My old article on Loudspeaker Imaging

[dbe]

Pragmatism...


gets me into trouble very often. Using terms that are specific to recording guys, acoustics and/or to the circle of people that I've grown around has a way of biting me in the butt. Remember my poor choice of the term "intermodulation distortion" on the MadBored? Man, I paid my dues there! IM is different things to different regimens, EEs' especially.

One of the coolest uses of QSound is on a Suzanne Ciani recording, I can't remember the name right now, though. It is a synth doing sounds that are supposed to be raindrops. They are all around you, even behind. Eyes closed in a dark room and It is magic.

The Polk SRS screws with HF phase a great deal, if memory serves me. Sounded impressive, but wrong.

Dave in ABQ

[dbe]

Kind of makes me...


> Before you mention it yet another time, I
> have done the reverse phase thing, as well
> as just any out-of-phase listening, so I
> have heard the image delocalization it
> generates. Yeah, it's pretty cool.;-)

seasick ;-) But if you turn up a sub to replace the phasey bass it can be fun, if only for a little while. Kind of like acid but without the paranoia.

John k?

Dave

[dbe]

First Watts...


> Yes. And on this system, I didn't get any
> special kick out of vinyl playback, even
> though he had a decent set-up there. I think
> what makes systems stand out in terms of
> Hi-Fi components other than speakers, are
> parts made with a great attention to fine
> detail, keeping the components discrete,
> separating left and right channels, using
> expensive parts.

> The only reason digital may sound worse to
> some people is that the gadgets have been
> mass-produced for a very undemanding market.
> Stuffing a lot of electronics into tiny
> packages is great, but there is a cost.
> Meanwhile a lot of analog stuff is very high
> quality, made with the utmost attention to
> niggling little details that most people
> don't notice, but some do. The reason is
> much bigger, but harder to describe than
> digital vs. analog or tubes vs. silicon.
> Those are just tyhe stereotypes that people
> recognize. I'd love to hear a Nelson Pass
> amp, for example. But I can't afford one.

sound amazing, as do his 1000s'. At the RMAF, Ray Kimber had a large ballroom for his exhibit of his IsoMike recordings. Three hugh Sound LAb U1s' per side in front with 2 U1s' per side in the rear EACH driven by a PAss X1000.5! These were fed the original hard drive recordings in SACD. Unbelievably real. I could hardly drag Gayle (the spousal unit) out of her seat. She told me that it was better than anything she had EVER heard. Being married to me, she has heard alot of systems and has great ears that aren't jaded by years of listening to the god, bad and ugly. I respect her observations.

You hit the nail on the head about great systems. It takes a great deal of time to hit the synergy (fine detail) that it takes to get the "magic". I'm not sure about "expensive parts". I am sure about the right (better) part for the application. Sometimes they are expensive, but it is more about voicing the system with complimentary parts to achieve a final goal that drives my choices. To top it off, everyones' goal ia little different. That is what makes this addiction fun for me.

I love this stuff.

Dave

[cjb]

Some of us still lurk from time to time...


Especially if there's something worth reading. It's a shame though that something I find worth reading I read the first time seven or eight years ago. IMO most people don't know good imaging because they have either never heard it or have never paid attention. But once you realize it's there you can't un-hear it. And if you don't have it now I'll bet you're cooking up a design to get it back. Unless you're like my brother. He just doesn't give a damn. Loud as HELL but CLEEEEEEN! That's his motto.
To each his own I suppose. -CJB

[Wolf]

Nice to hear from you! *NM*

[Wolf]

Re: Ooohhhh...so there it is.


> You mean the 1mH, 4 ohm, 160uF, right?
> Alrighty.

> Did I miss this in previous posts too?

The resistance includes the dcr of the inductor. The 20AWG Jantzen is .7 ohm, so the resistor is a 3.3 ohm resistor. Total of 4 ohms.
Yep- it's been in other posts before.
Later,
Wolf

[Wolf]

I like what I have heard, and experienced...


> Is one that I got from Andy G several months
> before they were available in the US. I
> ended up with a zobel that is 110 ufd, 1 mH
> and 4.7 ohms on these drivers. It measured
> just as good as the 160 ufd, 1 mH and 4 ohm
> version. Great tweeter, but not everyones'
> cup of tea. It lacks that 3rd harmonic
> excess detail signature that most tweeters
> have in abundance.

...every time I've heard an XT tweeter.
Later,
Wolf
> Dave

[Wolf]

Re: It could be flatter...


> and I could be better looking, too :-) Maybe
> not...

Heehee.

> It looks pretty darned good. Interesting
> that you got this reponse from the AR
> crossover. I wondered how they would look.
> Now I know.

> How are you feeling?

Despite the toob in the rt ear, I'm alright. I will probably have it removed, instead of waiting for it to fall out.
Later,
Wolf

[unknownuser]

Re: Vinyl Theories and other interesting diversion


> Then you should write it up and submit it
> for publication.

Na, been there, done that, all you get is the cheap swag, and you have to pay for that (“Dad went to AES '07 and all I got was this crummy pocket protector"). If I still worked in the industry, maybe I’d milk it.

> I'd want to do some hands-on computation on
> how digital algorithms work as well as
> hearing some good vinyl first-hand, but on a
> superficial level the discussion of
> "fractal" information being
> largely lost in digital is pretty intuitive,
> and something that's been on my mind for
> years. In the digital formats, low-level
> information Jeff is referring to will not be
> a separate signal stored at the 5th and 6th
> bits (corresponding to around -60dB); it has
> to ride on top of the total amplitude
> summarized by the 16-bit word length. If you
> calculate out its contribution to the total
> amplitude it could be anywhere from 0.1 to
> 0.001 dB at the signal peaks. How can you
> reasonably expect that to be retrieved when
> each bit is used to resolve everything over
> the 6dB range to which it's been assigned?
> It's actually amazing to me that digital
> resolves any sort of depth perception at
> all. Analog does not have this limitation.

I guess my earlier point went missed. The inherent resolution of any carrier mechanism is mathematically related to the SNR. See Shannon’s work from the 40s or 50s. Its pretty straight forward. SNR sets dynamic range sets resolution available. Doesn’t matter is its analog or digital. Shannon showed you can transport info well into the noise if the noise is chaotic. Analog has groove noise, digital has dither, both can and do carry info below the noise floor.

On top of that, the ear can hear usefully below the noise floor quite easily. A big reason why the noise doesn’t mask since the signal is that (usually) the signal is very un-noise like (not chaotic). Analog has no inherent superiority in this regards. The math is the math. We can’t get hung up on bits.

Not questioning analogs inherent superiority in the imaging dept, but my theory is as good as I can come to a rationale.




(Originally posted by: DDF)

[romanbednarek]

Re: That XT25...


> Great tweeter, but not everyones'
> cup of tea. It lacks that 3rd harmonic
> excess detail signature that most tweeters
> have in abundance.

It is funny that you say that because my friend accused the XT19 of sounding a bit dull despite a flat frequency response measurement. He prefers the more "lively" sound of the OW1 but after getting used to the XT19 and doing some side-to-side listening comparisons I would have to conclude that the XT19 sounds more natural and neutral while the "life" of the OW1 sounds a bit artificial (but it does add a lot of air and depth to the presentation). I recall DLR having to shelve down the top octave of the OW1 because of this (I suspect).

It really is funny how distortion can be interpreted differently depending on the frequency range that it exists in.

[kene]

Re: Vinyl Theories and other interesting diversion


> I guess my earlier point went missed. The
> inherent resolution of any carrier mechanism
> is mathematically related to the SNR. See
> Shannon’s work from the 40s or 50s. Its
> pretty straight forward. SNR sets dynamic
> range sets resolution available. Doesn’t
> matter is its analog or digital.

Yes, this part is pretty straight-forward. What I'm trying to discuss is going a bit more in depth. What you've discussed so far is perfectly applicable for simpler waveforms where all harmonics are approximately equal in amplitude. The issue I'm talking about is if you have two signals, one at 0dB and another at -60dB, which would not be unreasonable for a good deal of ambient information that accompanies music we listen to, and attempt to represent it using the current 16 bit format, the individual words do not see two signals. It sees an absolute amplitude only. If you calculate how much the total amplitude will change as a consequence of the -60dB signal being removed you will see a drop of only 0.004dB. Does current digital have the ability to resolve a change in amplitude this small from a single 16-bit word? This is why I'd like to work with the algorithms myself, but my overwhelming suspicion is that digital won't come remotely close. What must happen instead is several words have to go by until the instantaneous amplitude of the larger waveform drops down to the viscinity of -60dB where the smaller signal can be resolved. This is a limitation of the digital format that does not apply to analog, and to me this rings as a critical factor. It also sounds almost exactly like what Jeff seems to be describing when he refers to "fractal" resolution in analog.

[unknownuser]

Re: Vinyl Theories and other interesting diversion


>Does current digital have the ability to resolve a change in amplitude this small from a single 16-bit word?

Yes, it does, because it supports the -60 db signal. That's the basis of my point, and Shannon.


(Originally posted by: DDF)

[kene]

Think I figured it out.


> Yes, it does, because it supports the -60 db
> signal. That's the basis of my point, and
> Shannon.

Well, I'm not certain you saw how I had been framing the issue, which is fine because I finally remembered some of the reconstruction of amplitude with digital and realized that some of my suppositions were not applicable. This is why I keep saying I would like to work with the D/A conversion algorithms directly myself sometime. This was still worth discussing though.

Despite all that I rethink it all and the issues I see still end up collapsing into the more familiar "16bits => 100dB sampled at 22kHz just doesn't cover it" conclusion. The issue was never with the Shannon stuff you're referencing, at least as far as you've mentioned here; I've known all that stuff for years. Unless he was analyzing issues of digital audio back in the 50's. I wouldn't expect such. ;-)

Insofar as I'm reading you though, well...how would higher ambient noise levels enhance the ability to distinguish depth? If it were really about noise allowing the brain to exercise interpretive liberties in creating ambient cues that aren't really there then I would expect the presence of noise to uniformly instill all sounds with a similar sense of depth, even with purely artificial sounds that contain no spatial information whatsoever. At some point or another this noise should have a homogenizing effect which would impair the ability to resolve things as being closer or farther away. Now I've never listened to a "good turntable" so I don't know if maybe this is exactly what happens, so...well you tell me. Expound on how it might work.

[kene]

Is it just me...


> It lacks that 3rd harmonic excess detail
> signature that most tweeters have in abundance.

> Dave

Or is there also quite an abundance of "Dave"s around here? "Ken"s too, enough that I sometimes think I need to find myself a handle. What's up with that?

[kene]

Re: So much of what's been said...


> A few
> rare recordings do seem to have more depth,
> but it seems to be only those exceptional
> recordings that let this come through.
> "The Trinity Sessions" by the
> Cowboy Junkies comes to mind.

Talk about a fantastic recording. My copy was stolen out of my car a couple years ago (along with my other 75 most-loved CDs); I was just about to burn backup copies of all my discs too! I still need to get it again, along with some 30 other discs.

> I have also never had width to the side
> beyond the left/right position of the
> speakers with the exception of certain
> limited frequency ranges that reflected off
> of a side wall that were related directly
> the the wavelength. False localization
> issues. I do not see how any 2-channel
> system could image to the sides beyond the
> speakers themselves if there is no
> reflection. I have never heard one to do
> that, anyway.

With purely acoustic recordings I haven't heard anything incredible. Most good recordings will give me just a wee little bit outside the width of my speakers, but almost never have I ever had my system in as good a room setup as I would like, so I don't rule out the possibility of side-wall reflections playing a role.

But there are some heavily-engineered discs out there that will employ certain tricks to generate illusions of sounds coming from farther beyond the speakers. Part of how they do this is by employing phase cancellations. Shoot a selected sound, say from the left speaker, and then shoot a phase-cancelling signal from the right, with appropriate sound muffling, to minimize/cancel what your right ear gets to hear. You can tell there are a lot of things that can go wrong so really not a lot of systems (in their rooms) pull it off well in my experience.