What's Left In Speaker Design To Reduce Distortion/Increase Detail Retrieval?

[gnarly]

I think that for sure is the minimum requirements. Of course some will say you need to do it blind (I don’t necessarily agree with that if you are testing speakers in your own room, as opposed to doing objective research on sample sizes). Some will even argue needs to be “double blind” but I think that’s even more useless.

However, what about one on the left, and one on the right? Back and forth A and B and blind. All of Olive’s testing is single speaker mono, would seem to me (and I have done this a fair amount), that would get you even closer on a preference between two top choices. Determine preference. Then, switch speakers, left to right, right to left, do it over again just to make sure there isn’t a room/corner bias issue going on (which can and does happen, more with comparing a specific driver with another driver in same model of speaker).

Doesn’t that give you the truest preference and comparison? From there you take the top choice pair and listen in stereo to see if anything about imaging is a deal breaker.

I’m asking as this has seemed to work for me pretty well.

i tend to agree with all that.....

As an avid DIY speaker builder, I'm continually comparing new designs and different processing strategies.

I've learned that the only fair way indoors to A/B, is indeed one speaker at a time.
But the comparison is really only valid for A and B to be in the exact same room position....which means not only sighted, but also the problem of short term memory loss switching out A and B.
So I tend to do as you suggest; put A at left, B at right, and listen for a while back and forth. Then swap A & B's sides and repeat listening back and forth.
Only thing I've found that works better is to set up A & B outdoors, where the speakers can be next to each other. That test rules (and always sounds better clarity/timbre wise than indoors, if the speakers have any guts).

I have to admit i don't put much stock in either blind or double blind.
My opinion is, if it's close enough such a test is even needed...they're the friggin SAME, and who cares which one is "best"...

The speakers I build are 4, 5, and 6-way actives, all with dsp processing on each driver/amp section.
Things can get real complicated even setting up one speaker alone, much less devising ways to make quick and valid A/B comparison.
It's become pretty easy to hear when something is wrong on a new test. And then I quickly run a transfer function on the speaker to hunt down the issue.

After I know everything is technically correct (and all speakers under test would win the Good-Spinorama Seal of Approval ,lol) .....
what's more interesting to me, is long term listening comparisons....in the primary room, down the hall in my office, outdoors, everywhere.

Whenever I think I've hit a new level of detail/clarity in a DIY prototype, I make a second for stereo, and then go though A/Bing the pairs in stereo.
Long process for sure...but hey, hobbies can be fun huh?

 

[MattHooper]

Anyone who is a member here would acknowledge that if you really want to increase confidence levels, you want to control for variables. The ideal is the double blind testing often referred to for research concerning speaker preferences. But part of being reasonable is being practical, so given almost none of us are in a position to do those type of tests with speakers, we have to make do. Most often we can't hear whatever speakers we want in our own room. Life is life.

A scientific level of scrutiny is the most reliable, but it's often not practical nor necessary for making reasonable (and satisfying) purchases for speakers.

 

[Kvalsvoll]

Loudness is an essential property for ability to hear detail, if detail means sound very low in level compared to average and peak levels. When we turn down the volume, those low level parts fall below threshold of hearing, and thus can course not be heard.

So how loud is normal-loud? Not louder than can be listened to for some time, and louder than most typical hifi speakers are capable of:

How loud is our music – Audio

I have some sound samples that illustrates this, with a normal level signal on top of very low level signals. Then you can hear that a certain volume is required to be able to hear the low level signal. Just need to find them..

 

[Travis]

Anyone who is a member here would acknowledge that if you really want to increase confidence levels, you want to control for variables. The ideal is the double blind testing often referred to for research concerning speaker preferences. But part of being reasonable is being practical, so given almost none of us are in a position to do those type of tests with speakers, we have to make do. Most often we can't hear whatever speakers we want in our own room. Life is life.

A scientific level of scrutiny is the most reliable, but it's often not practical nor necessary for making reasonable (and satisfying) purchases for speakers.

Double blind is great for drugs. Even food and beverages. The gold standard for drugs is double blind testing. The person in the lab coat doesn’t know if they are giving the test subject a placebo or the real drug. Test subject therefore cannot pick up cues from guy in lab coat as to what they are getting, or try to influence the guy in lab coat to give them a potentially life changing drug instead of placebo.

They don’t even do double blind testing for drugs they are working on for people with terminal diseases or advanced stage terminal cancers. It’s unethical. Everyone who qualifies for the trial gets the drug, they factor out placebo effect statistically.

Audio comparisons, can certainly benefit from blind testing, depending on what you are testing, and what you are testing for. However, in a blind audio test, whether it is 4 speakers being compared on a turntable behind a veil, or DACs, or amps, etc. the concerns of single blind testing a don’t apply. I just haven’t seen any studies involving audio to show that if the person administering the test knows what A and B is somehow gets picked up by the test subject and influences their answers/responses. I would like to see the studies invoking audio/sound that showed there is a statistically significant difference between doing the comparisons blind vs. double blind.

In the case of trying to narrow down two top contender products, typically speakers, I guess someone could go through the charade of putting up acoustically transparent (no such thing, they all change sound to some degree) panels and place and hook-up speakers/equipment and then find another person to administer the test to have it double blind. I just don’t think it’s going to matter at all that your audio buddy sets up the speakers for you is also the one who administers the test.

 

[Travis]

I have some sound samples that illustrates this, with a normal level signal on top of very low level signals. Then you can hear that a certain volume is required to be able to hear the low level signal. Just need to find them.

I am guessing that most people would not be able to set up an A/B comparison of speaker 1 vs. speaker 2 in the same room, or even amp 1 vs. amp 2 (which is much, much easier) within 1 or 2 db of each other let alone the .1 db that is really required.

How much time is required to “calibrate” the speakers to a master gain control mentioned in the article. How long would it take to loudness match two different speakers in the same room so that is you switched back and forth between A and B they would be matched?

 

[Blumlein 88]

Double blind is great for drugs. Even food and beverages. The gold standard for drugs is double blind testing. The person in the lab coat doesn’t know if they are giving the test subject a placebo or the real drug. Test subject therefore cannot pick up cues from guy in lab coat as to what they are getting, or try to influence the guy in lab coat to give them a potentially life changing drug instead of placebo.

They don’t even do double blind testing for drugs they are working on for people with terminal diseases or advanced stage terminal cancers. It’s unethical. Everyone who qualifies for the trial gets the drug, they factor out placebo effect statistically.

Audio comparisons, can certainly benefit from blind testing, depending on what you are testing, and what you are testing for. However, in a blind audio test, whether it is 4 speakers being compared on a turntable behind a veil, or DACs, or amps, etc. the concerns of single blind testing a don’t apply. I just haven’t seen any studies involving audio to show that if the person administering the test knows what A and B is somehow gets picked up by the test subject and influences their answers/responses. I would like to see the studies invoking audio/sound that showed there is a statistically significant difference between doing the comparisons blind vs. double blind.

In the case of trying to narrow down two top contender products, typically speakers, I guess someone could go through the charade of putting up acoustically transparent (no such thing, they all change sound to some degree) panels and place and hook-up speakers/equipment and then find another person to administer the test to have it double blind. I just don’t think it’s going to matter at all that your audio buddy sets up the speakers for you is also the one who administers the test.

So, what audio/sound is unlike every other endeavor? You need blinding for best results except for audio. Huh-huh.

 

[Blumlein 88]

I am guessing that most people would not be able to set up an A/B comparison of speaker 1 vs. speaker 2 in the same room, or even amp 1 vs. amp 2 (which is much, much easier) within 1 or 2 db of each other let alone the .1 db that is really required.

How much time is required to “calibrate” the speakers to a master gain control mentioned in the article. How long would it take to loudness match two different speakers in the same room so that is you switched back and forth between A and B they would be matched?

Amps are dead easy. Measure voltage with a tone.

Speakers would be something different, but what can be done isn't especially time consuming.

 

[Kvalsvoll]

I am guessing that most people would not be able to set up an A/B comparison of speaker 1 vs. speaker 2 in the same room, or even amp 1 vs. amp 2 (which is much, much easier) within 1 or 2 db of each other let alone the .1 db that is really required.

How much time is required to “calibrate” the speakers to a master gain control mentioned in the article. How long would it take to loudness match two different speakers in the same room so that is you switched back and forth between A and B they would be matched?

Fairly quick, most time spent will be setting up the measurement equipment. For a calibration with sufficient accuracy for listening and end up with systems that will have reasonably comparable loudness for same master volume setting.

To compare 2 speakers AB, that is tricky. Frequency response for different speakers rarely match within 1-2dB, and in-room differences are usually even larger, due to larger differences in off-axis radiation compared to on-axis. So you end up using some sort of smoothing or average across a reduced bandwidth somewhere in the midrange. Then the differences in off-axis radiation will affect perceived loudness because decay will be different, perhaps not much, if the room is fairly dead. So this will be approximate, and likely not good enough to prevent loudness difference alone invalidate any attempt for blind testing, there will be audible difference.

The quickest way to do this is to use one speaker as reference, measure its frequency response. Then measure the next speaker, and eyeball the necessary trim level correction needed, by looking at the frequency response graphs. This should be good enough for AB listening.

Amplifiers are easy, because the signal can be measured directly on the amplifier output. But also not so interesting, because both measurements and listening tests have confirmed all reasonably good amplfiers are audibly transparent when used within operational limits.

 

[MarkS]

Here is a protocol that does not require loudness matching, but still does not introduce any bias towards the component with more gain:

After each switch, the volume starts at zero, and the listener turns it up to the listener's preference, using a control that does not give any visual or tactile feedback as to the volume level. In this way, the final sound level for each test will be the same, to within the listener's ability to set it.

 

[MattHooper]

So, what audio/sound is unlike every other endeavor? You need blinding for best results except for audio. Huh-huh.

I didn't take that to be Travis' point.

If I interpreted him correctly, he was making the point that, though double blind testing is often touted as the gold standard, it doesn't rule out the relevance
of single blind testing in regard to audio equipment. (Though I'd say he stated that a little too strongly for my taste).

Insofar as that is his point, I'd agree, and I'd think many here would agree including Amir (who has a video explaining how to do single blind testing for audio equipment).

 

[Blumlein 88]

Here is a protocol that does not require loudness matching, but still does not introduce any bias towards the component with more gain:

After each switch, the volume starts at zero, and the listener turns it up to the listener's preference, using a control that does not give any visual or tactile feedback as to the volume level. In this way, the final sound level for each test will be the same, to within the listener's ability to set it.

This is guaranteed to muddy results. Done this way there is a window of a db or so within the listener's ability to set volume. And much less than this will bias the results to whichever ends up being the slightly louder one.

 

[MarkS]

If that's the case, then the differences are too small to matter, IMO.

The advantage is that this protocol requires no measurements. It can be used by any audiophile, the vast majority of which do not own a voltmeter.

I agree that it will be slightly less sensitive to small actual differences than carefully measured level-matching.

So it really depends on the goal. A technosavvy audiophile with the proper measuring equipment, and the technical capability to do it correctly, should definitely measure.

But someone who is not capable of doing a proper measurement, but who has an open mind and wants to know if those differences he thinks he hears (when adjusting the volume by ear anyway) are really there or not, can use this much simpler protocol.

 

[MattHooper]

This is not so much directly related to the technical discussions in the thread, but hey...26 pages later we can riff a bit. Since I mentioned earlier in the thread listening to some speakers with my brother-in-law (BIL), who was most impressed with the Kii speakers, maybe I'll just mention quickly this recent trip to another Hi-Fi store (our largest, oldest and most well known in the city). It was pretty disappointing overall.

We went there to hear, among others, the KEF LS60 speakers - one I'd been keen to hear given the hype but also because of my BIL's interest in active loudspeakers.

Unfortunately it was just a terrible set up. They were about 8 or 9 feet apart I'd estimate, but sitting in an inset along a wall, so they were placed a couple feet up on a "shelf" and almost right up against the wall. We heard some Pink Floyd, Diana Krall etc. Mostly useless because the sound was so bad, the bass was bloated and boxy, even the midrange was slightly obscured. I tried my best to move close, further, lower, higher to somewhat combat the issues, and the best I can say in trying to hear through the obvious problems was that they had promise. Big, spacious soundstage and imaging, decent bass depth, and I could discern that recognizable character in the upper frequencies that I hear from the KEF LS50 models: For me it's a very open, smooth, sparkly though somewhat thin high frequencies that have (to my ears) a convincing sense of timbre (they have timbral "color" vs sounding black and white to my brain). Though also that slightly smoothed over sensation, almost the tiniest metallic sheen, that makes string parts seem at once clean and clear, but a tad over-smooth in terms of bow texture etc.

But if they'd used the KEF active attributes to try to combat a challenging set up, they seem to have failed that challenge. Some day we hope to hear the LS60 in a better setting :-(

Given my BIL's current taste skewing to the Kii Audio speakers so far, I was curious what he'd think of the B&W 803 D4 speakers shown in a dedicated room, which I've heard before at this shop. The room is average size, somewhat treated (though didn't seem to be with much attention to detail). Speakers were reasonably far out from the wall spread apart, so should give a decent showing.

My impressions were once again: Holy cow these things give the impression of "transparency" as many audiophiles use the term: that is, really detailed, open sounding, virtually zero sense of "boxiness" to the sound that I could detect at any frequency, "quick" and snappy sounding (as in no sense of bloating or sluggishness, whether it was playing low bass or not)...just really alive and zippy. And wow did it sound different playing the very same tracks as we heard on the LS60s. Obviously a lot of that attributed to the crappy LS60 set up, but there was also a general total change in timbre and texture - strings had more obvious "string" texture on the B&Ws, not the slightly smoothed over sound of the KEF. I also played a few of my test tracks (e.g. Rollercoaster and Missing from Everything But The Girl's Amplified Heart CD).
As before I still found the sound too sculpted for my liking, the zig-zagging in the midrange up that gives that slightly sucked out and then brightened sound. To zingy for my hears for relaxed listening. But I still find them a fascinating listen in terms of how clean, vivid and unboxy the sound is.

Though I figured pretty quick my BIL wouldn't be super impressed...and he wasn't. He didn't really care for them much at all and found them disappointing. A lot of it had to do with the bass. At least in that set up it didn't have a lot of punch or warmth or richness. The Kii 3s had more reach-out-and-sock-you bass feel, which my BIL clearly enjoys.

And speaking of that, we ended up with a listen to (IIRC) the Sonus Faber Olympica Nova III speakers - not their highest tier by far, but still quite pricey - something around 20 grand if I recall? Hooked up to McIntosh tube amplification. Both my BIL and I found this to be the most satisfying sound we heard that day. Totally different from the KEF or B&W - darker, less overtly airy, more "rich" in the mids and certainly much more full, round warm in the bass (which helped the sense of filled out mids I believe). I'd say the set up perhaps could have boosted some of the bass, though they were no closer to the wall than the B&Ws, but with the same tracks (e.g. Missing) the bass just had the solid, landing on the floor, travel along to wash over you sensation lacking in the B&Ws - not exactly neutral. And everything sounded more relaxing to listen to (though still a bit steely around the 3 - 7K region or so, e.g. vocal sibilance and guitar picking). The SF speakers rendered things like drum cymbals as more "complete" sounding than what I heard from the KEFs (and that I've heard from hearing the KEF LS50s many times) - in other words cymbals and anything in the region where the KEF tend to thin out, sound somewhat bigger with more body and solidity. My BIL recognized some of the sound signature - presumably the warmth of the sound and the bass - from owning some older Sonus Faber speakers.

While the SF speakers had a more relaxed sound, as mentioned I heard flashes of the type of electronic hardness that I'm just not accustomed to hearing at home.

Anyway, just thought I'd drop that here since there wasn't anywhere else it seemed to kinda fit....

 

[Travis]

So, what audio/sound is unlike every other endeavor? You need blinding for best results except for audio. Huh-huh.

Not what I said, go back and read it. I said blind testing would certainly be beneficial or required, but not double blind testing. People throw out the term “double blind” (dbt) testing all the time on here and it seems that they don’t even know what it means. You don’t need double blind testing to have a valid sound comparison, straight blind testing is all that would be required.

Again, the reasons for having DBT testing for drug testing/trials don’t apply with audio. Straight blind testing will more than suffice except in the most unusual of circumstances (the vendor who sells the products gives off cues that point towards the product that is twice as expensive for example). The fact you might not be getting the drug under test that could improve/change your life vs. a placebo doesn’t come into play with trying to see which pair of speakers you prefer in your room. As long as the test subject doesn’t know which side f the two speakers is A or B, you will have a valid blind comparison. You are just comparing two pieces of equipment.

So it most definitely should be blind if possible, but it doesn’t need to be double blind.

The reality is that it is extremely labor intensive and technically complex to have proper placement of two different speakers with matched sound levels. Then, after running those tests you need to switch the left and right and run the tests again to make sure there isn’t a room/corner bias.

 

[Travis]

Speakers would be something different, but what can be done isn't especially time consuming.

It’s incredibly time consuming if you want less than 1 dB level matching, and to get something approaching .1dB difference in a room gets more complex. Unless you have two laboratory grade and calibrated microphones, proper stand for height adjustments and distance plus two certified and calibrated spl (dB) meters.

If someone has all of that laying around then yeah, it should be pretty straight forward. The time is going to be relative to the number of times they have been through it. It will get quicker and quicker as they go along.

 

[Galliardist]

Not what I said, go back and read it. I said blind testing would certainly be beneficial or required, but not double blind testing. People throw out the term “double blind” (dbt) testing all the time on here and it seems that they don’t even know what it means. You don’t need double blind testing to have a valid sound comparison, straight blind testing is all that would be required.

Again, the reasons for having DBT testing for drug testing/trials don’t apply with audio. Straight blind testing will more than suffice except in the most unusual of circumstances (the vendor who sells the products gives off cues that point towards the product that is twice as expensive for example). The fact you might not be getting the drug under test that could improve/change your life vs. a placebo doesn’t come into play with trying to see which pair of speakers you prefer in your room. As long as the test subject doesn’t know which side f the two speakers is A or B, you will have a valid blind comparison. You are just comparing two pieces of equipment.

So it most definitely should be blind if possible, but it doesn’t need to be double blind.

The reality is that it is extremely labor intensive and technically complex to have proper placement of two different speakers with matched sound levels. Then, after running those tests you need to switch the left and right and run the tests again to make sure there isn’t a room/corner bias.

I disagree.

If the subject knows what they are being tested for, bias can kick in.

If the person conducting the test with the listener knows which of the test setups is being listened to at any time, they can influence the subject without realising it.

I’ve used straight blind testing myself a couple of times, but such tests should be taken as anecdotal or at best indicative, as far as I’m concerned.

 

[MattHooper]

I disagree.

If the subject knows what they are being tested for, bias can kick in.

If the person conducting the test with the listener knows which of the test setups is being listened to at any time, they can influence the subject without realising it.

I’ve used straight blind testing myself a couple of times, but such tests should be taken as anecdotal or at best indicative, as far as I’m concerned.

Double blind testing arose in response to tests in which the researcher/experimenter can plausibly influence the results. It does not mean that single blind testing
can't be sound (otherwise there wouldn't be countless single blinded tests searchable in any scientific database). As per this piece from The Lancet:

Blinding in randomised trials: hiding who got what​

https://www.thelancet.com/pdfs/journals/lancet/PIIS0140-6736(02)07816-9.pdf

"Many investigators and readers naively consider a randomised trial as high quality simply because it is double blind, as if double-blinding is the sine qua non of a randomised controlled trial. Although double blinding (blinding investigators, participants, and outcome assessors) indicates a strong design, trials that are not double blinded should not automatically be deemed inferior. Rather than solely relying on terminology like double blinding, researchers should explicitly state who was blinded, and how."

What matters is not so much whether trials were double or single blind, but rather the details in how the experiment was carried out. If, in a single blind test, it is implausible that the experimenter had a way to influence the trial results (beyond outright fraud), the single blind test is sufficient.

I think an example could even be a recent blind test I did and wrote about here, between my pre-amplifiers. The preamplifiers were in a room down the hall from my listening room, so I could not see what was being switched, so there was no chance for me to glimpse what was going on during switching. The person administering the test (my son) used a random number generator to create the switching pattern, so he had no say in the switching pattern. The switching was set up so there could be no audible "tell" in the switching from the listening room. The only communication during the test was me yelling "switch" when I wanted the switch to occur. I had zero audible or visual feedback from my son. As far as I can see, such a scenario readily takes care of the problem of possible influence of the administrator upon the results of the test, hence double blinding was not necessary.

Amir's video on blind testing also suggests well run single blind tests are sufficient.

 

[Travis]

If the subject knows what they are being tested for, bias can kick in.

If the person conducting the test with the listener knows which of the test setups is being listened to at any time, they can influence the subject without realising it.

Again, it’s extremely important to keep things in the context of what was being discussed as it was relating to blind testing.

What I said was if a consumer who had narrowed things down to 2 or 3 top contenders for speakers and really wanted to satisfy which was going to be his/her preferred choice, they should do blind testing (I also pointed out that responses that “DBT is the gold standard” don’t really apply to this situation so don’t worry about it.

So again, what you would be testing for is: “which pair of speakers sounds better to you in your own home so you can make a purchasing decision.”

The “subject” is going to know what they are being tested for - whether they have a preferred sound of a pair of speakers: Speaker 1 or Speaker 2. There is just no getting around that.

With the right equipment (matching amplifiers with gain control) you can set this up so that subject/listener is comparing left vs. right, the person conducting the test is simply switching the speaker selector switch (or balance control) on a preamp from left to right, right to left. The speakers are veiled so the listener/subject can’t see them, they were placed and set up by someone else.

I doubt that having the person switch the preamp switch or balance know back and forth is going to provide you cues that would influence the listener. Essentially, all that needs to happen is that the person giving the test needs to say nothing, and not give any feedback of their own. However, if it’s a real concern, have that person leave and the listener can have a spouse, child, other friend do it.

If straight blind testing hasn’t worked for someone in the past it’s because either: the test giver didn’t know when to shut up and kept providing their own feedback, or worse, they wanted to jump in and take the “test” in front of the test subject after they finished the test (“well I want to see which one I prefer, here you switch it back and forth for me”) and then proceeds to tell the subject their preference, or that they have no preference.

 

[cavedriver]

I want to advocate for the next CAF show that we set up a room like some audio clubs have done where you place a black fabric screen between the speakers and the listeners, and a test operator moves wiring, switches switches, etc. according to some kind of plan. Should be pretty close to a "well executed single blind test". Optionally a readout at a nearby computer would let listeners know what they just heard if they so choose and without interacting with the test operator. If the entire test could be conducted using software, so much the better. For example a laptop could manage via a script the settings on a miniDSP. A schedule could tell people that from 1 to 2 pm two different sets of interconnects, or speaker cables, or whatever are being demonstrated, and then from 2 to 3 it would be different speakers, and from 3 to 4 different amps, and so on. That way people could wander in and out of the room whenever they want and without having to force any interaction with the testing methodology.

 

[Galliardist]

Again, it’s extremely important to keep things in the context of what was being discussed as it was relating to blind testing.

What I said was if a consumer who had narrowed things down to 2 or 3 top contenders for speakers and really wanted to satisfy which was going to be his/her preferred choice, they should do blind testing (I also pointed out that responses that “DBT is the gold standard” don’t really apply to this situation so don’t worry about it.

So again, what you would be testing for is: “which pair of speakers sounds better to you in your own home so you can make a purchasing decision.”

The “subject” is going to know what they are being tested for - whether they have a preferred sound of a pair of speakers: Speaker 1 or Speaker 2. There is just no getting around that.

With the right equipment (matching amplifiers with gain control) you can set this up so that subject/listener is comparing left vs. right, the person conducting the test is simply switching the speaker selector switch (or balance control) on a preamp from left to right, right to left. The speakers are veiled so the listener/subject can’t see them, they were placed and set up by someone else.

I doubt that having the person switch the preamp switch or balance know back and forth is going to provide you cues that would influence the listener. Essentially, all that needs to happen is that the person giving the test needs to say nothing, and not give any feedback of their own. However, if it’s a real concern, have that person leave and the listener can have a spouse, child, other friend do it.

If straight blind testing hasn’t worked for someone in the past it’s because either: the test giver didn’t know when to shut up and kept providing their own feedback, or worse, they wanted to jump in and take the “test” in front of the test subject after they finished the test (“well I want to see which one I prefer, here you switch it back and forth for me”) and then proceeds to tell the subject their preference, or that they have no preference.

I remain unconvinced that blind testing for the purpose of purchasing a main component, particularly speakers, is a good idea. You will be looking at them after purchase every day for years and maybe decades. I’ve also read several accounts where people who tried this were disappointed with their systems within weeks, or chose to “upgrade” within months.

I could just be missing all the counter examples, or some detail of the testing. And it’s all anecdotal evidence without a controlled test.

It’s counterintuitive since the blind test should tell us the best choice as we are primarily concerned with the best sound, after all. Something seems to weigh against it though.