Law of Accelerating Returns


Notwithstanding this coming from the pen of Robert Harley, I think there's a good point being made here. There are many threads here dealing with the law of diminishing returns. However, I think the way Harley puts it is perhaps more applicable to our hobby - the smaller the differences, the more important they are to those who care about such things. Read it - it's only one page.
http://www.theabsolutesound.com/articles/from-the-editor-the-law-of-accelerating-returns/
chayro
I've always thought of this phenomenon in abstract terms as an example of Xeno's Dichotomy paradox. As one approaches the (unattainable) goal of the Absolute sound, progress is experienced not as the distance travelled from the point of origin, but rather as a closing of the gap that remains. If one's system is 90% of what it can be, then a 5% progress to 95% is perceived as a 50% improvement-- closing the remaining gap by half. At 95% of goal, progress to 96% of goal is perceived as a 20% improvement in performance. This is why relatively small incremental improvements can be meaningful.
I am not sure Zeno would approve of this but if one's system is 90% then a 5% improvement from that starting point would put it at 90.5

Please check my math, it was never my strongest subject in school.
"10-04-14: Dgarretson
I've always thought of this phenomenon in abstract terms as an example of Xeno's Dichotomy paradox. As one approaches the (unattainable) goal of the Absolute sound, progress is experienced not as the distance travelled from the point of origin, but rather as a closing of the gap that remains. If one's system is 90% of what it can be, then a 5% progress to 95% is perceived as a 50% improvement-- closing the remaining gap by half. At 95% of goal, progress to 96% of goal is perceived as a 20% improvement in performance. This is why relatively small incremental improvements can be meaningful."

The only way you can assign percentages as a representative of performance is if you know what the end result is, and that it can be attained. If you don't know what 100% is, then 50% and 20% (or whatever), is meaningless.
Zd542,
I think that in this context, using percentages to measure differences is meaningless. I think that that was is meaningful is that small perceived improvements do seem very significant to the listener.
It's always been argued here by some on A'gon that small levels of improvement don't warrant the praise and accolades given. I've been on the side of giving credit where it's due, no matter the level of improvement, and to state it in glowing terms as it's the beauty of the results that justify the praise.

I like Harley's perspective.

All the best,
Nonoise
I don't live in the reality that audio reviewer live in. It's important to remember that TAS is a magazine that has taken to calling $15,000 power amplifiers and $18,000 loudspeakers "bargains". There was a time when the magazine didn't have to resort to obscure and convoluted reasoning to justify what its writers heard.
From the article:
The second reason why The Law of Accelerating Returns applies to an audio system is something Meridian Audio co-founder Bob Stuart calls The Increasing Importance of the Smaller Difference. He posits that humans are naturally inclined to make very fine discriminations, such as differences between two champion dogs of the same breed. Dog aficionados aren’t interested in the differences between dogs and cats, or between different breeds of dogs, or even between mediocre and stellar examples of the same breed, but they are fascinated by the very finest differences among the breed’s elite dogs, some of which are invisible to the untrained eye. The smaller the difference, the greater that difference’s importance to those who care about such things.
I think he's making a valid factual observation, but I also think that the observation does not support the stated conclusion.

I would put it that as reproduced sound more closely approaches the sound of live music, two things occur:

1)Small differences become more perceivable.

2)Perceived differences become more important.

I think that both of these points can be illustrated by considering the situation of two components or systems being compared while using a terrible sounding recording, and then while using a great sounding recording. With the first recording, Garbage A may sound different than Garbage B, but who cares? With the high quality recording, though, differences between the components or systems are likely to be both more perceivable and more important.

None of that means, however, that “the smaller the difference, the greater that difference’s importance to those who care about such things.”

Regards,
-- Al
The article begins with a false premise that the cost of audio equipment is directly related to its quality. Part of the joy is finding products that are fantastic values. At the lower end of audiophile equipment I think finding an objective sonic improvement that the vast majority can agree on is far more likely than with the extreme high end where I think it is more about preference and differences than objective undisputed differences.

I think he has a point to be considered, but is something that is more in parallel than at odds with the law of deminishing returns. They are not mutually exclusive ideas.
"10-04-14: Onhwy61
I don't live in the reality that audio reviewer live in. It's important to remember that TAS is a magazine that has taken to calling $15,000 power amplifiers and $18,000 loudspeakers "bargains". There was a time when the magazine didn't have to resort to obscure and convoluted reasoning to justify what its writers heard."

I don't see what the big deal is about that. The main purpose of the magazine is to review and discuss high end/expensive audio gear. Its to be expected that they review these type of components. As far as calling 15k amps and 18k speakers a bargain, why not consider taking it a face value? The comment may be valid. For example, I think the Vandersteen Model 5 is a very good value when compared to speakers that are far more expensive like Wilson and Avalon. Its all relative.
I think Mceljo is correct. I like the idea of looking and finding the best VALUES. I've tried cable which costs about as much as a small car and have had better results with much less expensive wires. That's just one example.

The often astronomical cost of high-end components can be rationalized and justified in many ways. After being even a small player in this game for many years, I'm convinced that past a point, as audiophiles pursue "accelerating returns" it's really about "STUFF". It is no longer about the music. So often, today's great stuff is tomorrow's trade-in.

There are no "laws" involved. if you spend x amount on a system and it sounds good, and then you spend x+ to "improve it" what law is at work besides the law of thinner wallets? It's seems ludicrous to me to try to mathematically quantify the level of improvement and somehow correlate it to the amount of money spend.

I suppose in theory, if you're system is "99%" there, and you get another .5% maybe you'd think that was a substantially accelerated ROI. :) Yeah, maybe.

I think a valid comparison can be made with high-performance car engines, something I know more than a little about. X amount of dough will yield X horsepower. To get maybe 5% more HP (no not THAT HP!) will cost exponentially more than one's initial investment. Is it worth it??? I guess a lot of that depends on how deep your pockets are.

Please, no snarky comments using the word "accelerating!" :)

Robert Harley seems like a very nice man but he does have a magazine to run. The never-ending quest for the "Absolute Sound" can help the economy but it can also make you crazy.

It's been said many times before, but think about how many concert tickets or disks a person could buy if he/she didn't buy that bazillion dollar next component.
Hey that's a pretty good argument for buying more audio stuff. Let me try it on my wife and report back.
Sounds_real_audio - the math depends on if the 5% improvement is relative to a perfect standard (100%) or the current level of performance (90%). The total in one case is 95% and in the other it is only 94.5%!
I get the notion that incremental improvement in one aspect of the system can improve the overall sound of the system (and reveal other weaknesses, shortcomings or colorations). But once a system reaches a certain level, getting those incremental improvements usually does cost more money. (Let's leave aside low and no cost tweaks and set up issues for a moment and assume all that has been addressed).
I know this sounds like $=improvement but the reality is, the better phono cartridges within a given brand, to take one example, tend to be the more expensive ones. Is a ten grand cartridge a worthwhile investment? Perhaps, in a system with other issues already sorted. Is it going to materially improve a system with other needs? Probably not as much- or put another way, probably not as much as addressing the other shortcomings first.
So, one brings the entire system up to a level that arguably has no pronounced 'weak' link; sorts 'mains' power, room, set up, etc. Gear is top level, whatever that is, for the sake of argument. Now add that top tier phono cartridge. Will it be a noticeable improvement? Maybe. If properly matched with arm, tonally compatible with the rest of the system and the listener's bias, etc. To get a system to a level where all that makes sense costs time, effort AND money. And it is at that point that somebody could reasonably say, 'it may be better, but it's not really worth it to me to extract that last '--nth' from my stereo. And i'm not spending 10k on a god damn cartridge, no matter how good it is!"
So, I'm having a hard time accepting the idea that these incremental improvements - investing more to extract that last iota- don't cost more money. I think the reality is, they do. I also don't find % characterizations about improvement to be very helpful. It lends a false sense of being able to quantify and compare differences that are largely unquantifiable, though real.
If the message is, pay attention to all the details in set-up, placement, room, mains power and the like before going out and buying the next piece of expensive gear (with the expectation that your system is going to drastically improve as a result), I'm totally on board with that.
Have certain components made a dramatic difference in my system? Yes. But, the system was already pretty dialed in and everything else was at a pretty high level of performance, so I knew where my 'weak link' was.
Where have I seen the biggest difference in sound at this point? The pressings. I'm satisfied enough with what my system does in my current room that I've been spending most of my time on sourcing the best pressings of the music I like. And there, you can find dramatic differences in sonic result. From flat and lifeless to vivid and punchy and alive sounding. Unfortunately, many of these are old pressings that are known as better sounding, and finding them in unmolested condition is neither easy nor cheap. I suppose someone could have the same reaction there, too- I'm just not going to bother buying that XXX pressing, even if it is better. No way is a record worth $XXX dollars. And I get that too.
I read Harley's essay and I wondered how high was up, and if the merry go round was going too fast to ever get off.

But then again, I've never spent the kind of money it would take to assemble a "world-class" system.

I also wonder if, when you introduce a "better" component into a system and the change is dramatic, revealing alleged "weaknesses," whether that could just as well be a matter of incompatibility.

There are simply too many factors and variables.

Here's a hypothetical. Your system finally sounds fantastic after years of changing this and upgrading that. You live with it for a few more years and gradually it doesn't sound so great any more. You're bugged. You start auditioning stuff, reading the mags and forums more closely, and then start rationalizing another major chunk of bux for whatever.

But at that point does the average audiophile get a hearing test? Nah.

I'd love to see some sort of at least semi-scientific data about the percentage of audiophiles who get to the point of just saying no to more "dramatic" changes. But then what's the fun in doing that? Who wants a hobby where you stop getting any new shiny stuff. :)
If you use live acoustic music as a reference point for 100% then it becomes reasonable to assign a percentage to mark progress toward that goal. It's admittedly a subjective measure. Percentages as a way to describe improvement have been common vernacular between audiophiles for a long time. Percentages are typically used to describe progress from the origin. But that approach doesn't account for the eye-popping improvement("accelerating returns") that is sometimes reported after making a small tweak near the finish line. The Zeno paradox is a thought experiment that helps account for this.
I guess I'm stuck because:
1.no reproduction system consistently sounds like real music; yes,
sometimes a convincing illusion on the right material, but listen to enough
different source material and you can hear the 'machinery at work'- the
illusion falls away, and you are left with a very good reproducing machine,
but one that cannot fool you all the time. I think that's the nature of the
beast. Gear is generally better today than it was 20 or 40 years ago, but I
can still hear the 'seams.'
2. we all (or better if I just speak for myself, "I") have different
ideas of what real should sound like. A little more bass. I wish that vocal
were a little more prominent. Why, when things get complex in the musical
program, does it sound congested? Some of this may be system-related,
and some, in my experience, is the source material. But, at bottom, I think
the ideal - DGarretson's 100%- may be different for each of us, and that is
going to dictate what each of us strives for, assuming an unlimited budget
and access to gear.
3. Most people don't have an unlimited budget and access to gear. We mix
and match using commercially available gear, with some tweaks, to try and
achieve that special 'synergy' that says 'real' to the person in control of the
system and the checkbook. This results in that endless merry-go-round of
equipment swapping, or the realization that last month's 'astounding
improvement' has paled by the time the credit card bill arrives.
4. DGarretson- it looks like you got off this merry-go-round by modding
commercially available equipment. I assume you did this for at least three
reasons: out of the box, the gear could benefit from an improvement that
the manufacturer didn't offer; cost-effectiveness, and subjective 'tuning' to
get the system to sound more 'real' by your lights. Most or at least many
audiophiles (and I'll include myself in this camp) aren't blessed with the skill
to start tearing down electronics and replacing the innards. So, we are left
with trying to find that elusive synergy among commercially available
'boxes' along with 'tweaking' in the easy ways- isolation, positioning, cable
changes, gear swaps, tube rolling, etc.
5. I guess my point is that the realization of that elusive 100% is a very
personal, subjective thing. And that once you are close, getting it to that
last 'nth' degree is neither easy or cheap.
6. For me, the midrange is always the killer. I can live with more limited
bandwidth or even limited dynamics, but if the midrange isn't grain-free,
vivid and 'in the room' it's going to sound 'reproduced' to me. Others may
conclude that bandwidth and soundstage are the markers for real.
7. So, where does that leave us? The 100% is at best an impossible goal. It
will never really sound like real music all the time. Do we give up? Most of
us here, are here because we aren't satisfied with 'good enough.' So, we
continue the 'quest' within the limits of time, energy, knowledge base,
access to gear and budget. But, and this is my hypothesis, not some
absolute statement, knowing your biases- recognizing what makes the
illusion better for you and where you can brook compromise- makes the
journey a whole lot easier. (And one's knowledge and one's preferences
may change over time too, which just adds to the length and difficulty of the
journey).
8. I have stepped off the merry-go-round, not because my system can't be
improved- it could benefit by better, deeper bass and a larger room to set a
bigger stage for the speakers (and probably by going to a larger set of
horns as well). But I am at a point where the differences in gear at this
point, for me, in my system and room, make less difference than the sound
of particular pressings and masterings. And, since the objective at the end
of all of this isn't just gear, but what the the musical result of the whole
system is, including the source material, I've found much happiness in
buying lots and lots of records, most of them old. This has, in turn, led to an
exploration of cleaning methods and their effectiveness, which is a whole
other subject.
I think Whart's reasoned and very well-written post says it all at least for me.
Where have I seen the biggest difference in sound at this point? The pressings.
Whart

Whart nailed it. And if you take a look at his system, it makes even more sense.
A "bargain" is more of an absolute description of worth as opposed to the relative term "a good value". A bargain also has a much wider time window than a good value.

I agree the the Vandersteen 5a is a good value, but it is not as good a value as the Quatro, nor does it represent the bargain value of the 2ce. The loudspeakers cover a ten-fold price span, but they don't represent a 10x increase in performance and I doubt the products generate a 10x increase in listener enjoyment. I speculate and I could be wrong.
When a setup is good I agree most of the difference from there comes from the recording unless perhaps you move into a better venue/room. That' s the real way to avoid those diminishing returns.
"I agree the the Vandersteen 5a is a good value, but it is not as good a value as the Quatro, nor does it represent the bargain value of the 2ce."

Why? How do you go about judging something like that?
If you double the performance, then double it again, how many percentage points is that?
Zd542, go listen to the lineup and then look at their pricelist. To me it's obvious, but as I said earlier, I could be wrong.