|
Post by duke on Jun 5, 2017 0:11:17 GMT -6
I'm sorry but .001% THD ... one hundredth of a percent ..... that's something the ear will never hear 🤠I don't necessarily disagree with you, but THD in and of itself does not do a very good job of predicting subjective preference. The ear is quite tolerant of very large amounts of low-order harmonic distortion, and quite intolerant of even very small amounts of high order harmonic distortion, and the THD metric does not distinguish between the two. I remember one researcher telling me that in one of his studies something like 30% second harmonic distortion was found to be statistically undetectable on music (with untrained listeners), and barely detectable on test tones. On the other hand very small amounts of high-order harmonic distortion are detectable and objectionable (the word "grainy" comes to mind), but I can't remember the specific higher orders and corresponding amounts. I do remember that the ear has a non-linear perception of this and other types of distortion, meaning that it seems as if the distortion increases as the volume level increases, but such is not really the case. Instead, what's happening is, nasty-sounding distortions become increasingly detectable and objectionable the higher in volume level they are above the ear's detection threshold.
Several researchers have proposed alternative distortion measurement paradigms that are vastly more useful for predicting subjective preference than THD, but the metrics are more complicated to generate, and to the best of my knowledge, none of these alternative approaches has gained any traction. So unfortunately for the foreseeable future it looks like the audio industry will continue to measure distortion with a yardstick designed more for test instruments than for human hearing. Just in case anyone is interested, here are links to three papers by a couple of the leading researchers in the field of distortion perception:
www.gedlee.com/Papers/Distortion_AES_I.pdf
www.gedlee.com/Papers/Distortion_AES_II.pdf
www.gedlee.com/Papers/AES06Gedlee_ll.pdf
There are implications from all of this which tie back into John's original post and offer insights as to why hardcore audiophiles might prefer a zero-global-negative-feedback high-THD tube amp over a high-global-negative-feedback low-THD solid state amp.
|
|
|
Post by duke on Jun 3, 2017 15:28:04 GMT -6
CD treatments are interesting in that they knock the CD more off balance resulting in additional servo motor activity which generates heat that can affect the clock crystal depending on how the board is laid out. This can result in the disk playing ever so slightly faster. I hadn't thought of that. But the final step in this particular treatment is "thoroughly wipe it all off".
|
|
|
Post by duke on Jun 3, 2017 14:01:34 GMT -6
Once at a high-end home audio show, I participated in a room where one of the other participants was demo-ing a CD treatment product. Once an hour he would do a demo where he would play a minute or so from two CDs that were identical except that one had been treated and the other hadn't. Sometimes he would give his spiel before the demo, sometimes after, and usually not at all.
Briefly, the spiel was this: The manufacturing process leaves the plastic surface of the CD with small imperfections which are well within the capability of the CD player's error-correction system, but because error-correction is inherently limited to interpolating an approximation of the missing data rather than actually replicating it, the less error-correction needed, the better the sound quality. The surface treatment addresses those imperfections, thus less error correction was needed, ergo better sound.
Here is what I noticed: When he gave the spiel before the demo, he often made a sale. When he either gave the spiel after or didn't give it at all, he made fewer sales if any. So here is my take-away:
If people have a reason to believe that they might hear a difference, they are much more open to actually hearing one. If they do not have a reason to believe they might hear a difference, they are much less likely to hear one. In other words, our pre-existing beliefs play a much bigger role than we would like to admit.
Did I hear a difference? Yes. I think his first sale of the show was to me. But I had heard the spiel, so I was open to the possibility that this was science and not snake oil (my default setting was of course "snake oil", and besides one does not want to appear gullible in front of one's peers - it's socially safer to "reserve judgement".). So did hearing the spiel taint my audition, or un-taint it? Could be either one, or even both.
I think that, as Martin has suggested, the real proof is in the long-term pudding, because expectation bias can all too easily be the hidden 9/10ths of the iceberg in a quick A/B test if the difference isn't dramatic. So here is a possible alternative approach: Assuming the new thing doesn't obviously suck, try living with the new thing in the system for a while, then take it out and put the old back in, and see whether you miss it and still miss it after some time has passed, or are relieved that it's gone, or simply don't care one way or the other.
|
|
|
Post by duke on Jun 3, 2017 13:16:06 GMT -6
When it comes to those things, it's hard to get me to shut up. Never found a set of - what I assume would be called "expanded D'Appolito" boxes -- Tried Dunlavy and Duntech - and - it's early, sorry (on my first cup) - others. These things just disappear. The D4M's (little guys) are just for the occasional "paranoid" client that wants to see what something is going to sound like on something more like a home system -- But even those are ridiculous. You can get *so much* sound out of them - they they ask where the subs are - then I invite them to look behind the desk to find there aren't any. Advantages...? They sound awesome, assuming an awesome recording. They sound crappy with crappy recordings. Disadvantages...? Really - I've got nothin'. Did a blog thing at one point - (a moment passes) fount it. www.massivemastering.com/blog/index_files/tyler_acoustics_decade_series_d1.php Thank you, that's very helpful. I read your blog, GREAT information... and I gotta tip my virtual hat to you for having the courage to go against the grain and give Ty's speakers - an unknown in the studio world - a shot. I didn't realize that Danny Ritchie had done the crossover, but he's an absolute master of the art.
|
|
|
Post by duke on Jun 3, 2017 0:55:14 GMT -6
I need new pictures... It's a Sterling Modular (D...?) Hi John,
Very cool to see not one but two pairs of Ty Lashbrook's speakers in your studio! Do you find any particular advantages and/or disadvantages to the MTM configuration of the Decades for this application?
How do clients react to them?
Thanks!
|
|
|
Post by duke on Jun 2, 2017 18:59:54 GMT -6
Good grief, the response with that "new crossover" is just a hair over plus or minus 1 dB from 100 Hz to about 16 kHz! I realize it's not the final version - doesn't matter, that is still very impressive work. Is that the actual measured response, or the response predicted by a modeling program, presumably using measurements you input? This was a simulation based on the actual ZMA data. I have not done actual sweeps yet, just listening tests and I find that the speaker is as good or better than Amphions, but still not quite right to my ears. I found Amphion to be much too bright for my liking. These speakers are less bright, and really, really bring out the best in classical music or older recordings. I haven't really had time to mess with the speakers in a while since I've been really pushing to get the studio going again after a few months off for renovations. The 2-way speakers in this thread sound good, but there is still something in the sound that I want to adjust, but I was having a hard time figuring out what it was, so that's where I stopped last time. When I worked on the metal coned 3-way speakers, I heard a similar sound and found it's coming from the tweeter. In order to feel better about these speakers, I'd like to figure out what the thing I hear is, and see if I can change it. It's obviously from the metal-domed tweeter across multiple speakers, so it's inherent to the tweeter design. I've also talked with folks about taking over the design once I finished tweaking the speaker, so I don't know if I'd necessarily release it as DIY just yet. At one point I thought about selling pre-made crossovers for 1cuft boxes, but there didn't seem to be a market for DIY speaker builders in pro-audio. Very impressive job of hitting the target with that crossover design. Obviously you have mastered the stage of "being able to hit the target"... the next higher up stage is "figuring out where the target really is," because you can nail what you thought was the target, and then it turns out, it still doesn't sound quite right.
And kudos for being honest enough with yourself to be able to listen critically to your own creation... well, I suppose that sorta comes with the territory around here.
I have found that there is virtually always enough discrepancy between my modeled response and my subsequent measured response to call for a bit of crossover tweakage. Be sure to measure carefully - bad data is worse than no data - and then adjust your sights, so to speak, before you tweak the crossover, so that your actual measured response is what hits the target. Even if you haven't yet figured out "where the target really is", at a minimum you need good measurements to tell you where your design really is.
You mentioned something not right that is coming from the tweeter. I hope you don't mind me asking a few questions - feel free to ignore them!
Does the tweeter on your two-way have a flat faceplate? How about the tweeter on your 3-way? What are the ballpark crossover frequencies between midwoof and tweet on the two-way, and between mid and tweet on the three-way, if you don't mind? Any significant dips in your impedance curve?
Thanks!
|
|
|
Post by duke on Jun 2, 2017 13:44:16 GMT -6
Hm, interesting read. My brother and me found two chassis that were meant for a dipole sub and for a second we considered building a stereo pair, but decided against. I really am excited how this will improve our monitoring. Last time, we tested a sub with the Sonus Faber speakers our mouths stood open. Letting the sub do the work improved sound so much... Main reason we stood by our mono sub idea was the positioning in the room. AFAIK the sub we build has to stand a meter away from the walls, and it is not small either... Two of them would complicate things for us... Dipole subs are very interesting. They do some things really well, such a pitch definition, and some things not so well, such as visceral deep bass, but imo there is a work-around for that. And since you egged me on with the words "interesting read", instead of saying "sure wish I could have those five minutes of my life back"...
Once again, you can just skip ahead to the last paragraph. Well, second-to-last.
Let's start by dividing the bass region into two zones: The "modal zone", where we get those familiar and unwelcome room-interaction peaks and dips; and the "pressure zone", down below the modal zone, where the wavelengths are so long relative to the room dimensions that all of the first reflections occur within 1/2 wavelength of each other, which tends to put all of the in-room bass energy in-phase. In the pressure zone, sound behaves like changes in static air pressure. The longest room dimension is actually the oblique dimension, going from one bottom corner to the opposite top corner, on a 3-D diagonal. We will come back to these concepts later.
Another concept we want to look at is "de-correlation". This concept is less well-known, but it is the primary acoustic smoothing mechanism for in-room bass in both dipole bass systems and distributed multiple monopole bass systems. We get good de-correlation when the in-room bass energy is adding in semi-random phase. De-correlation is desirable because it results in smooth bass, because no particular frequencies have enough energy to stick out like sore thumbs ("peaks"). To help you visualize what we mean by the term, good "de-correlation" would be like the foot-falls of a highschool football team running out onto the field. On the other hand, strong "correlation" would be like the foot-falls of the Wehrmacht on parade. The thing to remember is, good de-correlation = smooth in-room bass.
Now the good news is, in general, dipole bass really is smoother than monopole bass! In 2002 researcher James M. Kates published a paper in the Journal of the Audio Engineering Society showing that dipoles have smoother in-room bass than monopoles. This wasn't news to the many audiophiles who love the excellent pitch definition of well set-up dipole speakers, but it was long-overdue officially-sanctioned peer-reviewed validation.
Now most people who talk about the "speed" (in-room smoothness) of dipole bass attribute it (mistakenly in my opinion) to the figure-8 radiation pattern exciting fewer room modes. But if only few-and-far-between room modes are excited, that's an indication of strong correlation (Wehrmacht)! And we'd expect those few-and-far-between peaks and dips (especially the peaks) to stick out like sore thumbs. Kates' data shows smoother in-room bass for dipoles, which implies good de-correlation (the opposite of only exciting a few modes), so something else must be going on. Here's what I think is happening:
The primary smoothing mechanism for dipole bass is, the phase relationship between the frontwave and the backwave energy, plus the radiation pattern shape (figure-8). Because of the radiation pattern, the normal-polarity frontwave goes off in one general direction, and the reverse-polarity backwave goes off in the opposite direction. After a few room boundary bounces the twain shall meet again, but their phase relationships are now generally semi-random - which means good de-correlation - highschool football team!
[Not that I'm necessarily the world's biggest fan of highschool football, whose players were often the dispensers of wedgies and noogies to us little nerdlings, but at least football gave them someone else to hit and hit hard.]
In general, it takes two intelligently-distributed monopole sources to get the same approximate in-room smoothness as a single dipole source. The energy from the dipole source sums in-room in semi-random phase because the frontwave and backwave start out in opposite polarity and are launched in opposite directions, and the energy from two intelligently-spaced monopole sources also combines in semi-random phase because of their physical distance from one another and dissimilar distances from most of the room's boundaries. So the good news is, a single dipole sub will probably be as smooth in-room as two well set-up monopole subs.
Unfortunately there is also some bad news. Aside from the significant amounts of EQ that dipole subs will need in order to reach down deep.
Okay we just saw that the same mechanism - semi-random phase summation - is at play in smoothing the in-room response of both dipole bass systems and multi-sub monopole bass systems in the modal region. But these two approaches totally diverge as far as what happens down in the pressure zone. If the energy in the pressure zone is all in-phase (like with four monopole subs), then it sums in-phase. This can result in boom, because in-phase summation is inherently 3 dB louder than the semi-random-phase summation we were getting up in the modal zone. And because of the ear's heightened sensitivity to relative loudness at low frequencies (which is shown in the bunching up of equal-loudness curves south of 100 Hz), a 3 dB emphasis below 40 Hz would sound bigger than a 6 dB emphasis up in the midrange region.
But with a dipole, down in the pressure zone the outcome is totally different. In the region the wavelengths are more than twice as long as the longest room dimension, where the net effect is like the summing of static air pressures, we have half of the pressure in one polarity and half in the opposite. So the net sum would be complete cancellation (making it a "no-pressure zone"?). In practice the cancellation isn't complete, but it is severe, and as a result dipole bass tends to have less "impact" or "solidity" than monopole bass unless a) the room is very large (pushing the "no-pressure zone" down very low in frequency) and b) the dipole bass system actually does go very low (which requires a lot of air-moving capability plus a very large baffle and/or a lot of EQ).
One "best of both worlds" approach would be, a dipole sub (or subs) active through the modal region, and then a monopole sub (or subs) blending in down low for the transition into the pressure zone. This would combine the inherently smoother modal-region behavior of a dipole sub with the superior extension and solidity of a monopole sub.
Smallbutfine, having read your signature, I understand finances are an issue (and that it's generational), and that you struggle with feelings of rejection. Since you already have the frames, completing a dipole sub shouldn't be too expensive, and I can make suggestions for a DIY companion pressure-zone monopole sub if that would help. You would then have an incredibly cool subwoofer setup, and if that doesn't make them love you, then at least they will envy you. Which is almost as satisfying.
|
|
|
Post by duke on Jun 2, 2017 12:10:15 GMT -6
Sorry, I don't buy this at all. Blind tests have inherent biases, period. Other far more knowledgeable engineer members here have said as much in other posts, so we'd be rehashing the same argument. So I'm not gonna argue my point this time, it's a rabbit hole. You want to use coat hangers or radio shack spools, be my guest ;-) People who want to hear there's no difference are only too glad to believe in blind tests to reaffirm their beliefs. I've often argued about cables or power cords with people who never even tried switching out a cable once. Ask Jim Williams or Bob Olson. * by the way, I've never found Monster Cables to sound very good. There are other brands at the same price points that are far superior. ** oh, and Duke, I've never found that audiophile wife you mentioned eiither, and I've known more than a few audiophiles in my time. Welcome to the forum! Your example earlier in the thread of jitter is an excellent one, showing that focused human perception can be far more discerning than the consensus theories and best measurements of the day.
The late Richard C. Heyser was one of the sharpest minds in the field of audio measurements. He's the one who invented time-gated ("quasi-anechoic") loudspeaker measurements, and brought Fast Fourier Transform analysis into the game. Here's how he concluded one of his fairly technical articles: "You out there, Golden Ears, the person who couldn't care less about present technical measurements but thinks of sound in gestalt terms as a holistic experience. You're right, you know."
|
|
|
Post by duke on Jun 2, 2017 1:06:00 GMT -6
Johnkenn wrote: "I keep hearing about audiophile wives who share their husband’s passion – everyone I speak to seems to know another audiophile whose wife enjoys nothing more than discussing the finer points of software room correction – but I never actually meet one. I start to wonder if it’s a kind of urban myth."
Unicorns. That's what they are. I didn't believe in them either, but then I met one and tricked her into marrying me. When I met her, I'm not sure she even knew what an "audiophile" was, which gave me an unfair advantage. But I could see that her appreciation of sound quality was way out at the tail end of the bell curve. She is the first female I met who instantly fell in love with the seven foot tall electrostats I owned at the time. I'm pretty sure they are why she married me. Like I said, unfair advantage.
One day I wondered if there were audible differences between balanced interconnects. I happened to have six different balanced interconnects on hand, so I set up a simple blind test. It was like a double-elimination basketball tournament. I switched out the cables and told her when I'd made a change, and put on whatever piece of music I was using (NOT Diana Krall). She would stop reading her book and listen intently and scribble down some notes, then she'd pick the winner of each "duel" based on her notes. And going back over her notes, I could see that she consistently described each cable whenever it showed up in the different duels. Her favorite was by a company called Creative Cable Concepts, a company that I had no connections with - they were just loaner cables from a friend.
I trust her ears more than my own. Her ears have more taste buds than mine do.
|
|
|
Post by duke on Jun 1, 2017 23:10:15 GMT -6
So here's the expected response of the new crossover. I still found the top end to be a slight bit "hot", so I added a 2R series resistor to the tweeter to knock it down a couple dB. Here's the old one for reference: Good grief, the response with that "new crossover" is just a hair over plus or minus 1 dB from 100 Hz to about 16 kHz! I realize it's not the final version - doesn't matter, that is still very impressive work. Is that the actual measured response, or the response predicted by a modeling program, presumably using measurements you input?
|
|
|
Post by duke on Jun 1, 2017 22:07:53 GMT -6
Duke I will say this I hate DSP for solving room problems except when it comes to the very Low Freq, DSP makes good bass so much easier!! Eric, I don't have much first-hand experience with DSP but what you say makes sense to me. In general I prefer acoustic solutions to acoustic problems, with EQ generally as a last resort... but acoustic solutions can be hard to implement at very long wavelengths.
|
|
|
Post by duke on Jun 1, 2017 21:06:45 GMT -6
In the higher end of the Home Theater world, two subs are a must. How would you actually get a signal to two subs anyway? I don't need one, my Avantone Abbeys go down to 35 Hz. Ericn probably knows more ways to do it than I do.
My hands-on experience has only been on the playback (home audio) side. Most of the time the preamplifier has a second set of fullrange outputs or a "sub out" that can be used; when it doesn't, I have to summon the mystical powers of math and design and build a "voltage divider network", which derives a line-level signal from the power amplifier's output.
From there, I use a line-level input into a single external amplifier to drive multiple passive subs, if it's a mono subwoofer system. For a stereo subwoofer system, I use two external amplifiers, one for the sub(s) on each side of the room. The external amplifier that I use has the needed controls built in.
|
|
|
Post by duke on Jun 1, 2017 20:52:44 GMT -6
I don't know what you guys are referring to by "wave guide". What ChaseUTB said.
As used in this context, a "waveguide" is a type of horn - specifically, a very gentle horn whose primary purpose is radiation pattern control, rather than acoustic amplification.
Basically the waveguide forces all of the tweeter's output into the same angle, in this case maybe 110 degrees, I'm not sure. Anyway left to its own devices the tweeter would want to be omnidirectional at the bottom end of its range, and then of course it would beam at the top end. The waveguide can't do anything about the beaming on the top end, but it can force the bottom end into a narrower pattern.
This introduces a potential problem: The bottom end of the tweeter's output is now louder because all of its energy is being concentrated into a smaller angle! So this has to be equalized in the crossover. Fortunately it's not too difficult, and along the way power handling is improved a bit.
Imo the most benefit from a waveguide comes when the crossover between woofer and waveguide occurs approximately where the woofer's radiation pattern has narrowed to match that of the waveguide. This is called "pattern matching". Without knowing the radiation pattern width of the Amphion waveguide I can't really calculate whether they are employing pattern matching, but they are imo doing so many other things right that I assume they are.
|
|
|
Post by duke on Jun 1, 2017 20:37:33 GMT -6
That is my thinking about good waveguides in general, but I didn't know John's space was small and reflective, which would indeed make their benefits more keenly felt.
Good pattern control not only reduces detrimental early reflections, but also make the spectral balance of the reflected energy similar to that of the first-arrival sound. This has multiple benefits, not the least of which is, reduced listening fatigue. (I can explain why if anyone is interested.)
Hey man you seem to know you monitors well... I have on stage stands and they have rubber ish pad over the metal however it resonates which make me believe I need to decouple.. would sorbothan hemisphere work well? Looking for a good solution that doesn't involve hundreds of dollars ... thanks if it's too off topic feel free to private mesg me Nah, I just talk a good game. I'm student of speaker design, and see some things in the Amphions that make a lot of sense to me.
The sorbothane hemispheres sound like a good idea to me. If your metal stands are kinda thin, you try covering the bottom of the plate that holds the speakers with Dynamat. This will add mass and damping to the metal.
And if someone with actual experience comes along and tells you something different, take their advice over mine!
|
|
|
Post by duke on Jun 1, 2017 20:17:01 GMT -6
The only thing I disagree with is that subs don't need to match, you are forgetting that they act above the crossover point , and here the acustic summing of like drivers is your friend, comb filtering of unmatched your enemy, you are already dealing with the phase interaction of the diss-similar phase / frequency response of the subs and mains, add in another player and your gains at the lowend are at the price of mid bass. Well if that's all we disagree on, we're off to a pretty good start!
If the subs are in stereo, then I totally agree with you, the two channels need to match for exactly the reasons you describe.
If we have just one subwoofer channel, I think it depends on how high and how steep the crossover is, and maybe on positioning as well, as to whether we can find settings that give a good blend between multiple dissimilar subs and the mains. But I have no objection to the rule of thumb being "the subs should all be the same", and then sometimes an exception can still work well. The person who taught me about distributed multisub systems, Earl Geddes, used one big bandpass ubersub that went very deep, and then multiple smaller subs. Now that I think about it, his big bandpass sub was probably way down in level by the time we got to the crossover region anyway.
|
|
|
Post by duke on Jun 1, 2017 19:59:59 GMT -6
In John's small reflective space the wave guide of the Amphions gives them a huge advantage! That is my thinking about good waveguides in general, but I didn't know John's space was small and reflective, which would indeed make their benefits more keenly felt.
Good pattern control not only reduces detrimental early reflections, but also make the spectral balance of the reflected energy similar to that of the first-arrival sound. This has multiple benefits, not the least of which is, reduced listening fatigue. (I can explain why if anyone is interested.)
|
|
|
Post by duke on Jun 1, 2017 19:00:59 GMT -6
I was going to post this in jeromemason's “PAIR of subwoofers” thread, but by the time I was done my post had pretty much turned into a full-blown hijack, as you can see from its length. If you're short on time, just skip down to the last paragraph.
Having totally failed to master the technical intricacies of digitally quoting on this site, I hope you will forgive me for reverting to analog:
jeromemason wrote: “Just wondering the thoughts of some that either have or had dual subs going...
“I read an article where two guys were pitching using 4 or more subs in a room...”
Briefly (or in retrospect, perhaps not!), here is a look at the situation with a single sub in a home studio room, and then we'll look at how an additional sub (or subs) can make an improvement. Imagine you have one subwoofer in your room. Its interaction with the room boundaries results in a rather drastic peak-and-dip pattern. You can move the sub and change the peak-and-dip pattern, but you cannot eliminate it by positioning alone. You can mitigate it by a few dB with large bass traps. You can equalize it to be fairly smooth in one location, but in the process you will have made it worse elsewhere. And it is possible that the response still will not be sufficiently smooth at both ears. Quoting jazznoise:
“In small rooms, the issue is that room modes can be so big that they can cause different tones to 'sit' left or right. Interaural time delay as a listening mechanic doesn't work well for bass, it's intensity and phase at low frequencies that allow your brain to attribute directionality. If your right sub puts out a 50 Hz tone which nulls to your right but rings to your left, the localization is gonna be messed up.”
In a small room with a single sub, there's a good chance you'll have at least one frequency region where the situation jazznoise describes applies, and EQing that single sub won't fix it. It's an acoustic problem, and arguably is best addressed with an acoustic solution.
At low frequencies, sub + room = an inseparable system, from the standpoint of perception. The ear/brain system literally cannot hear the sub apart from the room. We cannot detect the presence of bass energy from less than one wavelength, and we cannot determine the pitch until we have heard several cycles (wavelengths). Mentally compare your sub's distance to your various room boundaries with how long low-frequency wavelengths are, and you'll see that by the time you actually hear the low frequency energy, the room's effects are already in full swing.
Back to the room-induced peak-and-dip pattern for a moment. Smooth bass is “fast” bass, because the peaks literally take longer to decay. To put it another way, speakers + room = a minimum-phase system at low frequencies, which means that if we smooth the frequency response, we have simultaneously smoothed the time-domain response, and vice-versa. In-room peaks spoil the pitch definition because, as the energy decays over time (and remember the peaks take longer to decay), the spectrum actually shifts in the direction of the peaks.
If these room-induced peaks and dips were very numerous and bunched up close together, it might look terrible to the eye on a graph, but perceptually it would in effect be a “continuum”, because the ear/brain system tends to average them out (and indeed this is what happens with the reflections at midrange and high frequencies – perceptually they merge into a continuum). But if the peaks and dips are further than about 1/3 octave apart, the span is too great for the ear's averaging mechanism to come into play, so they stick out like sore thumbs. One counter-intuitive implication of this is, the problem isn't that we have too many room modes at low frequencies – it's that we have too few! (Bigger rooms have greater model density at low frequencies, which is why they usually have more natural-sounding bass; the smaller the room, the fewer and further apart the peaks and dips in the bass region.)
If we look at a set of equal-loudness curves, we'll see that they bunch up south of 100 Hz or so. This is telling us that a change in SPL in the bass region makes a disproportionately large change in perceived loudness. For example, a 5 dB change in SPL at 50 Hz makes approximately as much of a difference in perceived loudness as a 10 dB change at 1 kHz!
So to recap, we have three problems that can arise from using a single sub in most rooms:
1. Poor pitch definition due to room-induced peaks;
2. The peaks (and to a lesser extent the dips) perceptually stick out like sore thumbs because they are not only large but few and far apart; and
3. The ear/brain system has a heightened sensitivity to SPL changes south of 100 Hz, so the subjective detriment of these peaks and dips is greater than we would infer simply from eyeballing our in-room frequency response curves.
At the risk of oversimplifying, what usually happens in practice is, we turn down the level of the subwoofer until the peaks are tolerable. At that point it's usually a net improvement, but arguably still far from ideal.
So let's bring in a second subwoofer, and put it in a very different location from the first one. This second subwoofer generates a room interaction peak-and-dip pattern every bit as bad as the first one, but the peaks and dips are at different frequencies! And this difference happens throughout the room. So the net result is a significant smoothing of the frequency response. The only way that adding a second subwoofer could result in response just as bad as the first sub would be, if their in-room response curves were identical. And unless they are in the same place, that will not happen.
In general, the in-room smoothness increases proportionally with the number of subs we have distributed around the room. So two subs will be roughly twice as smooth as one, and four subs will be roughly twice as smooth as two. And this improved smoothness holds up throughout the room, which means that if you still need to EQ to fix a remaining problem, chances are you are fixing a global problem instead of fixing a local one but making things worse elsewhere. (I realize the mix position is by far what matters most, but when the time comes to “impress the customer”, really good bass throughout the room is not going to hurt.)
Now when we measure the in-room response at low frequencies, the peaks and dips will be smaller, more numerous, and bunched up closer together (more like what we'd see in a significantly larger room). The subjective improvement will probably be greater than we would infer from eyeballing our before-and-after curves, for two reasons: The ear/brain system's one-third-octave averaging mechanism can probably come into play; and just as the ear/brain system is disproportionately sensitive to SPL variations south of 100 Hz, it is also disproportionately appreciative of the improvements we will have made.
Let me point out one more implication of all this which is somewhat counter-intuitive: From a perceptual standpoint, it is the trailing edge of the notes (the in-room decay times) that matters most. A little bit of smearing of the initial arrival times from multiple subs at low frequencies doesn't matter because the ear/brain system doesn't react fast enough to detect it. But the ear/brain system is very good at hearing amplitude changes at low frequencies (remember the equal-loudness curves bunching up), so we want to get the amplitude as smooth as we reasonably can. And in doing so, we will also be getting the decay times as even as we reasonably can.
So even though it doesn't make sense at first glance, several small subs intelligently distributed will generally give you better perceived transient response than a single uber-sub.
The good news for those of you who already have a sub is, multiple subs do not need to all be identical. If your current sub goes plenty low enough, you probably add get one or more less expensive subs and still see significant benefits from the smoothing you'll get in the modal region. For those of you who don't have subs but are thinking about getting one or more, if you opt for “more”, you can get away with smaller subs than you might have otherwise been thinking of. And you don't have to buy them all at once.
One final argument for maybe putting something like this on your down-the-road wish-list: Humans are wired to have a positive emotional response to low frequencies moreso than to mid and high frequencies. So if you want to enjoy your incredibly cool jobs even more...
For those of you who took my advice and just skipped to the last paragraph, in general the more subs you have (and set up intelligently), the better the in-room bass will be, both from a frequency response standpoint and from a time-domain standpoint. In effect, you will be using the room's acoustics to work with you instead of against you, both acoustically and psychoacoustically. The net result will be, perceptually you will have pretty much gotten the room out of the way, so you'll hear what's on the recording down low more clearly and therefore make better mix decisions. And when the time comes that you need your system to make a good impression on a customer, better is better.
|
|
|
Post by duke on Jun 1, 2017 7:00:17 GMT -6
|
|
|
Post by duke on Jun 1, 2017 2:07:28 GMT -6
A passive radiator is in effect a type of port, and so the final rolloff rate will be 24 dB per octave, just like for any ported box. However depending on the design specifics, the initial rolloff rate may be much more gradual, and 12 dB per octave initial rolloff is possible. Gradual initial rolloffs usually start somewhat above the tuning frequency and extend somewhat below it, before accelerating to 24 dB per octave. My guess is that Amphion's engineers specified woofer parameters that would give their speakers an unusually gradual initial rolloff. In doing so they probably would have traded off some efficiency and/or lowest possible -3 dB point, but imo those would be good tradeoffs to make for this application because they would tend to preserve pitch definition (which can be compromised in ported designs).
Amphion may or may not disclose the resonant frequencies and other parameters of their woofers, but their published impedance curves reveal the tuning frequencies: About 50 Hz for the One15, and about 42 Hz for the One18. The lowest point in the dip between the two bass impedance peaks is the tuning frequency.
EDIT: For some reason my attempts to quote ChaseUTB have been in vain. Obviously I haven't figured out how this thing works yet.
|
|