Speaker Shootout - two of the most accurate and well reviewed speakers ever made - Page 41 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 1624Likes
Reply
 
Thread Tools
post #1201 of 1494 Old 08-27-2017, 02:51 PM
AVS Forum Addicted Member
 
noah katz's Avatar
 
Join Date: Apr 1999
Location: Mountain View, CA USA
Posts: 21,563
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 1095 Post(s)
Liked: 380
Quote:
Originally Posted by Floyd Toole View Post
You are not alone in thinking that. But nobody has been able to convince "management" that there was sufficient business to warrant the investment. SFM is complicated for a market in which customers expect magic solutions at the press of a button or icon.
Quote:
Originally Posted by John Schuermann View Post
As Dr. Toole says, it's not nearly as simple as running something like ARC or Audyssey - you really have to know what you're doing.

What exactly is required of the customer?

There may be no end to possible subwoofer locations, so let's simplify and assume that there are, say, only three available locations and that's where they're put.

Presumably the user takes measurements at various mic locations within the listening area, then lets the SFM processor crunch the numbers to calculate and apply the subs' proper relative levels and phase to the subs.

That's pretty much how all EQ systems work, so what more does SFM setup require?



Quote:
Originally Posted by John Schuermann View Post
Good questions for me to ask at CEDIA.

Great!
noah katz is offline  
Sponsored Links
Advertisement
 
post #1202 of 1494 Old 08-27-2017, 03:10 PM
Member
 
Floyd Toole's Avatar
 
Join Date: Jun 2017
Location: California
Posts: 189
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 197 Post(s)
Liked: 771
Quote:
Originally Posted by noah katz View Post
What exactly is required of the customer?

There may be no end to possible subwoofer locations, so let's simplify and assume that there are, say, only three available locations and that's where they're put.

Presumably the user takes measurements at various mic locations within the listening area, then lets the SFM processor crunch the numbers to calculate and apply the subs' proper relative levels and phase to the subs.

That's pretty much how all EQ systems work, so what more does SFM setup require?






Great!
SFM, as explained in the AES paper and my books, involves making transfer-function (amplitude and phase) measurements from each possible subwoofer location to each of the seats chosen for preferred treatment - or all seats, if you wish. This can be done with a single subwoofer moved around the room. That data gets stored in a laptop. The next step is to select some number, preferably four, of the measured locations and let the optimization algorithm do the tedious task of finding the best combination of delay, level and one band of parametric equalization in each of the subwoofer feeds. The goal is to minimize seat-to-seat variations so that any global EQ that may be needed benefits all of the listeners. The result of the optimization can be seen, so it is possible to repeat the exercise with different combinations of sub locations to find the best choice, but it will be hard to beat the four corners. Once the best combination is decided on, SFM can be run again with the subs that are intended to be used - they need not be identical.

The simple way is just to decide on four locations, put the subs in place and run SFM. It will do the best it can, which may or may not be the best possible solution.

This differs from "garden variety" EQ in that it is transfer function, not steady-state response, that is used for the optimization. Obviously, with patience, this can be done by trial and error, or perhaps someone else has developed an optimization routine. So long as it does not violate the Harman patent, that too can work.

I don't yet know how well the SDP75 implementation will work. There are several examples of the original system in my books. The gains in bass uniformity, absence of room resonances and efficiency are impressive.
andyc56, Madmax67 and duckymomo like this.
Floyd Toole is online now  
post #1203 of 1494 Old 08-27-2017, 03:30 PM
Member
 
Join Date: Aug 2017
Posts: 68
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 36
>> transfer function, not steady-state response

If transfer function is level + phase and steady-state response is level only, that implies that phase is important in ways other than its effects on (and seen in any analysis of) level.

I could go reread and restudy your earlier analyses and descriptions, but safely assuming laziness on my part, can you put in a sentence or so how phase is important other than its effects on level?

Last edited by davidrmoran; 08-27-2017 at 03:32 PM. Reason: elab
davidrmoran is online now  
 
post #1204 of 1494 Old 08-27-2017, 03:51 PM
AVS Forum Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 24,978
Mentioned: 108 Post(s)
Tagged: 0 Thread(s)
Quoted: 5074 Post(s)
Liked: 3547
Quote:
Originally Posted by Floyd Toole View Post
This can be done with a single subwoofer moved around the room. That data gets stored in a laptop. The next step is to select some number, preferably four, of the measured locations and let the optimization algorithm do the tedious task of finding the best combination of delay, level and one band of parametric equalization in each of the subwoofer feeds.
You're describing using SFM as a subwoofer placement tool as well as calibration/optimization tool. IF four subs in four corners is hard to beat when it comes to seat-to-seat consistency, then might as well use that placement, so SFM ends up being used just for optimization. In which case, should it be any more complicated to use than, say, a JBL BassQ box (spread four mics, push a button, walk away)?
andyc56 likes this.

Sanjay
sdurani is online now  
post #1205 of 1494 Old 08-27-2017, 04:28 PM
AVS Forum Special Member
 
corradizo's Avatar
 
Join Date: Aug 2009
Location: near Chicago
Posts: 2,582
Mentioned: 66 Post(s)
Tagged: 0 Thread(s)
Quoted: 1160 Post(s)
Liked: 865
When multisub is discussed it is usually about placing each of the four subs in the corners of the floor in the room. How about distributing the subs to the front corners of the floor and front corners of the ceiling instead? Would this have the same effect?
corradizo is offline  
post #1206 of 1494 Old 08-27-2017, 04:38 PM
Member
 
Join Date: Aug 2017
Posts: 68
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 36
yes, almost certainly
davidrmoran is online now  
post #1207 of 1494 Old 08-27-2017, 04:47 PM - Thread Starter
AVS Forum Special Member
 
John Schuermann's Avatar
 
Join Date: Apr 2003
Location: Colorado
Posts: 4,770
Mentioned: 51 Post(s)
Tagged: 0 Thread(s)
Quoted: 1370 Post(s)
Liked: 1427
I have yet to do a SFM calibration myself, so don't have hard answers. I have been playing with the JBL SDP75 processor, which will eventually have "SFM2" as a paid upgrade feature. I started writing about the my SDP75 experiences in the JBL Synthesis thread, where such discussion probably belongs. I start talking about it here:

Official JBL Synthesis / Pro / Revel Home Theater Thread

Lots more to come, including additional screen shots. I've already touched on some of the amazing versatility built into the piece, and why running a calibration is not nearly as simple as Audyssey or ARC. I will be updating the Synthesis thread with more impressions, so for those curious about the SDP75, you might want to start hanging out there as well
Madmax67 likes this.
John Schuermann is online now  
post #1208 of 1494 Old 08-27-2017, 05:11 PM
Member
 
Floyd Toole's Avatar
 
Join Date: Jun 2017
Location: California
Posts: 189
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 197 Post(s)
Liked: 771
David: Phase is important because the process involves superposition of sound fields - in particular, the standing waves which have amplitude and phase components. Once the sub to listener data are acquired, the algorithm predicts the sound from any combination of subs at any listener location, including variations of level, time and one parametric filter. The optimization is a brute force minimization of seat-to-seat variations. But we don't mind brutalizing computers

Last edited by Floyd Toole; 08-27-2017 at 05:26 PM.
Floyd Toole is online now  
post #1209 of 1494 Old 08-27-2017, 05:15 PM
Member
 
Floyd Toole's Avatar
 
Join Date: Jun 2017
Location: California
Posts: 189
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 197 Post(s)
Liked: 771
Quote:
Originally Posted by sdurani View Post
You're describing using SFM as a subwoofer placement tool as well as calibration/optimization tool. IF four subs in four corners is hard to beat when it comes to seat-to-seat consistency, then might as well use that placement, so SFM ends up being used just for optimization. In which case, should it be any more complicated to use than, say, a JBL BassQ box (spread four mics, push a button, walk away)?
I should have mentioned that room shape is allowed to stray from rectangular, mine has most of one wall open to the rest of the house. Corners are nice, but not necessary, another area of flexibility, and finally, the subs do not have to be identical as in the passive rectangular room solutions, so you can use what you have.The Bass Q box was a simplification, but it no longer exists.
Floyd Toole is online now  
post #1210 of 1494 Old 08-27-2017, 05:23 PM
Member
 
Floyd Toole's Avatar
 
Join Date: Jun 2017
Location: California
Posts: 189
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 197 Post(s)
Liked: 771
Quote:
Originally Posted by corradizo View Post
When multisub is discussed it is usually about placing each of the four subs in the corners of the floor in the room. How about distributing the subs to the front corners of the floor and front corners of the ceiling instead? Would this have the same effect?
Yes, and if you put them half way up the walls the dominant vertical room mode will be attenuated - which can also be done using subs at both ceiling and floor levels. Because of the efficiency gains in these schemes, some or all of the subs can be smaller. In my room the power distribution among the subs is 100%, 25%, 6.3%, 6.3%, So to have four high-power subs is a waste. The result is smooth non-reesonant bass. Figure 13.18 in my old book and Figure 8.22 in the new one.
Ericglo and corradizo like this.
Floyd Toole is online now  
post #1211 of 1494 Old 08-27-2017, 05:25 PM
AVS Forum Addicted Member
 
noah katz's Avatar
 
Join Date: Apr 1999
Location: Mountain View, CA USA
Posts: 21,563
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 1095 Post(s)
Liked: 380
Dr. Toole,

Thanks for the explanation; a few further questions if I may.

Quote:
Originally Posted by Floyd Toole View Post
SFM... involves making transfer-function (amplitude and phase) measurements from each possible subwoofer location to each of the seats chosen for preferred treatment... This can be done with a single subwoofer moved around the room.
Wouldn't the transfer function of the sub used be important then?

Does the process require making nearfield measurement of the test sub?

Either way, could a vented sub be used, or does the greater group delay rule one out?



Quote:
Originally Posted by davidrmoran View Post
If transfer function is level + phase and steady-state response is level only, that implies that phase is important in ways other than its effects on (and seen in any analysis of) level.
Changing phase of one sub relative to another changes how they sum and is a powerful variable to have control of, esp if it can be varied with freq.

I assume SFM can, or it would be the same as a time delay.

Or maybe not; being able to computationally test permutations of time delay of multiple subs would be a big step forward too.

[edit: I stepped away while writing and Dr. Toole beat me to the bunch.]

Last edited by noah katz; 08-27-2017 at 05:29 PM.
noah katz is offline  
post #1212 of 1494 Old 08-27-2017, 05:29 PM
Member
 
Floyd Toole's Avatar
 
Join Date: Jun 2017
Location: California
Posts: 189
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 197 Post(s)
Liked: 771
Quote:
Originally Posted by noah katz View Post
Dr. Toole,

Thanks for the explanation; a few further questions if I may.



Wouldn't the transfer function of the sub used be important then?

Does the process require making nearfield measurement of the test sub?

Either way, could a vented sub be used, or does the greater group delay rule one out?





Changing phase of one sub relative to another changes how they sum and is a powerful variable to have control of, esp if it can be varied with freq.

I assume SFM can, or it would be the same as a time delay.

Or maybe not; being able to computationally test permutations of time delay of multiple subs would be a big step forward too.

[edit: I stepped away while writing and Dr. Toole beat me to the bunch.]
The transfer function includes the performance of the subs in-situ, including boundary effects, etc. Subs can be closed or vented, or a combination.
Floyd Toole is online now  
post #1213 of 1494 Old 08-27-2017, 05:40 PM
AVS Forum Addicted Member
 
noah katz's Avatar
 
Join Date: Apr 1999
Location: Mountain View, CA USA
Posts: 21,563
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 1095 Post(s)
Liked: 380
Quote:
Originally Posted by Floyd Toole View Post
Yes, and if you put them half way up the walls the dominant vertical room mode will be attenuated...
Similarly, I thought mode cancellation required subs at different distances from the room surfaces related to the mode, so if all subs are on one wall they can't cancel the length modes between it and the back wall.



Quote:
Originally Posted by Floyd Toole View Post
Because of the efficiency gains in these schemes, some or all of the subs can be smaller. In my room the power distribution among the subs is 100%, 25%, 6.3%, 6.3%...

Well that's a fascinating twist.

6.3% corresponds to 12 dB efficiency increase; is that possible because those subs are located such that they (de)energise a strong room mode?

Wouldn't that gain be restricted to just those freq?

But I guess that's fine if the sub is only being deployed to smooth response.

Does SFM report this, so you know how big of a sub you need?
noah katz is offline  
post #1214 of 1494 Old 08-27-2017, 06:48 PM
Member
 
Floyd Toole's Avatar
 
Join Date: Jun 2017
Location: California
Posts: 189
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 197 Post(s)
Liked: 771
Quote:
Originally Posted by noah katz View Post
Similarly, I thought mode cancellation required subs at different distances from the room surfaces related to the mode, so if all subs are on one wall they can't cancel the length modes between it and the back wall.






Well that's a fascinating twist.

6.3% corresponds to 12 dB efficiency increase; is that possible because those subs are located such that they (de)energise a strong room mode?

Wouldn't that gain be restricted to just those freq?

But I guess that's fine if the sub is only being deployed to smooth response.

Does SFM report this, so you know how big of a sub you need?
To control width modes one needs subs positioned across the width of the room. Likewise with length modes. So all on one wall can only manipulate modes in that plane. The delays are in effect "distances".

The output required from an sub is related to, but not exclusively to, the distance from the key listening position(s). It is not a 12 dB efficiency increase, it is a level reduction of 12 dB. The efficiency increase is obtained from the combination of all of the subs operating in unison. If you look at the figures you will see that as more subs are added, they become more effective at reducing seat-to-seat variations, the nulls progressively disappear, leaving smooth steady-state room curves. SFM as it has been implemented tells you everything you need to know to set it up - all parameters for each sub feed. They would be automatically downloaded into the appropriate processor, or set up manually. My original system (ca. 2002) was a manual setup based on a Matlab computation by Todd Welti, the co-inventor.
Aja likes this.
Floyd Toole is online now  
post #1215 of 1494 Old 08-27-2017, 07:07 PM
Member
 
Join Date: Aug 2017
Posts: 68
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 36
Sorry, has the question been answered how the part of "phase" and "time" / "variations of time" that does not affect amplitude works?
The part that is outside (?) amplitude?

(I'm not talking of polarity issues; we want all the subs to be in the same polarity, sure.)

Don't mean to sound argumentative, just not getting it. If 8 subs in the 8 corners have a phase knob, what effects does its setting have on what we hear, at the listening or any other positions, other than level?
davidrmoran is online now  
post #1216 of 1494 Old 08-27-2017, 08:04 PM
AVS Forum Special Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 1,172
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 391
Quote:
Originally Posted by davidrmoran View Post
If 8 subs in the 8 corners have a phase knob, what effects does its setting have on what we hear, at the listening or any other positions, other than level?
I'm not Floyd, but you're responding to Floyd's answer to the question of Noah Katz, who asked this:

Quote:
Originally Posted by noah katz View Post
Presumably the user takes measurements at various mic locations within the listening area, then lets the SFM processor crunch the numbers to calculate and apply the subs' proper relative levels and phase to the subs.

That's pretty much how all EQ systems work, so what more does SFM setup require?
In this post, Noah seemed to suggest that SFM is no different from conventional EQ. Conventional EQ using IIR filters only considers the amplitude response to be in need of correction and applies the EQ to flatten it. But Floyd explained (in effect) that in order to calculate the amplitude response of multiple subs from individual measurements of each sub to each listening position, the phase response of each individual transfer function must be properly taken into account to compute the complex summation correctly, even if we don't care about the phase of the resulting response. SFM is therefore different from conventional EQ in this regard.

But AFAIK, once the superposition theorem has been applied to calculate the (complex, meaning amplitude and phase) response at each position, SFM only considers each position's final amplitude response and seeks to minimize its seat-to-seat variation. This is done without regard to the flatness of said response.

So what might be said of the phase vs. frequency of the properly combined sub responses at each listening position? Is it important from a subjective POV? I don't know, but since it is further corrected by EQ to flatten out the result, one might reasonably ask whether such EQ, when implemented with minimum-phase IIR filters, will also fix up the time domain. This question is best answered by attempting to determine whether the computed response is minimum-phase (with an allowed constant delay due to listening distance) over the frequency range for which we wish to apply the final EQ. This issue is discussed in the Room EQ Wizard documentation in connection with its facility for computing the excess group delay. This measurement can identify frequency regions where EQ using IIR filters is inadvisable.

Whether these effects are audible or not is a big question, but at least there is some way to determine whether EQ using IIR filters can fix the (maybe audible, maybe not) time domain flaws in non-minimum-phase frequency regions of the combined sub responses at each listening position.

Edit: In retrospect, this is probably more appropriate for John's other thread, "Official JBL Synthesis / Pro / Revel Home Theater Thread". Sorry about that.
sdurani, mitchco and Floyd Toole like this.

Last edited by andyc56; 08-27-2017 at 08:58 PM.
andyc56 is offline  
post #1217 of 1494 Old 08-27-2017, 08:26 PM
AVS Forum Addicted Member
 
noah katz's Avatar
 
Join Date: Apr 1999
Location: Mountain View, CA USA
Posts: 21,563
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 1095 Post(s)
Liked: 380
Quote:
Originally Posted by Floyd Toole View Post
The efficiency increase is obtained from the combination of all of the subs operating in unison.

Does that still happen with subs driven at -12 dB?

And it's surprising that driving some of the subs at such low levels is effective; is it that it's mainly deenergizing modes and not shouldering broadband bass duties?
noah katz is offline  
post #1218 of 1494 Old 08-27-2017, 09:59 PM
AVS Forum Special Member
 
Join Date: Dec 2014
Posts: 2,985
Mentioned: 36 Post(s)
Tagged: 0 Thread(s)
Quoted: 1227 Post(s)
Liked: 701
Quote:
Originally Posted by davidrmoran View Post
Sorry, has the question been answered how the part of "phase" and "time" / "variations of time" that does not affect amplitude works?
The part that is outside (?) amplitude?

(I'm not talking of polarity issues; we want all the subs to be in the same polarity, sure.)

Don't mean to sound argumentative, just not getting it. If 8 subs in the 8 corners have a phase knob, what effects does its setting have on what we hear, at the listening or any other positions, other than level?
I think there is a little confusion between phase and delay and the purpose of the phase knob on most subwoofers. The phase knob on most subwoofers is not used if the crossover on the sub is set to bypass or LFE. So you can not use the phase knob on most subs to make a poor man's SFM. You would need to be able to adjust the delay of the subs individually to do a poor man's SFM. The phase knob on a sub is used most commonly in 2 channel configurations where the processor does not have a crossover function and you are using the subs crossover control. The phase knob is used to get the speaker and sub in phase at the crossover frequency.
Floyd Toole likes this.
Ellebob is online now  
post #1219 of 1494 Old 08-27-2017, 11:11 PM
Member
 
Join Date: Aug 2017
Posts: 68
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 36
knob.

not switch

I am sure most know about getting polarity "right" wrt the satellites or whatever.
davidrmoran is online now  
post #1220 of 1494 Old 08-27-2017, 11:45 PM
AVS Forum Special Member
 
CherylJosie's Avatar
 
Join Date: May 2008
Posts: 1,375
Mentioned: 18 Post(s)
Tagged: 1 Thread(s)
Quoted: 732 Post(s)
Liked: 297
Quote:
Originally Posted by Floyd Toole View Post
Humans do not respond to phase shift - at most hearing a small "difference" in anechoic tests using real and contrived signals, but not able to express a preference. In normally reflective rooms there is no response.
[snip]
It isn't Revel that says those things don't matter, or me either. The investigative research was done by several others over the years. It's in the books. But that will never stop companies in a mature market from making claims that differentiate their products. It's business.
The qualifier I would contribute is that linear phase speakers might improve things markedly if the speaker designs across a multi-channel system are otherwise divergent from each other.

I might have sort of stumbled across this issue as a noob building my first multi-channel system from a mixture of 2-way and 3-way speakers that use typical consumer-grade 2nd order crossovers and invert the next driver phase at crossover. This builds in a systematic inter-channel phase conflict in two frequency bands such as maybe 800Hz-1KHz and 3KHz-20KHz.

I am considering removing the 3-way towers and down-sizing from 11.1 to 5.1, but I would prefer that my existing speakers that were marketed for multi-channel use were intelligently designed for it too.

Quote:
Originally Posted by Floyd Toole View Post
Phase absolutely matters in the design of loudspeaker systems because in the crossover region the output from the woofer (say) must add appropriately with that from the tweeter if there is to be a smooth transition between the two. It has to do with the performance of the overall system, achieving good on and off axis frequency responses in the crossover region.
If I understand this correctly, the audible importance of having good on-axis frequency response as well as smoothly varying off-axis response is due entirely to acoustic interactions with the room and its furnishings.

Phase mismatches also are interacting at the listener after reflecting off the room boundaries. It seems inescapable that any phase mismatch between channels must alter the perceived ambiance as well as the imaging.

At least the crossover for the M2 is in DSP so I suppose that helps some if it needs revision.

Dr. Toole, can you confirm that the crossovers of the shootout speakers prevent phase shift between them and differing models within their respective product lines that might be incorporated with them into a single multi-channel system? Does Harman publish speaker phase specs, simulations, or measurements?
CherylJosie is offline  
post #1221 of 1494 Old 08-28-2017, 12:10 AM
AVS Forum Special Member
 
CherylJosie's Avatar
 
Join Date: May 2008
Posts: 1,375
Mentioned: 18 Post(s)
Tagged: 1 Thread(s)
Quoted: 732 Post(s)
Liked: 297
Quote:
Originally Posted by John Schuermann View Post
Thought about live streaming the event, but am concerned about copyright on music and film clips (we had the same problem with our Home Theater Seminar).
YouTube lets me post video of bands that play canned house music between sets.

The standard free YouTube channel includes a system where the channel has to allow the video containing copyrighted work to be seen by anyone, and also allow advertising to appear on the video (readily blocked with an ad blocker btw) so that the copyright owner can monetize it. Many of my recordings have been flagged by the fingerprint checker but my account copyright status is still green. It only gave me issues when I tried keeping the video private and it relented when I made the video public.

No idea what the paid YouTube channel license agreement does though. I never tried live streaming either. Perhaps you can upload the video and see what happens? You have time to fix things. They do not shoot on sight.
CherylJosie is offline  
post #1222 of 1494 Old 08-28-2017, 12:11 AM
Senior Member
 
awediophile's Avatar
 
Join Date: Oct 2013
Posts: 390
Mentioned: 11 Post(s)
Tagged: 0 Thread(s)
Quoted: 282 Post(s)
Liked: 140
Quote:
Originally Posted by davidrmoran View Post
>> I believe the evidence strongly points to direct sound dominating perception

Not clear to me what you mean here by 'perception', but I sense that we are probably done. It is easy enough to show that direct sound to the ear is trivial (unless you are wearing the speakers on your shoulders) to our sense of things with some experiments and also some thought experiments. Imagine playing a piano, or sitting in front of a single loudspeaker say at 2' distance, as the platform where everything is located was moved into a gym, into a closet, into a bathroom, into a field, up into the trees, into a huge anechoic chamber, and then finally back into your listening room with a pass through your kitchen. Write a diary entry on the dozen changes you heard as you played chopsticks or Mozart or boogie-woogie. Or listened to same. The direct sound in all cases was unchanging. What you hear changes dramatically!
By perception in this instance, I mean perception of the anechoic source, which is independent from perception of the acoustic environment that it is within. If you record the piano in each of those different environments and then perform spectral analysis, the balance will vary over the place depending on the environment, yet in most of those instances, we will still hear the timbre of the piano accurately.

Exceptions may be: in the gym if we're sitting so far away that the reverb overwhelms our ability to parse the direct sound from the overall sound-field; and in the closet or bathroom where the direct sound itself is indistinguishable from a high density of early reflections or standing waves. In either of those cases, we cannot accurately gauge the timbre of the source independently from the acoustic effects and we end up hearing a lot of mud instead.

Quote:
Originally Posted by davidrmoran View Post
>>I sense you have not done the sketch or the maths for playback of a 220Hz oboe tone warming up, how many cycles, how many feet to the near corner and thence to you at your seat including the reflected paths, how much time elapses, etc. etc.
Well, the only near corner in my room is about 5 feet from the nearest speaker and about 14 feet from the listening position, so the total path traveled is between 3 and 4 cycles at 220 Hz. Hence, I said the situation is "borderline" in general and will depend a lot on the speaker, placement, listening position, and room. Another thing I know is that my toed-in and wider-than-typical speaker doesn't send as much energy in that direction. Though, the cabinet directivity drops off pretty rapidly below that frequency. So yes, I've thought about these things quite a bit.

As for the time it takes for "a 220 Hz oboe tone warming up", this is an ill-defined quantity. This highlights a central concept of frequency analysis: there is always a trade-off between temporal and frequency resolution. The time for sine wave at *precisely* 220 Hz to "spin up" is effectively infinite because we have already defined the frequency resolution to be maximal. If we instead concern ourselves with a band of frequencies, i.e. a 1/3rd octave band centered at 220 Hz, then I can give you a rough notion of time it takes for that frequency phenomenon to manifest: just a few cycles in this particular case.

Part of the reason this is important is that an oboe does not produce a sustained sine wave. Chances are, the precise frequency of the tone is fluctuating a bit as is the level. I believe our brains care much more about the attack, decay, release, and various modulations of the pitch, level, and timbre that occur as the note is sustained. All of these things are changes, transient things that manifest in the direct sound first, thus giving our brains something to latch on to and follow. Otherwise, we'd just get lost in the cacophony of the reflected sound field that we are surrounded by.

I would also point out that as far as your argument is concerned, there's nothing special about 220 Hz. If you measure an impulse response in the mid-field or far-field in a typical listening room and apply a window of 30-40 ms, i.e., the time you suggest as the approximate ear integration time, you will see a horrible mess of peaks and dips throughout the frequency range all the way to the very top. So you could ask the same question of 4 kHz or 10 kHz. Now imagine the sound of an oboe playing being convoluted with that mess of peaks and dips, and tell me if you think the result has *anything* to do with what you hear as a listener in that room.

Of course, the impulse measurement does hold a lot of useful information that can be extracted by using the right methods of analysis, but that is a subject for another discussion and probably another thread. I may have already overloaded some other readers with a bit too much technical information here.

Quote:
Originally Posted by davidrmoran View Post
Interesting phrasing. The usual phrasing is the ear being 'captured' by the first arrival. But not for timbral / spectral judging.

Another experiment that entails a little work. Which I have done (partly!). If direct sound mattered to anything other than localization, a given speaker would all sound mostly the same as you moved it from environment to environment. In a cubical room made of glass, a public library, a rug showroom, a closet, and outdoors. A public shower. A swimming pool. Do a little bit of this work and see what you think. Borrow some dressing mirrors and put them to the sides. of your speakers Then drape the mirrors with heavy carpet swatches or thick wool blankets. Next move the speaker into the corners. Then relocate them on the long wall. Place them so the woofer is equidistant from the three near boundaries (corner), and finally stagger them (this one is a bit subtler).

In most cases, to my ear, almost everything changes, and you can hardly tell it's the same speaker. I am not alone in this finding.

And nowhere is that more true than in the lower midrange, 220Hz. Well, maybe that's a lie. God knows the playback heard in your ear centered on 3k will change enormously. Ditto for 35Hz. Ditto for drum brushwork and rimshots centered on 3k-9k.

Just do the rug-mirror part, and the indoors-outside part.
I agree that this is probably true to some extent at 220 Hz, but is it really true for mid and high frequencies, i.e. above a threshold around about 250-1000 Hz? My understanding of Dr. Toole's work is that above some frequency, we mainly hear the speaker *despite* the room. That doesn't mean that we don't hear the room at all, but rather that the speaker sound and room sound are independently perceived. Dr. Toole, if you are reading, please correct me if I am misreporting your opinion here.

To go back to the piano thought experiment, we can hear it's the same piano no matter what environment we're in, provided that the acoustics aren't so bad as to interfere with the perceptual process. The actual sound we hear is totally different in each case, but the sound that we attribute to the source, the piano in the room, is always the same. Do you not agree?

The reasons above make a strong case for why direct sound dominates both localization and timbre perception even as early reflections modify that perception. If you don't agree with the above, then perhaps we have indeed reached the end of our productive discourse.

Oh, I see you mention 3 kHz. I guess that may be a valid exception, though I'm not sure rimshots or brushes are really sustained long enough to excite the ear resonance much. How about violins? They can hit 3k pretty hard with sustained tones. Though honestly, speaking from experience of playing in string orchestras growing up, I never noticed differences in the sound of live violins that I could not attribute to the acoustics of the room or the skill of the performer(s). OTOH, I think I've only heard two speakers in my life that did the sound of the violin justice: the Revel Salon 2s and my own.
awediophile is offline  
post #1223 of 1494 Old 08-28-2017, 12:46 AM
Senior Member
 
awediophile's Avatar
 
Join Date: Oct 2013
Posts: 390
Mentioned: 11 Post(s)
Tagged: 0 Thread(s)
Quoted: 282 Post(s)
Liked: 140
Quote:
Originally Posted by davidrmoran View Post
Sorry, has the question been answered how the part of "phase" and "time" / "variations of time" that does not affect amplitude works?
The part that is outside (?) amplitude?

(I'm not talking of polarity issues; we want all the subs to be in the same polarity, sure.)

Don't mean to sound argumentative, just not getting it. If 8 subs in the 8 corners have a phase knob, what effects does its setting have on what we hear, at the listening or any other positions, other than level?
In engineer speak: with multiple subs playing in the room, the response at any one location is a superposition of the individual responses. This result can be computed by summing either the impulse responses or complex (i.e. including magnitude and phase) transfer functions. These are interchangeable representations, so one may choose either at one's leisure. The key requirement for this to work is that the response measurements must contain phase/temporal response information and they must use a common temporal reference.

As you may be aware, when multiple sounds exist simultaneously and are summed, they may interfere constructive or destructively at different frequencies to varying degrees depending on the phase difference. So the phase at a particular listening location of each of the subwoofers ultimately influences the magnitude response at that location. It's no different from how multiple drivers influence one another in a crossover, except of course that subwoofer wavelengths are much longer and tend to be more strongly influenced by the room.

The "phase" knob on some subs is really just a delay control. Delay shifts phase indirectly by (IIRC) phi = -omega*t, where phi is the phase angle *change* at any particular location, t is the delay in seconds, and omega = 2*pi*f where f is frequency and pi is the mathematical constant. This means that phase cannot be altered independently for each frequency using such a control.

SFM's ability to optimize delay settings on each sub is essentially the same as optimizing the "phase" controls on subs, if such controls exist. However, to my knowledge, SFM does not have the ability to otherwise alter phase of each sub independent of magnitude and frequency, which requires more sophisticated filters and, in many cases, additional processing latency. SFM can and does alter the phase of the combined response at each seat, but I doubt it's designed to pay much attention to that, given the relative influence of phase response compared to magnitude response on perception.

If anyone wants me to go into detail or describe things in less technical terms, let me know, but I agree that this discussion should probably occur in a different thread.
Madmax67 and Floyd Toole like this.
awediophile is offline  
post #1224 of 1494 Old 08-28-2017, 01:17 AM
Senior Member
 
awediophile's Avatar
 
Join Date: Oct 2013
Posts: 390
Mentioned: 11 Post(s)
Tagged: 0 Thread(s)
Quoted: 282 Post(s)
Liked: 140
Quote:
Originally Posted by CherylJosie View Post
The qualifier I would contribute is that linear phase speakers might improve things markedly if the speaker designs across a multi-channel system are otherwise divergent from each other.

I might have sort of stumbled across this issue as a noob building my first multi-channel system from a mixture of 2-way and 3-way speakers that use typical consumer-grade 2nd order crossovers and invert the next driver phase at crossover. This builds in a systematic inter-channel phase conflict in two frequency bands such as maybe 800Hz-1KHz and 3KHz-20KHz.

I am considering removing the 3-way towers and down-sizing from 11.1 to 5.1, but I would prefer that my existing speakers that were marketed for multi-channel use were intelligently designed for it too.
This is an excellent observation! Though, phase linear speakers is kind of an overkill approach to achieving phase match between different speakers.

I compliment you because I see recommendations all over the place that people use identical speakers or speakers from the same product line in order to achieve "timbre matching". The funny thing is that if speakers were accurate, they would already be timbre matched. The real struggle is phase matching, and that's something different speakers from the same product line rarely achieve because they usually use different crossovers.

Thankfully, phase mismatch can usually be mostly repaired using simple IIR filters. However, the methods for doing so are advanced and require knowledge of the speaker and/or accurate measurements to get right. I've done it in my room for my fronts vs. surround speakers, and it's well worth it. Phantom imaging and even perceived tonal balance of phantom sources is much improved.

This falls in the general category of DSP and room EQ answers: Can it improve sound quality? Yes, absolutely. Will it improve sound quality? It depends on how it's done. How do I do it right? Umm, it's complicated. I don't know of any room EQ systems that address this properly. Audyssey definitely doesn't. Even though Dirac does attempt to some extent to linearize phase, it's not necessarily prioritizing phase matching *between* speakers. If you run it, maybe you'll get lucky. I have no idea if Arcos or Trinnov try to correct phase mismatch.

I'm sorry that I can't give a better answer than this. If people bug me enough, I be talked in to trying to explain it in a different thread. Or give me a few more years and maybe I'll be ready to sell something.

Quote:
Originally Posted by CherylJosie View Post
If I understand this correctly, the audible importance of having good on-axis frequency response as well as smoothly varying off-axis response is due entirely to acoustic interactions with the room and its furnishings.

Phase mismatches also are interacting at the listener after reflecting off the room boundaries. It seems inescapable that any phase mismatch between channels must alter the perceived ambiance as well as the imaging.

At least the crossover for the M2 is in DSP so I suppose that helps some if it needs revision.

Dr. Toole, can you confirm that the crossovers of the shootout speakers prevent phase shift between them and differing models within their respective product lines that might be incorporated with them into a single multi-channel system? Does Harman publish speaker phase specs, simulations, or measurements?
I think I can answer most of your questions here.

Yes, smooth off-axis performance is important because of acoustic interactions, and also because someone may want to enjoy the sound somewhere other than on-axis. No one listens in an anechoic chamber (on purpose) and you shouldn't either, so this is a key requirement for a good speaker.

Phase mismatches don't matter much for reflections because the reflections themselves introduce new phase shifts. Any delay, such as the time it takes for sound to travel through the room and interact with various boundaries, also shifts the phase of the sound relative to other sounds. Nevertheless, image location perception is strongly influenced by the first arrival of sound, so phase matching of speaker responses is critically important for the best phantom imaging. If the speakers are different distances from the main listening positions, then setting the delay for them accurately is also very important.

About ambiance. The word serves as a broad term that encompasses a few different things, some of which can be perceived in mono (including reverb, actually) and some only in stereo (such as the enveloping aspect of reverb). A key characteristic of reverb in real life is that the phase of the incoming sound is essentially randomized in all directions. So when captured in a recording (or often synthesized) in stereo, reverb presents as signals whose phases are constantly varying independently between the two channels. The term for this is decorrelated sound. Phase matching between speakers is irrelevant for accurate presentation of decorrelated sound.

AFAIK, the crossover of the M2 is not linear-phase, so phase matching between the M2 and other speakers is still a problem to address.

And lastly, the shootout we did was done in mono, so phase didn't really come into play there. If a shootout were to be done in stereo and for some strange reason we opted to use different speakers for left vs. right, then phase matching would definitely have an impact on judgments. But that's just silly.
awediophile is offline  
post #1225 of 1494 Old 08-28-2017, 07:55 AM
AVS Forum Special Member
 
localhost127's Avatar
 
Join Date: May 2009
Posts: 2,308
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 32 Post(s)
Liked: 72
Quote:
Originally Posted by awediophile View Post
Sorry, the part in bold should have said something like "these models don't work as well for low frequencies (but still above the modal region)" in small rooms. The fact is, you can't "spot treat" at a first reflection point, say using a 6"x6"x6" and expect to absorb enough low frequencies to matter. The sound can and does diffract around the panel.
your statement was in the context of the specular region. you continue to ignore that the specular region of a bounded space is not a static definition - it is dependent upon room/boundary dimensions. a large (flat/planar) boundary that is large with respect to a "low frequency" still adheres to the principle.

your comment regarding a 6"x6" panel is another strawman with no relevancy. and it was already noted (and renoted) in the original response that a diffuser must be large with respect to wavelength "to be seen" (thus to limit diffraction around). you seem to be mixing (or confusing) the context of the conversation to steer away from the original responses.

perhaps you could be more specific and actually illustrate how "these models dont work" - in detail - under the strict context of the conversation that has been provided.

Quote:
Originally Posted by awediophile View Post
Your statement that diffraction is a different concept and is outside the context of this conversation belies a misunderstanding of the physics. *Diffraction is the fundamental concept*. Sound is a wave. The specular-like behavior of high frequency sound can be understood as a consequence of the sound traveling via the multitude of alternative pathways destructively interfering and canceling out. See the classic Double-slit experiment for the optical analogy.
yes, diffraction is a fundamental concept of wave behavior but not in the context of the original statement being made. could you actually explain how diffraction is in play with respect to the original reply that you made: "Diffraction is the term people use to describe physics that defies the over-simplistic specular model".

what do you actually mean by this use of the word? and why do you continue to purposely ignore the strict context of which a statement is made under?

Quote:
Originally Posted by awediophile View Post
Mind you, it's not necessarily convenient to reason about sound as a wave all the time, and the specular model is very helpful, but the effects of diffraction must be kept at the back of one's mind at all times. They really do matter a lot for low frequencies.
again, missing the context of the original statement that is with respect to the specular region where boundary size is large with respect to wavelength.

Quote:
Originally Posted by awediophile View Post
That's not exactly what I meant. I meant that installing mid and high frequency diffusion on a bare wall does not degrade the sound field in the same as degrading as installing mid and high frequency absorption on a bare wall. Though really these statements are kind of lacking in meaning without specifics.
my original statement was that shallow diffusers are tantamount to thin porous absorbers - as each are only effective to the HF band and thus allow the lower-mid band to persist - thus impeding the listening position with a colored/EQ'd/LPF reflection. this is exasberated further by the fact that a typical loudspeaker will have less HF energy dispersed off-axis (eg, to sidewall) than lower/mid band energy.

what is your citation that a shallow diffuser "degrades the soundfield" more-so than a thin porous absorber - both of which are non-broadband treatments that color the specular reflection?


Quote:
Originally Posted by awediophile View Post
In the first paragraph, I thought you had missed my word "marginal", but I see you quoted me on that in your second paragraph. Yes, I am aware of the fact that real world diffusers also absorb some, but it is a marginal effect if they aren't poorly built. And in fact, certain types of diffusers (pseudo-Schroeder or non-Schroeder).
the absorption characteristics of phase gratings are fairly well understood and thoroughly explored by the work of trevor cox and peter d'antonio (RPG Inc) - and they are not inherently benign.

Quote:
Originally Posted by awediophile View Post
As for blackbird studio C, that is not a valid example of diffusers causing absorption. The -30 dB refers to the level of the ETC curve following the rapid early decay. It is not caused so much by absorption in the diffusers but rather by a spreading of acoustic energy throughout the time domain. That room is anything but anechoic.
sorry but you cannot simply hand-wave away an obvious fact. the first-order returns modified solely by the use of wooden diffusers attenuates the signal by -30dB. the losses from edge diffraction not "marginal". the room is not "anechoic" (it has a wooden floor and humans also deal with the integration of the diffused returns over 0.3s) - but it highly damped. it's an absolute valid example detailing how diffusers induce losses. simply saying something is not "a valid example" does not override real world measurements.


Quote:
Originally Posted by awediophile View Post
Also, I've heard it said that many people cannot stand the sound in that room and put up temporary curtains and other things to try to deal with it. Perhaps the problem is that the source and receiver to the diffusers are too near-field and they don't actually operate as intended, creating a sound full of weird resonances. Trust me, I've made that mistake in my space and it makes things sound very weird and wrong, almost robotic.
ok - your original statement was that: "Diffusers, when they are properly designed, only marginally absorb sound, so unlike absorbers, they are not permanently removing energy from the room within their active bandwidth. The made analogy is flawed." so you erroneously make the claim that diffusers do not "permanently remove energy from the room" and then instead of providing citation or evidence against a room that clearly details this, you attempt to distract and go off on a tangent about the room's response. whether or not people prefer working in the room or prefer absorption at sidewalls vs diffusion is entirely independent of the real-world behavior and absorptive losses the broadband diffusers impart - entirely contradicting your statement. just another distraction away from the original claim.

Quote:
Originally Posted by awediophile View Post
Oh man, I have some bad news for you. Diffusers do this anyway. Have you ever seen proper measurements of diffusers? Even the "good ones" typically have diffusion coefficients that struggle to break 0.5 and quite a bit of variation vs. frequency. Do you know that the QRD algorithm only theoretically guarantees good diffusion at a few key "design frequencies"? You know what the theory says about the rest? Zip, zero, zilch. Same thing with PRDs. Thankfully empirical measurements indicate that they work decent, but they are a long way from being optimal.
again, you continue to propagate the very well known and understood fundamentals of reflection phase gratings as some sort of "gotcha". thankfully empirical measurements also indicate an inherent lossy component (1/4wave resonance, viscous losses, edge diffraction, etc) - negating your original statement that diffusers do not permanently remove energy from the room.

if you have data to contrast in comparision to cox & d'antonio's work to support your statement, then please provide it.

Quote:
Originally Posted by awediophile View Post
Those nifty looking fractal diffuser designs? They come with all kinds of weird side-effects. Are these issues relevant to the listener? Good question. At best, we have to rely on opinions and accept that no room design will be theoretically perfect.

And of course, near-field vs. far-field is potentially a big deal. The picture you posted of the large, layered rear wall diffuser looks pretty cool, but the room is probably far too small for that thing to be operating in the far field. Does it make things sound good? Of course, and that's ultimately what's important.
the devices were presented as examples of ways of extending the bandwidth of a diffuser in response to the statement: "Can you point me to what those look like, broadband diffusers properly scattering sound down to an octave-plus below middle C? "

their proper use and deployment (eg, real estate constraints) is entirely independent of the fact that they exist and can be used to extend their bandwidth to create "broadband diffusers properly scattering sound down to X". phase gratings scale with wavelength - this is entirely independent of whether it is useful in a small residential size room. that's an entirely different conversation.
omholt and Tomas2 like this.
localhost127 is offline  
post #1226 of 1494 Old 08-28-2017, 09:12 AM
Member
 
Floyd Toole's Avatar
 
Join Date: Jun 2017
Location: California
Posts: 189
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 197 Post(s)
Liked: 771
Quote:
Originally Posted by awediophile View Post

I compliment you because I see recommendations all over the place that people use identical speakers or speakers from the same product line in order to achieve "timbre matching". The funny thing is that if speakers were accurate, they would already be timbre matched. The real struggle is phase matching, and that's something different speakers from the same product line rarely achieve because they usually use different crossovers.

AFAIK, the crossover of the M2 is not linear-phase, so phase matching between the M2 and other speakers is still a problem to address.

And lastly, the shootout we did was done in mono, so phase didn't really come into play there. If a shootout were to be done in stereo and for some strange reason we opted to use different speakers for left vs. right, then phase matching would definitely have an impact on judgments. But that's just silly.
Thank you awediophile for stepping in and answering the last string of questions - your responses were excellent. For me it is a rare treat to find someone who actually understands this stuff.

The loudspeaker phase response issue doesn't want to go away, even though it is about as important as speaker wire or power cords to what we hear - except, as you point out, when two loudspeakers are involved in creating a phantom image. In practice this is the essence of stereo so a stereo pair must be identical in magnitude and phase, but the actual phase response is unimportant. In multichannel soundtracks the center channel gets added and if it is involved with image panning that involves the L & R then there could be problems. In movies this is unlikely as such panning tends to be associated with rapidly moving sounds. I have no comment on multichannel music as there is so little of it - except in music videos where it is often screwed up by having the featured artist delivered by all three front channels - dumb - or in stereo, ignoring the center channel - a waste.

Even then it only works for the person in the sweet spot.

In immersive sound employing panning among some number of speakers (depending on the size of the intended "image") distributed over walls and ceilings, there are substantial arrival time differences among the multiple loudspeakers that can be simultaneously radiating the same sound, so the "purity" of phase is corrupted - even for the sweet spot. But one can safely predict that "calibrating" individual immersive loudspeakers will become a "feature" of high-end home theaters, even if the listener is substantially off the design axis.
Floyd Toole is online now  
post #1227 of 1494 Old 08-28-2017, 12:32 PM
Member
 
Join Date: Aug 2017
Posts: 68
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 36
>> l still hear the timbre of the piano accurately.
>> The actual sound we hear is totally different in each case, but the sound that we attribute to the source, the piano in the room, is always the same. Do you not agree?
>> The reasons above make a strong case for why direct sound dominates ... timbre perception


I do not know what 'accurately' would mean, nor the notion of sound 'always the same', since it's clearly not the case with my ears nor those of anyone I know.
So not only do I not agree but would add that it's not a matter of agreement or argument, according to the wisest texts and experiments and audio history of ideas. There's a reason any source sounds different in a tile bathroom from outdoors.

There has been the occasional soul in the past who has mistakenly asserted that timbral perception and tonal balance assessment come from / are dominated by direct sound. I haven't heard that said seriously for decades. Crossover work wouldn't matter in audio, really, nor would boundaries. What a waste for hearing to need integration time to make its judgments.

Right about summed LF magnitudes. Thought there was something more being asserted.
davidrmoran is online now  
post #1228 of 1494 Old 08-28-2017, 12:43 PM - Thread Starter
AVS Forum Special Member
 
John Schuermann's Avatar
 
Join Date: Apr 2003
Location: Colorado
Posts: 4,770
Mentioned: 51 Post(s)
Tagged: 0 Thread(s)
Quoted: 1370 Post(s)
Liked: 1427
If I'm understanding the above correctly, I believe the assertion is that if I take my old Story and Clark upright piano, and put it in a wide variety of venues, I would always be able to tell that what I am hearing is my old Story and Clark piano. I would most likely not mistake it for a Steinway Grand or any other piano (even one much closer in character to my Story and Grand), no matter what the environment - even a tile bathroom.

awediophile likes this.

Last edited by John Schuermann; 08-28-2017 at 02:53 PM.
John Schuermann is online now  
post #1229 of 1494 Old 08-28-2017, 01:01 PM
Member
 
Join Date: Aug 2017
Posts: 68
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 36
Quote:
Originally Posted by John Schuermann View Post
If I'm understanding the above correctly, I believe the assertion is that if I take my old Story and Clark upright piano, and put it in a wide variety of venues, I would always be able to tell that what I am hearing is my old Story and Clark piano. I would most likely not mistake it for a Steinway Grand or any other piano (even one much closer in character to my Story and Grand), no matter what the environment - even a tile bathroom.

That's what I believe that Dr. Toole and awediophile are arguing, and I agree.
shoulda never introduced piano, I see
davidrmoran is online now  
post #1230 of 1494 Old 08-28-2017, 01:01 PM
Newbie
 
Join Date: Jun 2017
Posts: 13
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 13 Post(s)
Liked: 3
Quote:
Originally Posted by Floyd Toole View Post

I have no comment on multichannel music as there is so little of it - except in music videos where it is often screwed up by having the featured artist delivered by all three front channels - dumb - or in stereo, ignoring the center channel - a waste.
I've been surprised by the number of albums available in multichannel from independent classical labels like Channel Classics, BIS, and Pentatone. Many are excellent.

Of course I wish there were more!
John Schuermann likes this.
Jesse44 is offline  
Sponsored Links
Advertisement
 
Reply Speakers

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off