Toole & Olive are examples of scientists pushing the frontiers of research into audio reproduction, but their are still holes in their studies — and they, so far, are mostly un-duplicated… one of the key features of a scientific finding. Very few audio studios have been verified by an independent study — one of the few i have heard about is Kunchner’s finding that humans can detect a 5 uSec delay (at least at higher frequencies). Reading the tribulations that he went thru to make the experiment work are very interesting from the POV of blind listening tests.
One of the key things i found that showed the short-comings of the Toole/Olive studies is the “assumption” they seem to make that all amplifiers of a certain quality sound the same. This is clearly a hole in their research, so what they have found should be a guide, but not a bible.
dave
Well said!
There's no shortage of subjectivists who are making extraordinary claims in order to sell their products. That's fine. I just wish they would back up their extraordinary claims with extraordinary evidence. "My wife came running from the kitchen..." does not – and should not – have the same validity as scientific evidence. I also think it illustrates somewhat outdated gender roles, but that's a separate discussion.
Tom
Oh? How so?
The research I've seen from Toole & Olive, including what's cited in Toole's book on sound reproduction, shows that listeners prefer audio equipment that measures well. I.e. they give such equipment higher ratings of fidelity. This is the case for trained listeners, audio professionals (studio engineers, etc.), audiophiles, and untrained listeners. The fact that they give two pieces of equipment the same rating does not imply that the pieces of equipment "sound the same" only that both pieces deliver an experience with high fidelity.
Interestingly, blind tests correlate well with the measurements. I.e. good measurements -> significantly higher fidelity scores by participants. If you make the listening test sighted, more expensive products and prettier products tend to score higher in the listening tests. You can find descriptions of the experiments and the results in Toole's book: Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers. At least to me, that's a pretty darn clear indicator of bias in sighted trials.
Tom
Last edited:
It so happens the speakers that do well in DBLTs are liked by ALL listeners. From the headbanging heavy metal fan, to your girlfriend/wife/mum, to the experienced speaker designer, to the classical recording engineer, to the nerd who designs his own microphones and makes his own recordings ....
We were quite surprised when this emerged.
Actually Floyd & Olive duplicated and confirmed some (but not all) OUR work which pre-dates them by a decade or so. At least they do reference us.
It also happens that the best speakers in DBLTs all have 'flat' anechoic responses. Our view of that is in AES E-Library >> Absolute Listening Tests-Further Progress
BTW, this isn't how we conducted our DBLTs. It was just an important step along a journey that still hasn't reached its end.
On LF bandwidth, the best performer in nearly 2 decades of DBLTs was a 6 ltr box. It has no real response below about 60Hz. Only a few recording engineers and speaker designers pick this up but still say it has "excellent musical bass definition bla bla"
I've never said measurements are useless and indeed have spent much of my previous life helping to develop tools which facilitate 3) It's an important part of my Room Interface Profile but there's more that no one has addressed.
These days, when I see a speaker with excellent measurements to my criteria (eg Gerrit Boers' The Totem of Tone, an active 3-way dipole and active subwoofer), I itch to put it in a DBLT against the best I've encountered. Alas, its VERY expensive to do proper DBLTs and I'll probably never have the tools again to do this. I take issue with much of how Floyd & Olive conduct theirs.
You should NEVER measure before you listen to a speaker. Even the best ears are prejudiced. There are people who's opinion I trust .. but only in DBLTs
Last edited:
Indeed thay have gotten very interesting results from measurements correlation. That they say any speaker measure done outside of an anechoic chamber should inform us of the limits of typical diy testing (and a lot of pros).
But the lack of inclusion of speaker/amplifier synergy in their experiments introduces a significant variable unlooked at.
Speakers & amplifiers should never be considered separetly. For instance, this summer, your amplifier connected to our big MTMs sounded fantastic, but, as i mentioned to you, i felt the damping factor too high for the speakers — as opposed to the ACA (run as bridged monoBloks) which had too large an output impedance. The SIT3 slots between the 2 and sounds best despite not measuring as good as your amp. On a different set of speakers things could well be different.
BTW: I got a comment from a friend with much better HF hearing than myself and he had some comments about something wrong up there — but he himself designs and builds amps and has his own predudices.
dave
But the lack of inclusion of speaker/amplifier synergy in their experiments introduces a significant variable unlooked at.
Speakers & amplifiers should never be considered separetly. For instance, this summer, your amplifier connected to our big MTMs sounded fantastic, but, as i mentioned to you, i felt the damping factor too high for the speakers — as opposed to the ACA (run as bridged monoBloks) which had too large an output impedance. The SIT3 slots between the 2 and sounds best despite not measuring as good as your amp. On a different set of speakers things could well be different.
BTW: I got a comment from a friend with much better HF hearing than myself and he had some comments about something wrong up there — but he himself designs and builds amps and has his own predudices.
dave
Can one access your research outside of the AES library? Authours usually have the right to distribute their own papers?
dave
When you start doing DBLTs, you quickly find that most (all?) self declared Golden Pinnae, dem who hear chalk & cheese differences between mains cables bla bla, are Deaf; give random results in DBLTs. You can present them with exactly the same thing thrice in an ABC test and they will still hear huge differences."My wife came running from the kitchen..." does not – and should not – have the same validity as scientific evidence. I also think it illustrates somewhat outdated gender roles, but that's a separate discussion.
The man-in-the-street performs better than most reviewers. The woman-in-the-street is significantly more perceptive (gives more reliable results) than the man-in-the-street.
This is from nearly 2 decades of DBLTs.
So you'd better listen when your girlfriend/wife/mum says of your new toy, "I don't like this as much as your old one."
I think the real hole of his study is he did not publish the study of the mastering studio listening environment. I strongly believe he researched about it but kept it secret.
Mastering studio processes have been esoterically hidden from our consumers purposefully, and unlike RIAA curve, we don't know what is the correct (suggested) response of the speakers in the actual listening room, so some people try to blindly believe that flat response is the right response while the mastering studios’ room and speaker responses are far from flat in most cases.
Hard to argue with those who want evidence that we have not yet learned to measure… but that the ear/brain (at least good ones) has no problem hearin gif it is missing or not.
If you lose information things will not sound as real, or as good. If you can listen to MP3s then it won’t matter to you.
dave
If you lose information things will not sound as real, or as good. If you can listen to MP3s then it won’t matter to you.
dave
planet10 said:
Can you link us to what methods were tried?planet10 said:
I can think of a couple that are possible in da 21st century with new-fangled digits.
Last edited:
There has been plenty of research on 'house curves' and with the advent of DSP, it's easy to experiment with. The B&K curve is a result of plenty of DBLTs, and one that I think sounds pretty good
Likewise, most people like what sounds most 'realistic' to them when it comes to music. Play enough material that they are familiar with, and i'm sure they will individually tend to prefer the best system
Out of curiosity, do you know which system this was? I'm interested to see what speakers people have preferred in the past
Likewise, most people like what sounds most 'realistic' to them when it comes to music. Play enough material that they are familiar with, and i'm sure they will individually tend to prefer the best system
Out of curiosity, do you know which system this was? I'm interested to see what speakers people have preferred in the past
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.