any suggestions for cables? - Page 28

actually, when i see repeating periodicity of the same diffuser in a large array, i immediately think of possible lobing issues and that, that may not sound good.

funny how your bias statement can go both ways

well well well WELL!

I may owe all of you a big apology.

Firstly, and amirs post subsequent to mine might further prove this, it could very well be true that (what I thought) about dbt's from harmon appearing in the thread earlier is untrue! That amir did not give a link to (what I recall) about harmon dbts on room treatments seems to back that up.

Hmm. So as I said, I thought we were all on the same page, now evidently not.

Sorry all. I just assumed.

Now I am not so sure what page I am on! haha. Boy, I must be getting senile. Happens to the worst of us, me.

I would have sworn that most of this 'contrary' argument was based on dbts conducted by harmon (which you can tell from the many ways I phrased the argument). I can clearly recall ages ago amir posting pics of a triangular shaped rotating screen with differing treatments on some sides and a bare wall on the last, so easily rotated in situ so the different wall regimes could be tested in the harmon suite. This was in counter to some peoples problems (including mine) of 'how do you make rapid changes from treated to untreated during a blind test?'.

The 'many ways I phrased the argument' becomes kinda moot if in fact there were no dbts conducted on this question by harmon don't it!

So, apologies for my assumption that those dbts were referenced earlier and hence I thought we were on the same page, and we (or more accurately I) need to confirm or deny my vague recall that harmon did in fact test the fr scenario in their room. If not, then what of those rotatable screens?

Man, old age eh?

More apologies forthcoming if there are no dbts at all!

Thanks Joe for the earlier 'support' in your post, but hey I really hope you are not including me in your last line there old chap!

I am afraid I fall firmly and entirely into 'speakers and room, forget the rest' camp. Not even a little toe makes it past that line in the sand

As I have said few times, I'd be willing to have a wager that not only have I spent the least of any in this thread on cables (certainly on a pound per pound of system basis), I'd also wager that my room is more heavily treated than anyone else in this thread.

My between the line points could have taken a bashing recently! Hopefully that will be cleared up soon.

Anyway, cheers. Just needed to clear up that little confusion

Nope; all of Toole's and Harman's research on the advantage of not attenuating first reflections is based on sighted listening tests. Speaker testing is based on DBTs, but not the acoustics research.

So here's the question to the subject under test: Do you prefer the sound when the second speaker you see over there on the side wall is simulating a first reflection or when it is silent? No possibility of bias?

Bias is always a possibility.

In some cases the audible difference is so obvious and gross that it can overcome bias. But that's probably pretty rare.

The other thing is that most ways of dealing with first reflections have more than just the desired consequence.

Well put. Not stated often enough.

you could very well be right about how they conducted the acoustics research, that is exactly what I am waiting for confirmation on. If those results were not blind then I will withdraw completely any and all of my earlier 'objections' and 'human observations'.

Just very quickly thinking about your question (so no rigour, have to dash) it is still possible in some way to get decent enough results? Sure we might see a speaker over 'there', but unlike a tv say we can't tell if any given speaker is getting a signal. (there could be a dozen speakers visible, maybe spread at different angles). Maybe you can press a button with prefer more, prefer less, whatever in response to what you hear (and you have no idea of which speaker or how much intensity, things like that).

In any case, any and all of my objections have been based on the idea they were blind. If not then all bets are off.

SOME of my objections still remain in the sense that (see just above two posts) there is still this lingering think that 'if two things sound different enough then bias can be overcome and blind tests are therefore not needed'. Obviously they will not agree with me on that, yet you (AS) have pointed out that bias can still be present with the harmon acoustic tests because it is sighted. Obviously if the signals are loud enough in the harmon sighted test 'it becomes subject to bias'. YET, it fulfills all of the preconditions needed 'to not have to be done blind'

Namely In some cases the audible difference is so obvious and gross that it can overcome bias. Direct question to you AS then, do you think that bias is present even in the face of gross differences if tests are done sighted? You will be going against the group, but what is your honest answer?

Anyway I can surmise that in the harmon tests the audible difference was sufficiently gross that bias was overcome. Yet those results are still rejected. Why? We saw that the conditions needed to prove dbts are not required were present, so I can (still) only conclude that the reason it is rejected is because of the conclusions.

So what is it guys. We only accept the conclusions we want?

Did anyone read the moulton link. There were very clear obvious differences in the speaker under test yet bias completely clouded his thinking.

The entire premise of blind testing loudspeakers shows that we ARE clouded by bias, even in the face of clear and obvious differences. That it was useful in developing metrics that can in large part indicate preference was a seperate benefit.
Edited by terry j - 1/25/13 at 1:56pm

very quickly

can we draw a line between 'normal audiophile preference testing' and research?

Where I am going is that they are two different activities are they not? We can assume there is a completely different mindset at least.

Is it not possible then that we can still get academic results in the research case even if sighted? At least we can assume the researchers are disinterested where we can safely bet the audiophlie is not.

I am thinking of *all* the commonly accepted bits of research we rely on, say fletcher and munson, hass affect blah blah blah. A whole screed of them collected over the years.

It would not be the case they were all done double blind surely. I mean I doubt the concept of double blind in audio tests was even around back then (when was back then btw??)

Just a quick question to ponder, would love your thoughts

Toole's paper: http://www.aes.org/tmpFiles/elib/20130125/13686.pdf

The research that Toole cited that Toole was listed as an author of (no idea if this conclusively means it was done at Harmon or not, nor if that really matters)

• [7] F. E. Toole and S. E. Olive, “The Modification of Timbre by Resonances: Perception and Measurement,” J. Audio Eng. Soc., vol. 36, pp. 122–142 (1988 Mar.).
• [32] S. E. Olive and F. E. Toole, “The Detection of Reflections in Typical Rooms,” J. Audio Eng. Soc., vol. 37, pp. 539–553 (1989 July/Aug.).
• [56] S. E. Olive, B. Castro, and F. E. Toole, “A New Laboratory for Evaluating Multichannel Audio Components and Systems,” presented at the 105th Convention of the Audio Engineering Society, J. Audio Eng. Soc. (Abstracts), vol. 46, pp. 1032, 1033 (1998 Nov.), preprint 4842.
• [58] F. E. Toole, “Loudspeaker Measurements and Their Relationship to Listener Preferences,” J. Audio Eng. Soc., vol. 34, part 1, pp. 227–235 (1986 Apr.); part 2, pp. 323–348 (1986 May).
• [68] F. E. Toole, “Loudspeakers and Rooms for Stereophonic Sound Reproduction,” presented at the AES 8th International Conference (Washington, DC, 1990).

Perhaps the most pertinent to your question is the second paper which can be grabbed here: http://www.aes.org/tmpFiles/elib/20130125/6079.pdf

Read section 2.1 for the test setup. Its also well worth reading at least the summary at the end.

Harmon has shown that even where speakers sound grossly different (ML, B&W, Harman), the test should be done blind to remove bias. So I don't think the differences being gross necessarily overcome bias. Seems like there is a lot of picking and choosing, along with a lot of rationalization, on when blind tests are required. I'd like to see a little internal consistency in the arguments being offered by posters.

Of course!

Sighted testing has a track record for activating just about any kind of bias that a human mind can exhibit.

I used to be an audio sales guy. Back in the day we had speakers from top name manufacturers that had many clearly audible flaws, along with other speakers that weren't too bad, even by the standard of the best modern speakers. People would see the right name on the speaker and it would seem to define for them what a good speaker should sound like. They would buy garbage even when far better products were sitting right next to them, and sometimes for a better price. It had the right name on it and they could see the nameplate.

You mean, you used to do what amirm does?

Test

Your test didn't sound good. Please upgrade your cables and try again.

Nah, it's his keyboard. It doesn't have enough resolution, so all the letters look the same.

It is more than mindset. In the type of blind testing that is talked about in forums all the time, the person is always challenged to identify something. The goal and there is only one, is to determine if he can perform that identification. To the extent that we tell the tester which is which, then the whole experiment loses all of its value. "X" will no longer be "X" in ABX. It will be "A" or "B" as the tester will know the identity. So a little bit of information biases the experiment completely. Although not quite as bad, we don't want to test speakers by showing people the units under test and then ask, "which one sounds better?" We may have handed them the potentially wrong answer by showing them the size, brand, etc. of the speaker.

Now look back at the Ando test where the listener is given one clip and then another with a reflection added and asked which sounds better. The fact that he knows he is evaluating two sounds doesn't give him the final answer we are asking him.

Can the wrong vote still be cast? Yes. For that reason, we don't rely on one person or one test. We look to objective measurements and see if we can correlate the results. We look to how we hear to see if we can correlate the results. We look to other tests done in similar area to see if we can correlate the results. If they all directionally point to the same conclusions then we run with them. Such is the research that I have been putting forward. And how I became convinced of its validity. It was not a single DBT from Harman whatsoever.

Here is an example from, Journal of ASA, "The Active Listening Room: A Novel Approach to Early Reflection Manipulation in Critical Listening Rooms." There, Naqvi, and Rumsey actually performed double blind ABX tests: "The listening test was based on the A/B/X methodology, which is summarized in [16]....In this experiment X was assigned randomly either to have artificial reflections or not, whereas A and B were the “with reflections” or “without reflections” comparisons, assigned randomly."

Here is the final statement in the conclusion of the paper:

"The findings of the pilot experiment were found to be in close agreement with the findings of Olive and Toole’s [6] and Bech’s [9] experiments in terms of the threshold of image shifts in IEC standard listening rooms."

So while Toole and Olive tests were not "blind" their results some 20 years later is confirmed just the same using ABX tests. Threshold shift by the way is one of the things we like about reflections: they broaden the point source of the speaker.

It is important to remember that the bulk of the research we are talking about is accepted as being proper in peer-reviewed journals of Audio Engineering Society and Acoustic Society of America. So clearly if there are major issues, no one has told them! If we take a position here that what is good for ASA and AES journals is not good for us, then that is a lot of pain for us to endure as we go and try to make other arguments in the future. Vast majority of our proof points would go out the window then.

Speaking of such, look at this peer reviewed paper from Zwicker and Zwicker, folks that are more often quoted on psychoacoustic matters than any:

”Using the loudness exceeded in 10% of the time as an indication of the perceived loudness, it can be expected that the speech is 1.2 times louder in the room with 0.6-s reverberation time and about two times louder in the room with 2.5-s reverberation compared with the loudness produced in the free-field condition. This increment in loudness is often very helpful for the intelligibility of speech in rooms as long as the reverberation time does not produce temporal masking, which reduces the audibility of faint consonants appearing in sequence to loud vowels.”

So now we see objective data backing us both with respect to measurements and how our auditory system works that we do benefit from reflections up to a point. Early reflections increase sound power and therefore help with intelligibility. Later reflections which linger on cover up softer sounds so we don’t want our rooms to be “too live.”

Check out this *introduction* section of the Journal of ASA paper, ” The influence of spectral characteristics of early reflections on speech intelligibility,” by Arweile and Buchholz, dated 2011:

” Early reflections (ERs) of a sound in a given environment are characterized by arriving at the listener’s ears shortly (approximately within 50 ms) after the direct sound (DS). They are integrated with the DS in the auditory system, i.e., within a certain time window their energy is added to the energy of the DS. With regards to speech intelligibility the DS and the ERs form the useful part of the speech signal, whereas late reflections are considered detrimental for speech intelligibility. Thus, the effective level of a speech signal depends on the energy of the DS and the energy of the ERs at the listener’s ears. ER [early reflection] energy increases the effective speech level and has been demonstrated to improve speech intelligibility (Lochner and Burger, 1964; Na´beˇlek and Robinette, 1978; Soulodre et al., 1989; Parizet and Polack, 1992; Bradley et al., 2003).”

For someone to say the same thing I did in such strong manner in the introduction of a paper in a peer reviewed journal, you have to assume there is incredible weight behind it. Dismiss it at your own peril as they say.

Understanding speech is critical in movie content where other sounds in the track can mask it (i.e. acts like noise in research). Likewise, when you listen to music, aren’t we looking for clarity and understanding of the singer? Why would we then cast a negative light on reflections in post after post here? What do we believe if it is not the real expert’s views per above? The list is even longer than that: think of masking which covers the distortion if it is within the sound of a louder main tone. That was determined using the methodology we are talking about. Without masking, audio compression would not work remotely as well as it does. And what would we do if we took masking out of our vocabulary in this forum as we argue other distortions in audio? Folks better think hard about what they are saying here .

Keep in mind that it is not like we are comparing sighted tests to double blind tests. Folks objecting have no listening tests to back their point of view whatsoever! The two examples given in this thread were both the opposite of what people thought. And at any rate, were not “DBTs.”

We have a choice: we can continue to believe something is right because we keep reading it on forums, or follow the path of research and science. The latter data is not perfect. It never is even if it were “DBTs.” But it is heck of a lot more perfect than believing what one reads on the Internet.

It is remarkable to me that we don’t remotely accept someone can become a doctor or a lawyer by reading online posts, but somehow we think we can become master acousticians giving advice to others on how that science works, without every setting foot in the journals of ASA or AES. We think all we have to do is read what is posted here. Here is a great example of a post just yesterday:
Bolding mine. Poster is honestly disagreeing with someone in the industry (Nyal) who actually has read some of the research I am talking about by saying he is putting forward what he has read on forums? I know he means well but surely you want to consider for a moment that you can’t trust everything you read on the Internet. And certainly not to the point of disagreeing with folks on that basis alone. For his proof points, he offers a power point by Dr. D’Antonio and an article in EDN magazine. No listening tests, no real papers like I am quoting here, no nothing.

He makes statements like, ” The problem is, "right" isn't necessarily fully defined unless your goal includes an acoustical model in which you are trying to achieve.” Acoustic model? Exactly how does he propose someone determine that? Folks in that thread are struggling to get REW to measure the sound in their rooms. Now they are expected to go and determine their desired “acoustic model?” What is going to happen is that folks are going to be told what the acoustic model is: they will be told to hate reflections and try to measure them with tools that show them improperly no less. Such is life on the Internet.

To be clear, folks there have done a phenomenal job documenting the mechanics of REW there. But they are about to throw out all of it and then some with the non-scientific folklore they have read online.

To treat or not treat first reflections is just a part of the room acoustic picture, and generally when anyone talks about room acoustics it's all encompassing.
I don't believe anyone here ever said " just treat your first reflections and you'll have perfect sound".
I know you're trying hard to put the room acoustic treament camp in the same camp as the snake oil pushers but it's weakest of weak arguments.

you really need to stop assuming the only factor with respect to speech intelligibility is merely gain.

we (ideally) do not have home residential rooms with such poor SNR that the direct signal itself cannot be heard in the presence of a noise floor. we do not require to have the early reflections fused with the direct signal to increase the perceived gain of the signal as we have AMPLIFIED reproduction setups - of which the gain of the direct signal is controlled via the volume knob.

do you live near a train yard? if you are having difficulty with the level of gain of the direct signal with respect to the ambient noise floor, then may i first suggest you focus on sound isolation prior to internal room acoustics.


SNR is only one aspect of 'intelligibility' (the gain relationship of the signal to noise). it does not address the other aspects of intelligibility pertaining to the ability to understand the articulation of the signal.

but don't let that stop you from your ever-expected producing of unamplified "speech room studies".

Thanks amir.

Re the tests (thanks for taking it up, for a while I thought it would be another 'important' point that simply would go missing in action) you mentioned one that I had not included in my 'list'...there are tons of them!...but I spose a good example of a non dbt in this area I was thinking of is the simple one, 'what is the normal range of human hearing'. Just grab people off the street (good chance it fulfils one of my posited requirements..'disinterested') whack headphones on them and tell them to raise a finger when they hear something.

Things get trickier later on when testing things admittedly, but the point is that perfectly valid and usable data can be gotten even if not double blind. So there has to be some evaluation when looking at the question of 'was it db or not'.

Anyway, cheers for that. However, I simply must ask you a direct question please.

In the discussion in the past about 'the harmon findings' on room treatment, I guess specifically FR treatment, were they done in the harmon room blind or double blind? At some stage I came to the conclusion that they were. That they WERE db became my stable point from which to argue. It seems to be getting stronger that I was wrong in that. I'd like to get it clear if you'll oblige. Thanks.

In a very broad sense my stance would not change that much, all of us at some stage will argue a point from emotional or 'religious' grounds (religious simply trying to mean 'I accept that as an article of belief and I am unable to really argue it without resort to 'it simply just is'..') and that will always be true even if there were no actual dbts on this question.

And the findings were (quoted from the summary of said paper [6], bolding is mine):

Sounds reflected within a room have a range of effects on the perception of sounds originating either from live sound sources or from loudspeakers in a stereophonic system. The first sound to arrive, the direct sound, has a dominant effect on the localization of the auditory event, even when the later arrivals are at the same or even a higher sound level. However, this does not mean that the later arrivals have no effect. In fact, depending on the specific demands of the listening situation, reflections may be problematic at levels much below that of the direct sound.

The rest of your post is a continuation of Toole's paper's mistake of only talking in detail about the effect of first reflections on speech intelligibility, an effect that has been well understood for more than 50 years, and hand waving away the differences between speech intelligibility and all other sources/factors.

In other words the listeners were subjected to 2 audibly different volume levels.
No wonder the results were "in close agreement", they were both subjected to different volume levels, blind or sighted.
Not surprising to see such example cited by someone who posted this:

yes - and im still waiting on his personal experiences in all of those ambechoic, LEDE/RFZ, FTB, BBC CID, or NE rooms of which must all have such poor speech intelligibility - as all have attenuated early-arriving high-gain sparse indirect specular reflections.


but then again, i bet the SNR in those rooms are probably sufficient enough (along with the fact that the reproduction setup is of the AMPLIFIED variety), that the gain of the direct signal is of sufficient levels with respect to the ambient noise floor such that it can indeed be "heard".

There are many 'findings" in that paper. Alas, the part you quoted has nothing to do with the specific portion we were discussing, namely the image shift threshold. Further, the point you have bolded is in complete agreement with the position I have been taking and that of Dr. Toole. The confusion comes from not appreciating the term "late" reflections which I have underlined for you above. In acoustic (research) lingo, late reflections are defined in delays of 50 to 100 milliseconds. These represent distances > 50 feet which is not the scenario in our homes with respect to first/strong reflections. To wit, here is data from measurements of the first six reflections by Alan Devantier (in Harman R&D) overlaid on the integration timing of reflections:



See the little black dots clustered together? They are all at < 15 milliseconds and with amplitude which fully allows the integration to happen (i.e. us not detecting reflections as separate events).

It is a very common mistake to throw all reflections together and talk about them as one. You don't want to do that. Late reflections (i.e. as the sound keeps bouncing in your room) can be detrimental. This is why we don't want to listen to music/movies in an empty box. We need to control the later reflections and bring them to a moderate level. This is an independent concept however of what to do with very strong first reflections.
Well, we can put a lot of fault at the feet of industry journals but I am pretty sure the odds of the peer review process not catching a "hand waving" article is pretty low! . While the Haas effect has been known for many years, but sadly constantly ignored here as people keep thinking there are "echos" to worry about with respect to reflections, how the mechanism works and its levels has been the subject of a lot of research. Dr. Toole and Olive's research is but one example of deeper dive. We for example want to know if there is a difference in its effect when the reflection angle changes (there is). We like to know if it changes in real rooms vs anechoic. We like to know what happens when we change the spectrum of reflection which occurs all the time due to speakers not having the same response off-axis than the do on-axis and of course, material such as carpets, walls and acoustic products change the spectrum just the same. And whether the excitation signal makes a difference. On that point, and to address your continued remarks that this is about speech, here is a graph from the paper:


That graph actually came from another research paper by Dr. Toole and Olives and if I am not mistaken, won a publication award from Audio Engineering Society. So I say we can safely say it is not hand waving .

There is also data like this:


We see that the detection threshold of the reflection changes with angle.

In addition to detection level of reflection we also have he image shift threshold and finally, when we hear two distinct events. All of this varies with the test conditions so no wonder so much research has bee focused on it.

All of this said, if you believe all of this is already known for fifty years, then join me in making sure people actually know that. Because as you see, there are constantly cries of denial about it.

BTW, I am still waiting for you to give us quotes from the surround sound abstract you quoted earlier. Shall we conclude that is never going to come and that if we want to search for hand waving, it doesn't get better than putting forward the abstract for a paper you don't have and hence have not read?

Edit: fixed a typo .
Edited by amirm - 1/28/13 at 10:36am

Even better example than mine .
Yes, you were wrong . The DBTs from Harman which we talk about frequently are in the areas of speaker preferences and room eq. I have been to both rooms that Harman uses for this and they are set up for speaker testing only. One uses the the speaker shuffler and then other, is a rotating triangle as you said, but it is used to test in-wall and car audio speakers and such, not acoustic panels.
Your point of view should not change at all. As I noted in my previous reply, even when presenting double blind tests, the data is not accepted any more than other research point. You know this is the case where research in peer review journals is couched as hand waving. I don't mind a top expert in the industry saying that and then going on and giving us their own data that shows different results. But simply declaring it while posting under an alias, with no industry or educational credentials, and in the same post confusing one of research data for another, points to what you said: "it simply just is."

i remember not too long ago when you had to be repeatedly informed on the fact that not all "reflections are reflections are reflections", as your commentary would assume. glad you are learning from your "mistakes".


boy, i wonder if he could tell me how those "1970" (laugh) models handled the later arriving reflections, and just who and what and when and why they were first to implement the QRDs via that from Dr. Manfred Schroeder. and just why the later arriving sound-field has such strict, specific criteria.




tell us again about the mix of non-broadband baby diffusers installed in your company's showroom?
Edited by localhost127 - 1/28/13 at 3:12pm

Cool then. It shows (I think) that we can evaluate the relevance of tests, even if not double blind or even blind in some cases.


Notwithstanding that (at least, if only) you and I know that the thrust of my arguments remain valid, sadly I have been arguing from a wrong position. And what is more, have been doing so for a while now. At times (and this is prob what pains me the most) when someone (from that now shown to be wrong viewpoint) 'squirmed' and avoided the question, I got sarcastic.

Sarcasm as a 'weapon' can be effective in the right place. Sadly, being wrong to start with is NOT a good place!

I have previously apologised 'if it turns out I was wrong'. Well, I was wrong so apologies to all of you, this time without any modifiers. There is enough misconception in audio, that I contributed to it saddens me.

I might just make an observation...back on to my 'interest in human behaviour' hahaha. I have often said the manner in which we argue/discuss can have a large bearing on the outcome. I mean who is gonna change their mind when actually being attacked? Anyways, in this case I think what really impinged on me was the subtle, methodical gentle prodding from X and AS. "What blind tests was that?"

I was always happy to admit I did not have a link as you well know, but as it was asked without scorn or rancour I think it was far more effective. It, the doubt, just slowly wormed it's way in if you follow.

The flip side of this coin? Well as I said for a while now I have been making references to 'the harmon blind studies on room treatments'......why did you not pick me up anywhere along the line? 'What the heck are you talking about terry'?

I got no problem with being corrected.

So you admit to being a pathological liar, and then you complain about being attacked? Christ, man, you deserved every bit of s**t you got, and my only regret is that I didn't dish out more of it myself.

Want to atone? Never post here again. Ever. Nothing less will do.

Liar?

not at all. Pity you have not got what it takes to see it. Mistaken, definitely.

Ah, dishing out ****. Yet you squealed innocent when presented to you.

Who then is the liar?

All can take it or leave it as they see fit. As you have made clear. No, I did not do it to 'atone', I did it because my integrity demanded it of me. (you can look it up if needs be)

Thanks for the chuckle tho, esp the last line, which demands the response ''"of course my leige, I am not worthy of you responding to me"" (monty python voice of course)

thanks. :D

No actually, I did jump on you at some point. You are a perfect example of the sarcasm I spoke of.

So yes, in addition to the general apology to all, a direct sorry to you. FWIW.

That's actually what the designer of my speakers during the early days of his career ( doing psychoacoustic studies ) did... he grabbed people off the streets to participate in the studies.