Flat Response

[Zilch]

Re: Flat Response

Lost ya on this one.

Scroll down to "Upgrade" here:

http://www.linkwitzlab.com/orion-rev3.htm

[MSaturn]

Re: Flat Response

Ligs: I'm pretty sure that's an 'underperforming woofer' or 'incorrectly padded tweeter' kind of thing, not a rolloff issue.

[dantheman]

Re: Flat Response

I think this is true, that the integration time for timbre is pretty short for fusing reflections with the direct sound (Zilch summarized these earlier, and I would add Kates thought it was 5 ms).

One qualifier: integration time of the ear is actually as long as 250ms, to integrate the perceived power.

http://www.google.com/patents/about?id=atgFAAAAEBAJ&dq=6178162

However, reverb itself doesn't universally create more intelligibility or detail retrieval. If its too loud, it lessens it. If too quiet, it doesn't optimize it.

I used to design teleconference rooms and there is a reverberation "sweet spot" for voice intelligibility. The room acoustic behaviour targets had identifiable characteristics for RT60, RT30, vs frequency. So, the frequency variation of teh reverb signature and nature of the decay curve over time all effect intelligibility. It will therefore also effect detail perception.

How the speaker loads the room (ie its dispersion) affects these acoustic characteristics.

I think this concept blends well with my earlier points. Not only should the axial frequency response match the room due to the "power response", but the reverb decay curve vs frequency should attempt to reach a particular target, for optimum detail retrieval. For each different room, the dispersion requirements on the speaker could very well differ slightly!

Just to remind people, that quote of me by DDF was in reference to why diffractive speakers sound worse in reflective environments and why wider polar patterns have more apparent detail. More chances to endure the joy or the pain.

That certainly seems right to me DDF. I haven't read the link yet, but I'd guess to get the sound power you'd need a longer integration time in the bass. Maybe we then hear the rest as reverberation of the original event, but it still integrates in absolute power. I'll check out the link right after I post here. I've got a book on this very subject that I need to get reading.

Dr. Geddes once said in June of last year:

I think that a strong case could be made that we hear LF ONLY in the steady state. It is well know that the ear has an integration time of about 10-20 ms. over which all sound arrivals are integrated into a single event. This corresponds to a period of about 100 Hz, meaning that a 100 Hz signal is basicaly not even recognized by or hearing until more than antire period has ellapsed. How is it then that we could "perceive" transients of these LF signals?

I only ever look at steady state signals at LF because I am convinced that this is all that we can perceive.

Above 500 Hz the situation is quite different and in fact changes 180 degrees - transients and <10ms impulses are the most important.

Do you remember where those optimal "sweet spot" RT60 and RT30 times were? That's a large part of the reason why I urge people to not ignore the room--EQ can't fix temporal issues though I'd bet it can minimize its objectionable characteristics. FR seems dominant in the bass to my ears during playback anyway so I'm not too worried about the rate of decay there. I'd just like to make it sound flat.

Going from 0.4+ to 0.3 second RT60 time it a substantial auditory difference in a small room, but I can't say intelligibility changes much either way--though it should be worse b/c early reflections were reduced. I've read a few books on studio design, but strangely they were fairly light on acoustics and perception. They more like "this is the right way according to.... and this is the right way according to ...." So I've sort of mimicked the general trend of the various "right ways" on the cheap. Of course there is also adaptation to our own rooms to contend with.

Thanks,

Dan

Oh, I don't buy the FM curve issue b/c of adaptation as well--see El Greco Fallacy. IOW we don't need to boost bass and treble various levels at different volumes. If an engineer with normal hearing mixed it to sound flat at a typical listening level, it should sound flat to you so long as you are basically free of gross hearing issues and your speakers and environment are similar--not exact, just remotely similar. Of course there are issues with various engineers being out of touch with the population on many fronts(like the not rolling off the highs that seems like what many classical music lovers want(except me)), but it's the best I can think to do to break the circle of confusion. This is true of all professions--some people just don't understand their market. Since going this route, it's been a long time since I heard a recording I couldn't listen to. Only the 1030A made some recordings sound bad and that was just b/c it exposed noisy mic preamps and overloading d/t it's detailed presentation. Things that really shouldn't be on a recording, but are there. Possibly with narrow pattern speakers, they just won't be as obvious. That's been my experience. That's where I'm trying to find my sweet zone and I believe I have.

Broader than this:

but narrower than this:

something like this:

or this:


There are many ways to improve spaciousness and perceived sound power with any of those designs. As long as the off axis is smooth, whatever bothers you can be fixed.

[MSaturn]

Re: Flat Response

There are even more considerations when it comes to things like dipoles. Thusfar I am building in a lobe that happens a few degrees off-axis, because I know the rear contribution will be greatest near my crossover point. It'll be interesting to see what actually happens.

[dantheman]

Re: Flat Response

You don't say if you've actually *heard* those speakers (in other than a substantially damped and deadened control room) . . . but it's not pretty. Like the B2030A they're OK in a *dead* room . . . in a more "normal" home environment (and sometimes in the control room) anyone interested in "neutral" would down slope 2-4 dB . . .

Or fix their built in problems--too little damping, too much diffraction.

Dan

[dantheman]

Re: Flat Response

Most people think that studio monitors sound awful.

I, on the other hand, think they sound fantastic.

On one hand, everyone else thinks that I'm mad. On the other hand, choosing audio equipment based on simple metrics ("Is it flat? Does it distort? What's the dynamic range?") is a whole lot easier.

You just have to keep them from sight:

http://www.audioholics.com/reviews/speakers/bookshelf/budget-bookshelf-shootout-2009/bookshelf-speaker-face-off-2009-results

Dan

[Deward Hastings]

Re: Flat Response

Or fix their built in problems--too little damping, too much diffraction.

The HF rolloff switch (on the B2030A) solves the "too bright" problem without any change in damping or diffraction. Same also (or so it seems) with the ORION rev3.x.

[Deward Hastings]

Re: Flat Response

By my thesis, elitist euphemism for a listening preference in decline.... ;)

No doubt . . . trash and tastelessness is on the upsurge, and you don't need to go further that the average car stereo or ipod to "hear" that. The desire for quality, either in content or presentation, is indeed in decline.

In this instance, it's a mere 28 resistors and 12 caps, apparently.... :rolleyes:

Of which fewer that a half dozen are "new" (the rest just changed values), and the total cost is perhaps 25 bucks.

[dantheman]

Re: Flat Response

Deward, you don't find it harsh at all?

Dan

[mgrabow]

Re: Flat Response

wow way too much to read, but after about 20 minutes of skimming through this, did anyone mention that some builders prefer a flat response so that the speaker is not coloring the music. The theory being every component has a flat response and you only hear the music. (well if your amp is linear says nelson pass)

[Deward Hastings]

Re: Flat Response

did anyone mention that some builders prefer a flat response so that the speaker is not coloring the music.

Well, you've gone straight to the heart of the matter. Some builders profess to prefer speakers with a flat-on-axis response despite the fact that often such speakers clearly do "color the music".

The rest of this thread is about "why" they sound bad, and what to do about it . . .

[Deward Hastings]

Re: Flat Response

Deward, you don't find it harsh at all?

I don't find *any* speaker perfect . . . (including ORION, and the B2030A). I do find that the response change that restores (perceptually) a more accurate tonal balance also substantially reduces the perception of "harshness" (which I find mostly already present in recordings, not significantly added, at least by those speakers).

[DDF]

Re: Flat Response

Toole also studied different rooms, and found that the ranking remained the same; that's in part the basis for the adaption component of this. Harman and others use different rooms, as well.

See section 6, Adaption, here:

http://www.harman.com/EN-US/OurCompany/Technologyleadership/Documents/Scientific%20Publications/13686.pdf

Citation:

S. E. Olive, P. L. Schuck, S. L. Sally, and M. E.
Bonneville, “The Variability of Loudspeaker Sound Quality
among Four Domestic-Sized Rooms,” presented at the
99th Convention of the Audio Engineering Society, J. Audio
Eng. Soc. (Abstracts), vol. 43, pp. 1088, 1089 (1995
Dec.), preprint 4092.

This is a great paper, thanks for pointing it out.

Speaker ranking remained the same when comparing multiple speakers across numerous rooms, but that is a relative ranking between speakers.

It doesn't imply that there is no benefit to adjusting the on axis response to adjust to different rooms, to obtain better perceived tonal neutrality.

Toole himself makes the example in this paper that if the first reflection is highly absorbed in the high ferquencies but not below, the sound will become dull.

One finding mentioned here that is most interesting is that a strong first reflection aids vocal intelligibility. Its sort of counter to the typical wisdom that the first reflection should be scattered or absorbed. It does mirror my experience in designing conference rooms, that a hard ceiling should be used to allow a strong first reflection and increase talker intelligibility. The rationale used there is that it increases the sound pressure level of teh early arrival sound. So, if the first reflection is strong but the room total absorption remains the same, this indicates that the first reflection adds with the direct sound to increase its effective level when compared to the background din of the total room sound.

The whole part about adaptation is no surprise. What's old is new again. Its been called the "cocktail party effect" and well known since the 1950s, where we use our binaural processes to listen in even negative SNR scenarios:

http://en.wikipedia.org/wiki/Cocktail_party_effect

[DDF]

Re: Flat Response

Do you remember where those optimal "sweet spot" RT60 and RT30 times were? That's a large part of the reason why I urge people to not ignore the room--EQ can't fix temporal issues though I'd bet it can minimize its objectionable characteristics. FR seems dominant in the bass to my ears during playback anyway so I'm not too worried about the rate of decay there. I'd just like to make it sound flat.

Broader than this:
....
There are many ways to improve spaciousness and perceived sound power with any of those designs. As long as the off axis is smooth, whatever bothers you can be fixed.

I did know the targets at one time but its been 15 yrs since I designed a conference or listening room. I'm sure I could dig it up though (PM me if you want me to try and hunt it out). ~ 400ms RT60 in the mids rings a bell.

The paper that zilch pointed out showed that a strong first reflection aids intelligibility. Might be a good starting point if added detail is desired. Sort of the anti-Kantor method.

Again what's old is new, but there's no surprise that the ear's integration time is frequency dependent (Geddes way oversimplifies with that quote). DRA's MLSSA makes use of this fact through its "Adaptive Window" whereby the FFT window used is variable in each measure, and is longer the lower you go in frequency. The intent is to better correlate the measurement with the perceived tonal balance in room. I've never seen it benchmarked against listening tests, but in theory it should have some merit.

[DDF]

Re: Flat Response

Yes, that's how a dead recording sounds, but not a dead living room (reproduction), correct?

Well, it shows both. When the room is left to provide all the reverb (with none in the recording) it sounds just bizarre!