Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[ErikN]

I gather that JRiver has its roots in the Audiophile community. With WMC's demise, there will likely be more users like myself who would have chosen it had it only supported basic system audio (gasp). Whereas I imagine many of you would have selected MC even if it didn't play video. I'd like to understand an audiophile's use of JRiver and, maybe, also an audiophiles approach to music in general. When possible, technical answers would help.

(1) What are the sources of your music? It seems like everything 'audiophile' boils down to minimizing error. It also seems correct to think that nothing gets better than the master copy. Do you find special studio recordings? Is it possible -- or worth it -- to capture 'high fidelity' recordings of live concerts? Are there 'audiophile' worthy recordings of mainstream pop / rock ?

(2) I assume it's a given to use a lossless encoding. How do you select one over the other? They are lossless so it is not quality. I'd be surprised if one is 50% smaller than the others. Is it compatibility?

(3) Now the crux. Where does JRiver fit in. You have a pristine recording, high sample rate, high bitrate, losslessly encoded. I assume this stays digital on an optical link to a high-end external DAC, hi-current amp, and speakers that can faithfully render God's laugh. In theory, a hands-off approach doesn't sound entirely wrong.

(4) Do you have speakers that can actually render the difference between 48Khz, 96Khz, and 384KHz sampled audio? Given that we can send/recieve signals from the edge of our solar system I will concede the ADC, DAC, and amp as doable. Making something that can faithfully move air like a violin, voice, saxophone, and drum with the same, low error rate as a $10K DAC seems amazing.

(5) In an answer to an unrelated question, one JRiver user noted having a 4TB audio library. On average, what is your encoded bit rate (i.e. if you ball-parked a 60 min album how much space would you want) -- assume your best quality stuff? How many albums/tracks do you manage within JRiver?

(6) Anything else relevant I should have asked?

[JimH]

I'll just give you a couple of quick answers.

Any lossless format will deliver the same sound to the DAC.  FLAC is probably the most widely supported, but it doesn't matter much what you choose since you can always convert without any loss.

For lossless, a good estimate of storage required is about one GB for 3 CD's.

There are compression settings available when encoding lossless, but they don't affect quality.  They only allow you to save 1 or 2 per cent in storage required, but they are slower, and not worth the trouble.

There is a wiki topic you might find useful:
http://wiki.jriver.com/index.php/Audiophile_Info

[blgentry]

(1) What are the sources of your music? It seems like everything 'audiophile' boils down to minimizing error. It also seems correct to think that nothing gets better than the master copy. Do you find special studio recordings? Is it possible -- or worth it -- to capture 'high fidelity' recordings of live concerts? Are there 'audiophile' worthy recordings of mainstream pop / rock ?

The "provenance" of a recording is probably the BIGGEST single factor in sound quality.  Meaning, what version of the album, mastered where, and most importantly, how it was recorded.  So this is a really good question.  Some albums are just recorded well to start with.  Like the super famous Fleetwood Mac Rumors album.  Very carefully done and the results are fantastic.  Still an audiophile standard today and that recording was made in 1977.

Today's "big deal" in recording quality is the so called Loudness Wars.  I'm constantly searching for the "best" version of albums I want to buy.  My primary criteria for most albums today is making sure I get an album with good Dynamic Range.  I use the Loudness War database to try to find the best version:

http://dr.loudness-war.info/

(2) I assume it's a given to use a lossless encoding. How do you select one over the other? They are lossless so it is not quality. I'd be surprised if one is 50% smaller than the others. Is it compatibility?

Yes, they are all essentially the same.  FLAC has the highest compatibility with most equipment.  WAV is the worst choice because it's tagging support is poor and not standardized.   FLAC and ALAC are going to be roughly 60% the size of WAV or AIFF, as FLAC and ALAC use LOSSLESS compression (like a ZIP or RAR file) internally to save space.  They still have 100% of the information, just like WAV or AIFF.  They are just smaller, like a ZIP file is smaller than the original file.

(3) Now the crux. Where does JRiver fit in. You have a pristine recording, high sample rate, high bitrate, losslessly encoded. I assume this stays digital on an optical link to a high-end external DAC, hi-current amp, and speakers that can faithfully render God's laugh. In theory, a hands-off approach doesn't sound entirely wrong.

JRiver is the player and the cataloging or Library software.  It's incredibly important in terms of organizing and presenting your music to you.  If you are an audio purist, you use as little processing inside of your player as possible.  No equalization, loudness processing, etc.  That's all up to the end user.  JRiver will send the pure untouched output, or the modified output directly to your sound card or DAC.  I prefer to use external DACs, as I think hey have higher attention to detail and more noise immunity.  USB is a great connection method.  Some will argue that SPDIF optical is a better output from computer to DAC.  I think USB sounds fantastic.  Just MHO.

(4) Do you have speakers that can actually render the difference between 48Khz, 96Khz, and 384KHz sampled audio?

That's a hotly debated topic:  Does high resolution audio sound different than 16/44.1 ?  I can only say that I think I hear differences on the very small number I have sampled.  I haven't done blind testing.  But I think I hear a difference.  You be the judge.

(5) In an answer to an unrelated question, one JRiver user noted having a 4TB audio library. On average, what is your encoded bit rate (i.e. if you ball-parked a 60 min album how much space would you want) -- assume your best quality stuff?

Jim is correct on this.  FLACs made from CDs run around 300 to 350 MB per album.  Some a little more, some a little less.  High resolution stuff is WAY bigger.  About 1 GB per album for 24/96.  Almost double that for 24/192.

How many albums/tracks do you manage within JRiver?

Me?  Not many.  About 300 albums right now.  Growing slowly as I build my collection.  That's music.  I'm building a video library with MC also.   Some MC users have 500,000 tracks in their libraries and it's said to perform well at that size.  Other software, like itunes for example, falls on it's face somewhere in the 50,000 range.  I don't think I'll ever have that much music.  But some people do!  Here's some speed and size stuff from JRiver's site:

http://www.jriver.com/speed.html

Interesting topics.

Brian.

[Hendrik]

(4) Do you have speakers that can actually render the difference between 48Khz, 96Khz, and 384KHz sampled audio?

Speakers deal in analog, the sample rate makes no difference at their end anymore - any analog waveform is entirely continuous.
But on that note - even if audiophiles might not agree or even want to hear this, but high sample rate is not beneficial. The science of audio sampling dictates that you can always represent a waveform with perfect accuracy up to the Nyquist-Frequency (half the sample rate). So a 44.1kHz audio file can accurately represent any audio up to 22.05kHz - increasing the sample rate is only going to increase the possible audio it can represent, not the precision or quality of the audio below the limit.

I like to link to this article that tries to clear up some of these points, it was born out of similar questions, written by an audio expert from Xiph (the organization behind the Vorbis and Opus audio codecs): http://xiph.org/~xiphmont/demo/neil-young.html

[ErikN]

The sample rate and bit width article was quite interesting.  In regard to speakers, my question could have been more clear. Is seems like a lot of effort goes into keeping a pure signal from studio source all the way to your home speaker. However, if I had to guess, it seems like the biggest source of error would be the speaker itself -- or I guess speakers. For example, consider the following -- which I'm just kind of winging:

Assume you have two channels with a tweeter, mid, and bass each.

After the amp you use an analog high-pass, band-pass, and low-pass filter to split each channel and send to your tweeter, mid, and bass. I imagine you must have noise/distortion from the filters and from the trying to move a physical, heavy diaphragm that far exceeds any distortion introduced anywhere else? Plus, are there setups that take into account the fact that, for example, a mid-range weights 20x? that of a tweeter. Wouldn't this result in some kind of phase distortion?

With a fair amount of expense you could probably avoid the filter noise by doing it digitally. This seems great expect now you need 3x DACs and 3x Amps. However, you are still left with the speakers.
 
Is there any technological advance on the speaker/rendering side that can match the level of fidelity kept by the result of the studio-to-home flow?
Maybe a surgically implanted bluetooth-to-cochlear-DAC (with JRiver personal-zone support of coarse) ?


A secondary question, is there a standard for mic placement in studio recordings? For example, if a studio places their mics 6 feet apart and 3 feet from the artists but the home's speakers are 15 ft apart and 10 feet from the listener does this matter?

[mwillems]

The sample rate and bit width article was quite interesting.  In regard to speakers, my question could have been more clear. Is seems like a lot of effort goes into keeping a pure signal from studio source all the way to your home speaker. However, if I had to guess, it seems like the biggest source of error would be the speaker itself -- or I guess speakers. For example, consider the following -- which I'm just kind of winging:

Up to here you're correct.  Speakers are by far the largest source of distortion in most audio chains (in competition with the room you're in).  Very nice speakers can have measured distortion profiles below 1%.  Most have significantly more.  

Assume you have two channels with a tweeter, mid, and bass each.

After the amp you use an analog high-pass, band-pass, and low-pass filter to split each channel and send to your tweeter, mid, and bass. I imagine you must have noise/distortion from the filters

Crossover filters are not typically a significant source of harmonic distortion because in most speakers they're passive and there is only so much distortion that can be introduced by a competently designed passive filter.  Passive crossovers do introduce phase distortion, but when properly designed this may or may not be audible (depending on your system and whose testing you believe).

and from the trying to move a physical, heavy diaphragm that far exceeds any distortion introduced anywhere else?

This is the real source of most speaker distortion: you're ultimately using a piston (or some vibratory substrate) to move air; the longer and faster the piston has to travel the more distortion you introduce.  One general way to reduce this kind of distortion is to use very sensitive speakers that also have high power handling so that the physical elements are moving as little as possible.  That's not a guarantee, but it's one of the ways that (for example) horn-loaded compression drivers manage to deliver such low distortion when compared to similar conventional cone or dome speakers [EDIT: This was snarled up and inaccurate; there's no replacement for displacement; with conventional speakers increasing cone area will improve distortion performance (because the element does not need to move as far to move the same amount of air); horns can also deliver lower distortion because the horn provides a more efficient coupling to the air allowing a speaker to displace more air while moving the same distance]

Plus, are there setups that take into account the fact that, for example, a mid-range weights 20x? that of a tweeter. Wouldn't this result in some kind of phase distortion?

Here you've lost me.  There are phase issues when you have a single driver reproducing a very wide frequency range (the same cone will have a hard time producing sound many octaves apart and there will be lag), but using separate tweeters and mid-ranges solves that exact problem because the mechanical elements can be driven separately.  The speker designer will match the sensitivities of the drivers (or pad one down) so that they both react similarly to the same strength of signal.

With a fair amount of expense you could probably avoid the filter noise by doing it digitally. This seems great expect now you need 3x DACs and 3x Amps. However, you are still left with the speakers.

As someone who has done exactly this (all filtering is done digitally with phase linearization, the high frequency and low frequency are fed to different amps and then directly to the speaker elements), I can tell you that it has several significant benefits and I would recommend it to anyone with the patience and budget for it.  However, though there are many benefits, distortion reduction is not one of the main ones. Bi-amping does reduce certain kinds of intermodulation distortion in the amplifier, but that has nothing to do with crossover filters.

The speakers are the real "enemy" if lowering distortion is your goal, and the only solution is finding speakers with very low measured distortion (or building your own).  Electrostatic or magneplanar speakers have very, very low distortion, but are expensive and can be fragile.  I'm a fan of compression drivers for mid's and highs, but doing that right requires fairly large format speakers.

A secondary question, is there a standard for mic placement in studio recordings? For example, if a studio places their mics 6 feet apart and 3 feet from the artists but the home's speakers are 15 ft apart and 10 feet from the listener does this matter?

There's no long-term studio standard for mic placement, and in fact differing mic locations are used for effect often in the exact same recording (there might be 8 mics in the room, two in front of the singer, one inside the bass drum, etc., etc.) and they all get mixed down to stereo at a mastering table.  Realistically you should be asking if there's a standard speaker arrangement in a mastering studio (which I can't speak to, but I expect there's variation even there).

The ideal home speaker arrangement (assuming a neutral room) is an equilateral triangle with the speakers at two corners and the listener at the third with the speakers pointing more or less directly at the listener.  Real life rooms and listening conditions often defeat that kind of arrangement.  Second best is an isoceles triangle with the listener on the "long" point, and as much toe in as your living conditions/spouse will allow.

[glynor]

There's no long-term studio standard for mic placement, and in fact differing mic locations are used for effect often in the exact same recording (there might be 8 mics in the room, two in front of the singer, one inside the bass drum, etc., etc.) and they all get mixed down to stereo at a mastering table.

A great example of this is the drums at the beginning of March of the Pigs by nine inch nails. I read a fascinating interview with Reznor one time about how he went through a ton of different mic arrangements for the drums for that piece, trying to find one that gave him the sound quality he wanted.

In the end, he did it with a single mic on a mic stand set in front of the kit.

[mwillems]

A great example of this is the drums at the beginning of March of the Pigs by nine inch nails. I read a fascinating interview with Reznor one time about how he went through a ton of different mic arrangements for the drums for that piece, trying to find one that gave him the sound quality he wanted.

In the end, he did it with a single mic on a mic stand set in front of the kit.

Exactly.  There are a million stories like that, and my personal favorite is Joy Division's Unknown Pleasures. There was a short book written about the albums production, but among the more ridiculous things: they had the drummer playing on the roof with a separate mic to give the drums a completely different sound than the rest of the band (and to introduce a line delay), and had Ian curtis in the basement singing into a telephone and close micing the receiving telephone handset.

And then you've got Andrew Bird doing single mic recording for an entire ensemble to attempt to recreate the "old-time" sound of the 20's and 30's.  If you can do something ridiculous with mic placement, someone's done it (including taking an angle grinder to a live mic).

[JimH]

... (including taking an angle grinder to a live mic).

That's a job I could relate to.

[ErikN]

... a mid-range weights 20x? that of a tweeter

Here you've lost me.

I probably drifted way off the mark. My thought process was: A and B are simultaneously given current by the amp. B has 20x the mass of A. It seemed like it would take longer for B to start moving with respect to A.  I'm happy most of my question made some sense.

The speakers are the real "enemy" if lowering distortion is your goal

If not signal purity what is the goal?  Putting aside the perfectly reasonable 'listening enjoyment' or 'to see if it could be built', there must be some objective goal when designing and piecing together a setup?

[Bccc1]

If not signal purity what is the goal?  Putting aside the perfectly reasonable 'listening enjoyment' or 'to see if it could be built', there must be some objective goal when designing and piecing together a setup?

I'm with you, purity should be the goal, but for e.g. fans of tube amps it certainly isn't.
From my limited experience the goal is defined by the frequency response, the phase*, the group delay, distortion (not only harmonic distortion, but also intermodulation distortion) and the directivity. IMO the last point gets too little attention. A linear freqency response without constant directivity may sound ok outdoors, but indoors it can't be linear.


*normally frequency response and phase are tied together, but with FIR-Filters you can alter them independently.

[mwillems]

If not signal purity what is the goal?  Putting aside the perfectly reasonable 'listening enjoyment' or 'to see if it could be built', there must be some objective goal when designing and piecing together a setup?

I agree with you that signal purity is certainly one of the main goals of constructing an audio chain, but (as Bccc1 noted) not everyone necessarily agrees.  I'll make three related observations to better explain why I offered the caveat above:

1) There are other kinds of "audio purity" that aren't conventionally thought of as distortion.  What I meant by distortion above was total harmonic distortion (not necessarily other types of audio inaccuracies not commonly called distortion).  A great example is time decay (usually expressed in a waterfall plot).  A speaker may be able to reproduce a signal perfectly (i.e. with very, very low harmonic distortion, intermod, phase distortion, you name it), but after the initial impulse the speaker takes some time to stop making noise.  That kind of delayed spectral decay can cause music to sound smeared or indistinct, or if certain frequencies take much longer to decay than others it can make a speaker sound "boomy."  This is related to phase distortion (because both are in the time domain), but they are distinct issues and most people don't refer to time decay as "distortion."  Similarly, deviations or ripples in frequency response aren't customarily called distortion, but represent an inaccuracy in reproduction.  But if you take "distortion" broadly as including all deviation from the original signal however measured, certainly that's one of the most important system goals.  A better term for the overall bucket might be accuracy.  

2) Also absolute lowest distortion (in the broad sense of total accuracy) is never the only goal in setting up an audio chain even for folks who are objectively minded like you and I.  Other perfectly objective goals could include: 1) acceptable cost, 2) fits in my home, 3) plays loud enough for my needs, 4) durability, and 5) will not cause my wife to divorce me.  You can always think of another way to lower distortion (unless you're the proud owner of kit that holds the world record for lowest distortion), but ultimately, even objectively minded folks have to keep their other goals in mind.  I agree that in an unconstrained environment signal purity is the goal, but in the real world it will always be "as much signal purity as I can afford," etc.  Creating the type of system I was talking about above (large drivers with high sensitivity and high power handling) tends to be expensive in both dollars and space, so that may not be a realistic option for many folks.  Similarly Electrostatics and Magneplanars are awesome speakers, but are (comparatively) expensive, large, and power hungry, and electrostatics in particular can be quite fragile.

3) Lots of people don't agree with the premise that accuracy is the goal.  Euphonic distortion introduced by tube amps, by vinyl, or even (to a certain extent) by speakers with paper/cardboard cones have made many people into devotees of those respective technologies even though they are necessarily higher distortion than other alternatives. Additionally, on a personal note, even though I'm a fairly dyed in the wool objectivist, I personally don't prefer a perfectly flat frequency response across the band.  If I play back very high frequencies flat (above 10KHz), I find that it sounds very fatiguing and I eventually get headaches.  I have a sensitivity to high frequencies I guess.  But I deliberately introduce a roll-off up there (deliberately deviating from the accuracy my speakers are capable of delivering) because it sounds easier on the ears to me personally.  Although at least one fairly famous mastering engineer (Bob Katz) has expressed that he introduces a slight roll off in the high frequencies himself, so I may not be too far off the beaten path.

I probably drifted way off the mark. My thought process was: A and B are simultaneously given current by the amp. B has 20x the mass of A. It seemed like it would take longer for B to start moving with respect to A.  I'm happy most of my question made some sense.

You can measure how quickly speaker elements "ramp up" (it's called an impulse measurement).  It's a complicated area, but the short answer is that the heavier speaker has a much, much larger "motor," so will not necessarily be slower to move (although some will be somewhat slower, depending on the design).  This is actually one of the areas where bi-amping (doing the crossover before the amps, and not using a passive crossover) actually does help: when an amp is coupled directly to a speaker element it dramatically improves what's called the "damping factor" which is how quickly and effectively the amp can start and stop the motor.  Again bi-amping has lots of benefits for system accuracy, just not many for what's classically referred to as distortion (at least in the speakers, as noted above it will reduce intermod in amplifiers).  

More info if you're curious: http://sound.westhost.com/bi-amp.htm

[blgentry]

If not signal purity what is the goal?  Putting aside the perfectly reasonable 'listening enjoyment' or 'to see if it could be built', there must be some objective goal when designing and piecing together a setup?

A.  Enjoyment really is the goal.  As others have stated, this is really up the individual listener's tastes.  Though the rub here is, the individual listener may have NO IDEA what they like until they hear it.  Listeners can be sort of unreliable, even to themselves.

B.  Objective measurements:  The best way to state the goal of perfect reproduction is to say the exact same signal out of the speakers as came in to the input of the system.  A more technical way to say this is "perfect impulse response".  Impulse response is essentially a perfect "click".  Which seems like the wrong thing to measure right?  The reason impulse is the right thing is it contains all frequencies.  So it measures the whole bandwidth from lowest to highest in one signal.   But why would that be any kind of big deal when we can use other measurements and get essentially the same thing?  Because impulse response also encompasses *time*.  Your impulse can't be correct unless your time response is correct.  Time response is a BIG deal that gets very little attention.

One of the most talked about new DACs is the Schiit Yggdrasil.  It's designer says that it has correct phase response.  Phase being the inverse of time, this means that it has correct time response.  People are going nuts over this DAC.  Is it because it has perfect time response?  I dunno for sure, but I think it's correlated.

You've asked a lot of very good questions. 

Brian.

[mwillems]

One of the most talked about new DACs is the Schiit Yggdrasil.  It's designer says that it has correct phase response.  Phase being the inverse of time, this means that it has correct time response.  People are going nuts over this DAC.  Is it because it has perfect time response?  I dunno for sure, but I think it's correlated.

Careful, phase is only a snapshot in time and can't be equated to the entire time-response of a system.  Phase refers to the periodicity of the waveform (how delayed is this waveform at this point in time with respect to the source).  Phase does not refer to or describe other time-domain effects, like decay (how long does whatever is being measured take to stop making noise once it has started).  

The good news though is that impulses can be used to measure decay too.  You can get a lot out of an impulse (you're right about that)  

[glynor]

Also absolute lowest distortion (in the broad sense of total accuracy) is never the only goal in setting up an audio chain even for folks who are objectively minded like you and I.  Other perfectly objective goals could include: 1) acceptable cost, 2) fits in my home, 3) plays loud enough for my needs, 4) durability, and 5) will not cause my wife to divorce me.

All of those goals are, for me, far more important than accuracy. Especially that last one. 

[mwillems]

All of those goals are, for me, far more important than accuracy. Especially that last one.  

I was fortunate in that I had blueprints drawn, parts purchased, and substantial initial carpentry done on my pair of washing-machine/small refrigerator-sized speakers before I got married.  

My spouse is a very tolerant person, but I suspect that had I attempted to start the project even a year later, it probably would have been a much harder sell and I might've had to settle for mini-fridge sized speakers 

[mojave]

(1) What are the sources of your music? It seems like everything 'audiophile' boils down to minimizing error. It also seems correct to think that nothing gets better than the master copy. Do you find special studio recordings? Is it possible -- or worth it -- to capture 'high fidelity' recordings of live concerts? Are there 'audiophile' worthy recordings of mainstream pop / rock?

The best modern recordings I think are live recordings on concert Blu-ray discs. 

(2) I assume it's a given to use a lossless encoding. How do you select one over the other? They are lossless so it is not quality. I'd be surprised if one is 50% smaller than the others. Is it compatibility?

I chose Monkey's Audio (.ape) since it was developed by one of JRiver's developers, Matt Ashland. It was also one of the only options when I started ripping my collection.

(3) Now the crux. Where does JRiver fit in. You have a pristine recording, high sample rate, high bitrate, losslessly encoded. I assume this stays digital on an optical link to a high-end external DAC, hi-current amp, and speakers that can faithfully render God's laugh. In theory, a hands-off approach doesn't sound entirely wrong.

I go from JRiver to an external 12-channel 32-bit DAC to amps to speakers can can faithfully render God's whisper.  My center speaker is active (no passive crossover) and my left and right may soon be active.

(4) Do you have speakers that can actually render the difference between 48Khz, 96Khz, and 384KHz sampled audio? Given that we can send/recieve signals from the edge of our solar system I will concede the ADC, DAC, and amp as doable. Making something that can faithfully move air like a violin, voice, saxophone, and drum with the same, low error rate as a $10K DAC seems amazing.

The more displacement you have available (diameter x excursion) the less distortion and higher accuracy you have in reproducing the audible frequency range.

(5) In an answer to an unrelated question, one JRiver user noted having a 4TB audio library. On average, what is your encoded bit rate (i.e. if you ball-parked a 60 min album how much space would you want) -- assume your best quality stuff? How many albums/tracks do you manage within JRiver?

I am somewhere around 9000 audio files.

I probably drifted way off the mark. My thought process was: A and B are simultaneously given current by the amp. B has 20x the mass of A. It seemed like it would take longer for B to start moving with respect to A.  I'm happy most of my question made some sense.

The Thiele-Small parameter of a speaker driver that relates to mass is called Mms. However, it is insignificant to the motor strength of the speaker driver. Imagine adjusting the weight of the space shuttle by 10 lbs. Will it make a difference in acceleration? Less mass with a smaller motor can also accelerate the same as a larger mass with a larger motor. So, while B is larger than A, it will also have a larger motor.

[mojave]

One general way to reduce this kind of distortion is to use very sensitive speakers that also have high power handling so that the physical elements are moving as little as possible.

Sensitivity doesn't change how much the physical elements move. Any two speaker drivers of the same size must move the exact same distance to produce the same SPL at any given frequency. This can be modeled with winISD or other speaker modeling software.

Higher sensitivity and power handling reduce power compression.

To the OP:
There is a physical limit to how loud any speaker driver/tweeter can play. This limit is increased by adding more displacement. Also, the lower the frequencies, the more displacement is required. A good way to think of it is with musical instruments. The smaller the instrument, the less bass it is able to play - or its bandwidth is higher in frequency. When one transitions from a violin to a bass, the bass capability increases. The same is true of brass and woodwind instruments.



Since the various frequencies require different displacement, one usually breaks up the frequencies into different size drivers since, like instruments, they each can play better within a certain frequency range. Displacement in a speaker driver is measured by multiplying the diameter of the driver by its X-Max. X-Max is basically the parameter for how for in or out a driver can go while still sounding good. X-Mech is how far a driver can physically move, but its distortion can go way up.

One can increase displacement by increasing the size of drivers or by increasing the quantity of drivers.

[mwillems]

Sensitivity doesn't change how much the physical elements move. Any two speaker drivers of the same size must move the exact same distance to produce the same SPL at any given frequency. This can be modeled with winISD or other speaker modeling software.

This isn't strictly true of all speaker types.  It's true if you're comparing flush mounted pistonic speakers (typical cone speakers), but speakers with different principles of operation do not necessarily need to move the same distance to displace the same amount of air.  For example, that's the whole underlying principle of horns, and why horn-loading a driver can increase its effective sensitivity, or why you or I can shout louder through a horn.  The horn couples the driver to the air more efficiently, so the driver doesn't need to move as far to create the same air displacement, and the horn (of course) doesn't move at all. https://en.wikipedia.org/wiki/Horn_loudspeaker

That's how compression drivers can produce ridiculous sensitivities at 500Hz or 800Hz while moving tiny fractions of a millimeter. There are several other non-cone speaker-types that similarly can have different results because they move air differently and more (or less!) efficiently than a pistonic cone.  

But you're entirely correct though that sensitivity and power handling are not the best things to look for to lower distortion (I was writing fast, and should've stopped to think more carefully).  Thanks for the welcome correction    (I edited my post above with a strikethrough to make it clear that what I said was inaccurate).

OP If you're comparing two cone speakers, mojave is 100% correct, larger cone area (either through adding drivers or using larger drivers) is the best route to lower distortion.  FWIW, though, IMO the best way to lower the distortion of a cone speaker is to replace it with a horn-loaded compression driver at whichever frequencies you can.  I'm still dreaming of a bass horn one day

[blgentry]

All of this kind of seems to be pointing towards a question and answer of:  What kinds of speakers behave in certain ways.  What speakers sound different?  What speakers measure differently?  This is going to be way over simplified, but here's my little take on speaker types and performance:

A.  Speaker's time domain response is important.  So having time aligned drivers is a good start.  Having crossovers that don't mangle the phase response is also important.
B.  You can achieve the above with several different approaches.  The only passive designs I know of that are time correct are ones with minimum phase crossovers, which happen to be either first order, or some whacked out all pass compensation (which I've never seen done passively).  The last time I checked (more than 5 years ago), the only passive designs that were minimum phase were:

1.  Thiel.
2.  Vandersteen.  The 3A in particular gets high praise at it's price point.

There may be others, but those are the only ones I know that use passive crossovers.

C.  Planar speakers tend to have very good impulse response and therefor sound "fast", detailed, and "open".  They *should* also have proper phase response, but I haven't studied them enough to say.  Due to their low moving mass (very very low compared to cones) they tend to not "ring" (keep playing after the signal has stopped) and have low distortion because they move very little relative to the sound output.  They have giant diaphragms, which makes this possible.  *Most* planar speakers also tend to make something resembling a plane wave, which has a different in-room acoustic response than a smaller speaker, specific to the reflection from the ceiling and floor.  Finally, planar speakers are almost always true dipoles:  They make the same amount of sound from the front and the rear.  So they produce a "fake" reverb effect in your room, which can add to their illusion of a sonic image.  In very large rooms this disappears and that effect is negated.  Planars are interesting speakers.  It's worth noting that planar headphones are becoming increasingly popular for these same reasons.

D.  Horn loaded speakers have a lot of the same impulse response characteristics as planars.  They tend to not "ring" which can reveal details that you don't hear on a lot of other speakers.  It's just hard to find horn loaded speakers that go down below ~2000 Hz.  At least ones that are true horns.  They exist, but tend to be very large.  Like the Community M4.  A beast, but it's got nearly flat response from 300 to 3000Hz.  Horns are rather directional and can sound funny off axis.  Some people really like their "fast" detailed sound.  Some think they are biting or piercing.  I've never heard a horn system that I would call "mellow" or "relaxed" sounding.  But I've heard ones that I liked.  Just with a good bit of extra highs.

E.  Regular old cone speakers and domes.  These can sound fantastic.  But they are the "norm", as opposed to more special type speakers I have described above.  Some cone speakers are bipolar, meaning they have one or more rear facing speakers.  In *most* designs this just introduces "fake reverb" from your room and doesn't help with accurate reproduction.  I've personally heard one design that uses a small rear facing driver that was absolutely incredible.  Mind blowlingly good.  It's in the top 5 speakers I've ever heard.  It was an early version of the Von Schweikert VR4.  Albert claims to be using the "inverse transfer function of a microphone" in his speakers.  I've heard a LOT of speakers.  The VR4 was totally amazing.  I see he now only produces insanely priced speakers.  Too bad.  The VR4s were special at a rather reasonable (for high end) price.

As I said, this is necessarily a very non-inclusive summary.  It's just some things I know about different types of speakers and their relative sound and performance.  I'm almost certain to get some disagreement here.  Hopefully a bit of *agreement* too. 

Brian.

[ErikN]

I love it when a few questions spawn such thorough responses. I think I will be spending a while rereading the posts, following links, and recovering atrophied memories of ECE mathematics.

Also absolute lowest distortion (in the broad sense of total accuracy) is never the only goal in setting up an audio chain even for folks who are objectively minded like you and I.  Other perfectly objective goals could include: 1) acceptable cost, 2) fits in my home, 3) plays loud enough for my needs, 4) durability, and 5) will not cause my wife to divorce me.

I absolutely get this. With the exception of (3), these are pretty much universal. A variant of (4) that I like is "I can repair, tweak, and modify myself". I like programmable, modular, or large enough that it can be fixed with a soldering iron, hammer, etc. 'Not user serviceable' or 'Void if removed' stickers are my enemy.

Are there resources that explain complete end-to-end setups with an explanation the various design decisions and, maybe, even how 'data' was collected to guide the decisions?

[mwillems]

There are quite a few.  Two recommendations:

1) One of the most public and highest quality "end to end" system design setups I can recall is Siegfried Linkwitz's ORION system:

http://www.linkwitzlab.com/design_of_loudspeakers.htm
http://www.linkwitzlab.com/orion_challenge.htm

It's an idiosyncratic system (it's a dipole), but reading through the materials on linkwitz's site have been very educational for me even though I never intend to build a dipole speaker.  This is primarily because:

A) Linkwitz is one of the leading lights in the field of audio design (he is the co-inventor of the Linkwitz-Riley crossover, among other things).
B) He includes lots of measurements and explanatory material
C) And he documents the build well enough that many others have reproduced it.

2) Zaph Audio's builds:

You can find a dozen high quality end to end builds here with serious documentation and very detailed measurements:http://www.zaphaudio.com/

Additionally Zaph has measured dozens and dozens of common speakers currently in production and published the results, so you can have some independent confirmation of (among other things) the distortion performance of a speaker element you're considering.  As an example here's 2 and a half dozen 5.5 inch speakers that he's measured:
http://www.zaphaudio.com/5.5test/

Similar results exist for tweeters and several other sizes.

Another resource I've found to be very helpful:

Rod Elliot's westsound pages (not about specific setups, but lots of issue papers explaining various audio topics with measurements, and several design guides) http://sound.westhost.com/articles.htm

[mojave]

Another good place to view speaker designs and how they came about is to read the forums at diysoundgroup.com. Tuxedocivic and mtg90 are the main designers so you can browse through some of their threads. Sometimes half the documentation is on diysoundgroup and the other half on avsforum. The speaker and subwoofer kits for sale on diysoundgroup are an excellent way to get involved in speaker building (and get a superior product) without needing all the know how.

This isn't strictly true of all speaker types.

You are correct when comparing different speaker types to each other. I didn't even think of this even though all my surrounds use compression drivers.  

I'm still dreaming of a bass horn one day

Jeff Permanian of JTR Speakers is a speaker genius, IMO. He is the one bringing speakers and subwoofers for the JRiver room at Rocky Mountain Audiofest. His residential bass horn, the Orbit Shifter LFU, is one of the best subwoofers I've ever heard. Actually when I heard it, two were in operation.  His passive Orbit Shifter Pro was recently tested by Josh Ricci at data-bass.com. It has the highest output from 40Hz and up of any subwoofer tested with maximum levels exceeding 140 dB! The LFU version is tuned lower to 25 Hz and testing at data-bass.com is scheduled for sometime in October.

Most of his speakers are horn loaded with horn loading down to 640 Hz or lower. He is bringing five of the 210RT's for RMAF with horn loading down to 400 Hz.

[ferday]

I've never heard any boxed sub ever come even close to a well done infinite baffle setup...and of course the hardest to make happen.  I dream of the day, but WAF is actually below zero on that one

[JimH]

I'll be there some time on Friday.  I'm looking forward to it.

[mojave]

I've never heard any boxed sub ever come even close to a well done infinite baffle setup...and of course the hardest to make happen.  I dream of the day, but WAF is actually below zero on that one

I have an IB system with 8 Acoustic Elegance IB15 drivers in two manifolds. Two of the drivers in the manifold face in and two face out. In theory this is supposed to reduce THD, but that's not why I did it. I had to do it to make the manifold fits. This is still my favorite subwoofer system, but I've heard a lot of subs (ported, sealed, horn) that come close and I wouldn't mind owning. I will say the IB system has the most WAF of anything I've had in my room.

[mwillems]

I have an IB system with 8 Acoustic Elegance IB15 drivers in two manifolds. Two of the drivers in the manifold face in and two face out. In theory this is supposed to reduce THD, but that's not why I did it. I had to do it to make the manifold fits. This is still my favorite subwoofer system, but I've heard a lot of subs (ported, sealed, horn) that come close and I wouldn't mind owning. I will say the IB system has the most WAF of anything I've had in my room.

My ultimate unconstrained plan for the bass horn is to build it into a room wall, which (like your manifold) once complete will have pretty high WAF (low WAF during construction obviously).  Those folded horns from JTR look pretty awesome, but I feel like to get useful loading down to 10Hz in a home you almost have to build it in a wall/floor/ceiling.  That box is tuned to 25Hz and is already enormous, and you'd need more than twice the "length" to get loading down to 10 Hz.  I can dream though

[mojave]

I'll be there some time on Friday.  I'm looking forward to it.

You can check it out any time you like, but you can never leave! 

[JimH]

I was just kidding.   

[ferday]

I have an IB system with 8 Acoustic Elegance IB15 drivers in two manifolds. Two of the drivers in the manifold face in and two face out. In theory this is supposed to reduce THD, but that's not why I did it. I had to do it to make the manifold fits. This is still my favorite subwoofer system, but I've heard a lot of subs (ported, sealed, horn) that come close and I wouldn't mind owning. I will say the IB system has the most WAF of anything I've had in my room.

Your setup sounds delicious.  Pretty much exactly what I want.  My WAF problem relates to the room...I literally have nowhere to use as a baffle unless I build a large false wall and maybe cut into the ceiling...and the house is already very small.  One day

[KingSparta]

You know North Korea, and south Korea have the largest speaker banks so they can play propaganda across the border.

the speaker banks are huge, and the Loudspeakers can be heard 10 miles away.