I can certainly relate to this. A kitchen range hood sounded more annoying than in a car driving on the freeway (which had 20db more measured SPL).
Guess I am also mixed about Toole's and Olive's work - and I recognize their immensely valuable contributions, particularly Toole's super book (with many subtle jokes in there too).
To Dan's comments about the yuccky choice of over-produced pop crap music I would add my criticism of "training." First, I have never seen a good description of this "training" and hope somebody will enlighten me. Folks in my line of work are very particular about the details. Could it be that they were trained to think HK or Infinity speakers were wonderful and say, Martin Logan's with sizzling treble that burns your ears on over-produced pop crap is terrible?
There is a phenomenon known as "adaptation level" that is ignored too (well, mentioned in passing somewhere). If you spend a while in reddish surroundings, your first moments out will seem greenish. There are few absolute standards in the human senses (mostly in the pain system, by the way) and our judgment is profoundly shaped by the context in almost every perceptual dimension. So if the listeners or "trained" people are listening to garden variety tweeting most of the time, a Martin Logan treble will seem quite exaggerated.
I know many of the more curious minds will want to read about "adaptation level" right away and some will assert it couldn't possibly apply to people like themselves with their powerful engineering minds.
I am sure nobody will wonder why I have chosen Martin Logan... if they have read Olive's comparison study.
To Dan's comments about the yuccky choice of over-produced pop crap music I would add my criticism of "training." First, I have never seen a good description of this "training" and hope somebody will enlighten me. Folks in my line of work are very particular about the details. Could it be that they were trained to think HK or Infinity speakers were wonderful and say, Martin Logan's with sizzling treble that burns your ears on over-produced pop crap is terrible?
There is a phenomenon known as "adaptation level" that is ignored too (well, mentioned in passing somewhere). If you spend a while in reddish surroundings, your first moments out will seem greenish. There are few absolute standards in the human senses (mostly in the pain system, by the way) and our judgment is profoundly shaped by the context in almost every perceptual dimension. So if the listeners or "trained" people are listening to garden variety tweeting most of the time, a Martin Logan treble will seem quite exaggerated.
I know many of the more curious minds will want to read about "adaptation level" right away and some will assert it couldn't possibly apply to people like themselves with their powerful engineering minds.
I am sure nobody will wonder why I have chosen Martin Logan... if they have read Olive's comparison study.
I suspect the Beolab 5 and the B2031A have a somehwat similar sonic signature that reminds me of the other is primarily due to the amplifier. I'm sure nobody is interested in investicating these more deeper. I listed to a conceptual article like this:
An externally hosted image should be here but it was not working when we last tested it.
Driven by a traditional amplifier, it sounded quite nice but not accurate. The measurements showed lots of diffraction. Since the source of diffraction was almost in line with the sound wave path, there is very little one can do to reduce the effects.
It's because you only see the data in the form they want you to see.😀😎Even if the measurements are great(certainly don't look bad), it just shows the bias of sighted listening. You just can't remove it. Hopefully that's what you were getting at--what I've been saying all the while.
...
Dan
Ah, but those measurements are POLARS! And in fact not just one view, but TWO! (..that's right, two for the price of one!) 😀
..eh, the data *is* useful, AND in and of itself it doesn't even appear to have any particular bias. It's just that placing to much emphasis on one category of measurement can sometimes lead you down the path of misery. 😱
"Noise" is something that has no meaning in sound reproduction and the topic at hand. Distortion and spatial perception has.
Last edited:
Heck, most of the Martin Logan stuff has a large beaming panel, chock full of comb-filtering (from the size of the radiating sources vs. the freq.), is a dipole, and is a line source at higher freq.s that transitions to a point source at lower freq.s..
In other words - it *is* exaggerated in the treble at most listening distances! 😛
..Yeah, when I started reading the work on loudspeakers with different directivities from the 80's, some of those "hinted at" conclusions were pretty bad. OK, they were *really* bad. Good thing I wasn't drinking at the time, or I could have been snorting the liquid. 😀
Last edited:
Scott, read Oliver's post. He states it better than I did. I was trying to be polite. He said it politely and stated how it doesn't correlate w/ what Dr. Toole states/demonstrates.
Also Ben brings up an idea about adaptation. Maybe had I endured that room for a few days--they'd have broke in. (again that's a joke)
Dan
Also Ben brings up an idea about adaptation. Maybe had I endured that room for a few days--they'd have broke in. (again that's a joke)
Dan
Here's some info on listener training:
Audio Musings by Sean Olive: Search results for listener training
I know there in more out there and I think it's been posted in this thread. I've wondered much about it as well. The results seem to suggest it works.
Dan
Audio Musings by Sean Olive: Search results for listener training
I know there in more out there and I think it's been posted in this thread. I've wondered much about it as well. The results seem to suggest it works.
Dan
In physics "noise" is usually defined as a signal, or portion of a signal that is nearly devoid of information. In mathematics this mean an autocorrelation of nearly an impulse. However, as Zwicker used to point out, "noise" means "power" to a motorcyclist, so this is "information". But this "information" is all in their heads - their is no real information content there, only what they perceive.
Thanks, on my reading pile ASAP. I've been reading Olive's blogs. Very worthwhile material. Both Toole and Olive are good writers.
Harman considered releasing the computer program they use for listener training to the public but that was some time ago and no news has come out about it
4534 quotes 4533 before it was corrected. As corrected, it's a point in Toole confirming Geddes -- we WANT the delayed, low IACC contralateral reflections as spaciousness cues, but not the early ipsilateral ones, which degrade imaging as well as spectral and spatial quality....
Don't know why bentoronto brought up "noise". It's just misleading. We can't measure "beauty". In sound reproduction we're interested in simpler things that seem to be complicated enough:
- How much distortion is each reproduction device allowed to produce and how can it be measured?
- How does a room's optimal reflection pattern need to look like?
- How does a meaningful standard for audio reproduction (and mixing) need to look like?
Perhaps Markus76 answer this question: if we were designing music reproduction for dogs, would the systems be different? For example, would noise be measured say 50-35,000 Hz? Would the direct/reflecting proportions be different and how would you decide for dogs (or for humans)?
How about this question: why don't music systems have reverberation controls so that the reverberation time in a music listening room can be adjusted to be the same as the long reverberation in a classic concert hall? Seems compellingly logical, eh.
How about this question: why don't music systems have reverberation controls so that the reverberation time in a music listening room can be adjusted to be the same as the long reverberation in a classic concert hall? Seems compellingly logical, eh.
Speaking of measurements, Olive has charts for "percentage correct responses." But it looks like the trained listeners were given 5 choices each trial (flat and four spectral distortion options). So is 50% the random rate or is 20%? Depending on the source, percentage correct ran from high 80s to low 50s.
On some music (like full orchestra and chorus or brass quintet or string quartet), trained listeners had lots of trouble telling any difference reliably even for 3 dB frequency bubbles and cuts. Olive says they often couldn't tell bass cut from treble boost and vice versa. Likewise trouble distinguishing a narrow hole at 500 Hz from a narrow boost at 2000 Hz. I wonder if that will have any significance to Markus?
Olive didn't use HK loudspeakers. His PowerPoint illustration shows ER-1 earphones. Below is their rather horrible compass but Etymotic suggests the "true" sound the human hears is flat. Ummm, must be some kind of HUMAN ears cooking the measurements, eh Etymotic?
Frequency chart
The ER-1 is/was a standard for hearing testing. That may not be determinative for our purposes or good for music but they are highly respected. Anybody remember the remarkable X-ray sound from Beyer DT48s that came attached to Nagra recorders? Plain peculiar for music listening but the ultimate headphones for field recording.
On some music (like full orchestra and chorus or brass quintet or string quartet), trained listeners had lots of trouble telling any difference reliably even for 3 dB frequency bubbles and cuts. Olive says they often couldn't tell bass cut from treble boost and vice versa. Likewise trouble distinguishing a narrow hole at 500 Hz from a narrow boost at 2000 Hz. I wonder if that will have any significance to Markus?
Olive didn't use HK loudspeakers. His PowerPoint illustration shows ER-1 earphones. Below is their rather horrible compass but Etymotic suggests the "true" sound the human hears is flat. Ummm, must be some kind of HUMAN ears cooking the measurements, eh Etymotic?
Frequency chart
The ER-1 is/was a standard for hearing testing. That may not be determinative for our purposes or good for music but they are highly respected. Anybody remember the remarkable X-ray sound from Beyer DT48s that came attached to Nagra recorders? Plain peculiar for music listening but the ultimate headphones for field recording.
Last edited:
Have you ever communicated with a dog about perception? That's an important difference between animals and humans. So my approach for a hypothetical reproduction system for dogs would be the classical "recreate the original wave field at the ears". That approach is known as wave field synthesis but has it's own problems. You can read about it and how it compares to stereo here:
http://hauptmikrofon.de/HW/Wittek_thesis_201207.pdf
The working hypothesis for sound reproduction still is that only a subset of cues need to be presented to the listener in order to recreate genuine perceptions.
There is a lot of artificial reverberation added to single signals in the mixing process. Adding reverberation to a summed multichannel signal (stereo is multichannel) played back by two or more speakers doesn't give the same perception as two speakers in a concert hall. Let alone real instruments in that very same concert hall. The reason for this is simply the differing reflection patterns of the concert hall and the listening room.
Yes, of course. I'll repeat my 3 earlier questions:
- How much distortion is each reproduction device allowed to produce and how can it be measured?
- How does a room's optimal reflection pattern need to look like?
- How does a meaningful standard for audio reproduction (and mixing) need to look lik
These questions might look very simplistic to you but if you really want to answer them you'll arrive pretty soon at a point where our knowledge ends. Important investigations have yet to be carried out.
- Status
- Not open for further replies.