visual image on our retina is upside down or that our eyes are in constant jerky (saccade) motion.
I hope I have put to rest one of the screwball aspects of this discussion.
Last edited:
I don't think my perusal of such a document would enlighten me much or necessarily add much to my grasp of playing music at home. Certainly reading Toole's discussions of these things left me numb.
If you have a point to make, could you please simply make it.
Thanks.
Firstly, just to clear up something, in #2, by sound you apparently mean reflected or reverberant sound.
I haven't read Toole, but I've studied recording studio design a bit over the years (for one thing, I bought the Master Handbook of Acoustics 3rd Edition about a year before the 4th came out). An LEDE (live end dead end) room for monitoring/mixing/mastering, which has become a popular design over the last couple of decades, fits these three points very well. The speakers are in the (acoustically) dead end of the room, reducing early reflections, while the live end is behind the listener, providing a bit of reverberation that prevents an overall dead-sounding room.
In my experience (and I imagine everyone else's here), just keeping the speakers several feet away from the walls, at ear level, pointed directly at the listening position goes a long way towards good sound compared to how speakers are placed by more casual listeners, on the floor next to walls or in corners.
I presume you mean as far as what microphones pick up in the recording. The studio is large, has a high ceiling, and the musicians and instruments are placed toward the middle, away from the walls, to delay the first reflections as much as practical. This can include a carpeted floor and/or the musician standing on a riser to reduce the level of and increase the time to the floor reflection.
When adding artificial reverberation, it's easy enough to add a pre-delay of any desired number of milliseconds to increase the time before the first "reflection."
While I understand the desire to make speakers with appropriate off-axis response for whatever room it might go into (after all this is the speaker design forum), it can be frustrating hearing about expensive audio systems put into rooms that may be inadequate to the task. Certainly a well-off person could afford to have a "proper room" built into their house, but even a smaller room can be greatly improved by an acoustically-minded arrangement of furniture, speakers, sound absorbers and diffusers. Of course, insisting a room "look good" where the desired looks are inconsistent with an arrangement for good sound is yet another common problem (often called the "WAF factor"), and is beyond being resolved by current technology.
Interesting.
Depends on the room (and speakers) if LEDE makes sense. Often leads to pretty dead sound when all you need is to attentuate the first 8-10 feet.
I had in mind the recording engineers room (geez, I don't even know what to call things) rather than the mics. Wouldn't the set-up of the musicians - whatever it was for better or worse - be a "given" and not something to be manipulated, unless multi-multi-mic'ing was being practiced?
You know, ESLs and OBs fit the description too because they are usually spaced from walls behind.
Yes, kind of a debate between the practical people who design speakers in isolation (and sometimes have Theologically Correct views on directivity) and those who think you must think system-wise by including the room. In turn, leads back to "flat" or, more to my concern, what are "levers" available to shape response to our liking besides simple tone controls and shelving filters.
Depends on the room (and speakers) if LEDE makes sense. Often leads to pretty dead sound when all you need is to attentuate the first 8-10 feet.
I had in mind the recording engineers room (geez, I don't even know what to call things) rather than the mics. Wouldn't the set-up of the musicians - whatever it was for better or worse - be a "given" and not something to be manipulated, unless multi-multi-mic'ing was being practiced?
You know, ESLs and OBs fit the description too because they are usually spaced from walls behind.
Yes, kind of a debate between the practical people who design speakers in isolation (and sometimes have Theologically Correct views on directivity) and those who think you must think system-wise by including the room. In turn, leads back to "flat" or, more to my concern, what are "levers" available to shape response to our liking besides simple tone controls and shelving filters.
Last edited:
Actually I think Toole was able to show that there was statistically significant consensus on which speaker was preferred among listeners with similar hearing. When the preference outlier's hearing was looked at they tended to be different than the rest of he group. He mentions something about this in his book and makes an analogy between preference and a pair of glasses, each person looking for a speaker that corrects for the particular defect in hearing they experience. Everyone's hearing deteriorates over time in different ways and anyone participating in toole's listening tests had their hearing tested regularly. I believe he mentions having to retire from regular listening tests due to his own hearing changing away from standard.
Statistically speaking, IMO it's highly probable that there is significant variance in hearing ability among those who post here which, in turn, could be the root cause of many disagreements being debated regarding the subtleties of flat vs non-flat response, early vs late reflections, etc., etc., etc...... .
It's almost as though these things are discussed and debated with an understanding that we are somehow comporised of a trained 'golden-ear' group of listeners.
It's almost as though these things are discussed and debated with an understanding that we are somehow comporised of a trained 'golden-ear' group of listeners.
Toole addressed that -- among those free of hearing deficits, preferences were surprisingly uniform; statistically, we all hear the same, basically.
Those with deficits were unable to reliably rank the same loudspeakers in the same order in subsequent tests. They were also the most adamant in defending the validity of their highly variable assessments, as if their very esteem were at issue.... 🙂
Those with deficits were unable to reliably rank the same loudspeakers in the same order in subsequent tests. They were also the most adamant in defending the validity of their highly variable assessments, as if their very esteem were at issue.... 🙂
Almost all Toole's data relates to variability, not central tendencies. Big difference. It is true however, to say that people (and groups) with more variability make poorer judges and that seems (out of proportion) important to Toole. But only in that sense is their judgment "poor." The handful of professional reviewers did not seem to uphold the honor of their profession.
I don't recall that stuff about "esteem" and it doesn't sound like Toole.
I have a hearing defect. It's exactly the point I'm trying to make. Does that disqualify me from any further posting here? I certainly hope not. Do my listening preferences matter? They do at least to me.
The bottom line is you don't know who is a golden ear amongst those who post here.
Toole may have used a control group of 'average' listeners and that was indeed needed to conduct controlled studies. HERE, it's a totally different ball game. People have different abilities to hear, make judgements about sound and off their opinions.
That's why we respect the data more than opinions. Those talking about how stuff sounds and what they like are talking about themselves, not loudspeakers; it's an important distinction to keep always in mind.
Geddes don't play that ... and Zilch don't either.
[Mostly.... 😉 ]
- Status
- Not open for further replies.