Head-Fi.org › Forums › Equipment Forums › Sound Science › Learning more about the science of sound
New Posts  All Forums:Forum Nav:

Learning more about the science of sound - Page 3

post #31 of 395

Would a objective graphing of audio players' performances be the best determinant of how good it is unaffected by perception and placebo bias?

I read somewhere that a clip+ graphs better than its x1000 price counterpart the Hifiman 801. 

 

For clarification by 'good' I mean the accuracy of reproduction.

post #32 of 395
Quote:
Originally Posted by uchihaitachi View Post

Would a objective graphing of audio players' performances be the best determinant of how good it is unaffected by perception and placebo bias?

I read somewhere that a clip+ graphs better than its x1000 price counterpart the Hifiman 801. 

 

For clarification by 'good' I mean the accuracy of reproduction.

 

Maybe, depends what you're really asking.

 

You can graph objective data that doesn't really have anything to do with perceived audio quality.  There are a lot of potential audio benchmarks you can run.  How do you define accuracy of reproduction?  Something that scores the best across a range of benchmarks (which ones?  how to quantify the extent of deviation from ideal?  how to weight different problems?)?

 

The benchmark results can be referenced against some kinds of previous psychoacoustic research on what matters for human audio perception, but that's never 100% clean and straightforward.

 

Maybe it's the best method, maybe not.

 

In my opinion the best thing would be to test peoples' subjective responses to the devices, controlling for and randomizing out the different perceptual biases, order effects, and so on.  Results are still affected by those things, but hopefully there wouldn't be a net bias towards one device over another.  Of course, that kind of testing requires a whole lot more effort and resources than just doing electrical bench measurements, especially if you want a reasonable sample size for making statistical inferences.

post #33 of 395

If I am not wrong, isn't it fairly straightforward to achieve the best audio reproduction if noise and distortion is low (below human hearing levels) and all the frequencies are present without rolling off. I am asking the question under the assumption that the subject is in a double blind test that they do not know whether they are listening to a hifiman or a sansa clip +.

 

I see no point in running a test to gauge peoples' subjective responses. You can't make any statistical inferences from such a test except for the fact that 99 percent if not every subject will obviously hear a drastic improvement in SQ from an expensive source with a 'wow' factor.

 

I am curious if it is indeed so cheap to produce a source device with optimal audio reproduction (where distortion is low and all the frequencies present etc which is what graphs tell you), if the perceived difference in SQ between devices can be explained by added colorations through EQing or other means. This then makes me wonder, with correct modifications, you could fine tune a clip + to sound like a hifiman?


Edited by uchihaitachi - 3/15/13 at 7:15pm
post #34 of 395

Subjective responses with all those factors controlled, as mentioned before (meaning double blind, nobody knows what they're listening to, and more).  I should have said "responses to the sound" rather than just "responses"; I guess it wasn't clear from context.  If I understand what you're asking, our interest is in just the sound produced, so any reasonable test would obviously control for the other factors.  Or maybe it's not so obvious, if you look at the average response in an audio forum.  

 

 

So what happens if one device has lower noise but higher distortion than something else, or something else like that?  How do you rank which is better?  Depends on the level of noise, level of distortion, type of distortion, etc.  It's not necessarily clear-cut, especially if the parameters are above levels that most experts would agree are audible.

 

Now, even for the subjective evaluation it would depend on the parameters of the test—listening volume, headphones used, and so on.

post #35 of 395

You could certainly gimp the clip to output amounts of noise and distortion comparable to the hifiman. It could be a gloriously pointless appropriation of rockbox, a third-party firmware for the player. An oft-cited experiment in the 80s had an engineer modifying a moderately priced solid state amplifier to perform exactly like an overpriced tube amp. This was done by fiddling with the ss amp until it's output would produce a (iirc) 70db null with the comparison amp. The result was successful in fooling even diehard skeptical audiophiles and so was roundly ignored in the subsequent decades.

 

If you look at the spec sheets for DACs, opamps, etc., they often include a turnkey design suggestion which provides for cheap solutions with performance well below the established hearing thresholds for distortion/bandwidth. Whether that's all there is to it is a different question. I think it's fairly well established that even the most reasonably priced components will produce perceptually neutral output in a competent implementation of a competent design. This is why it's fun to point to overpriced monstrosities like the hifiman by comparing them to entry level mass market products. Though, in all fairness, the difference would not be audible in most cases (and in cases where it would be, the difference wouldn't be caused by a, say 80db snr vs. a 90db snr).

post #36 of 395

Oh thanks for the clarification!

 

If a source has the lowest possible distortion and noise levels in addition to having all the audible frequencies present, would this not be considered the optimal audio reproduction by professionals?

 

But to listeners, certain levels of distortion and coloration would sound more pleasing. Like how they purposefully coloured tube amps etc.

 

Could this 'colouring' process be accomplished by simple EQing?


Edited by uchihaitachi - 3/15/13 at 7:40pm
post #37 of 395

ALso burn in is it a myth?

 

Sorry for the double post :S

post #38 of 395

There are several threads on this in this forum, but the bottom line is that we can measure very small pre/post break-in differences. The "rest" of the reported differences (or let's say almost all tongue.gif) seem to be "imaginary".


Edited by xnor - 3/20/13 at 1:23pm
post #39 of 395
The first 30 mins of unpacking the hd800s was awful. Once that had gone they started working much better. There is a school of thought out there that this phenomenon is due to their just getting cold and not being used for a long period.
Whether they need 30 mins or 30 days or 30 years is impossible to prove hence all the urban myths out there on the subject
Trev
post #40 of 395
Quote:
Originally Posted by uchihaitachi View Post

Oh thanks for the clarification!

 

If a source has the lowest possible distortion and noise levels in addition to having all the audible frequencies present, would this not be considered the optimal audio reproduction by professionals?

 

But to listeners, certain levels of distortion and coloration would sound more pleasing. Like how they purposefully coloured tube amps etc.

 

Could this 'colouring' process be accomplished by simple EQing?


You seem to believe in the 'tube sound' without knowing why. The truth of the matter is, there has yet to be done any double blind testing where a tube amp was correctly identified versus other solid state amps for the purpose of hifi. There exists an idea of the tube sound, but so far there are no scientific results supporting it.

 

Tube amps tend to distort more than modern solid states, yes, but even so, said distortion is still inaudible and thus the idea of tubes coloring the sound with their distortion has no basis in reality.

Coloration in terms of EQ'ing can not be generally pleasing because the results always depend on the headphone used. At best there may be some coincidental result where an emphasis on lower frequencies just so happens to be found enjoyable by some listener(s). Even if that were the case, it would be easy to simulate by just using an equalizer.

For what it's worth, most modern tube amps (like the valhalla, little dot 3 etc) aim for flat frequency response just as much as SS amps tend to do.

post #41 of 395
Quote:
Originally Posted by uchihaitachi View Post

If I am not wrong, isn't it fairly straightforward to achieve the best audio reproduction if noise and distortion is low (below human hearing levels) and all the frequencies are present without rolling off. I am asking the question under the assumption that the subject is in a double blind test that they do not know whether they are listening to a hifiman or a sansa clip +.

 

I see no point in running a test to gauge peoples' subjective responses. You can't make any statistical inferences from such a test except for the fact that 99 percent if not every subject will obviously hear a drastic improvement in SQ from an expensive source with a 'wow' factor.

 

I am curious if it is indeed so cheap to produce a source device with optimal audio reproduction (where distortion is low and all the frequencies present etc which is what graphs tell you), if the perceived difference in SQ between devices can be explained by added colorations through EQing or other means. This then makes me wonder, with correct modifications, you could fine tune a clip + to sound like a hifiman?


The last part depends. If the clip measures great and the hifiman measures badly, it will be hard to 'worsen' the clip's measurements to the hifiman's level (depending on which measurements we are talking about). The question is why you would want to ofcourse, and the answer is that you just might happen to prefer the hifiman even if it measures worse.

The thing is that the science of amping implies that in general, the best results are found if all measurements are as dictated. That is to say, the better the measurements, the better it's generally believed to make your headphones sound. Better measurements are analogous with real life phenomenons such as 'seeing better', 'hearing better', 'running faster', etc; that is, better measurements lead to superior results that are, basically without exception, experienced as unambiguous improvement. 

Your opinion may deviate from that but in that case, you'd be just as lucky to find whatever it is you're looking for in a ridiculously expensive tube amp as in your onboard audio. That is to say, a person might prefer to see less or hear less in a situation where it's generally found preferable to see more and hear more, just because they have that option (the reason isn't really important).


Edited by SunshineReggae - 3/24/13 at 1:24pm
post #42 of 395

So one could conclude that tube amps and other ridiculously expensive gear (that measures same as cheap gear) are for the design and look aspect and that it 'enhances' your listening experience from the way it is in a visual aesthetic manner and of course its price tag.

post #43 of 395
Quote:
Originally Posted by uchihaitachi View Post

So one could conclude that tube amps and other ridiculously expensive gear (that measures same as cheap gear) are for the design and look aspect and that it 'enhances' your listening experience from the way it is in a visual aesthetic manner and of course its price tag.


Gear can be expensive for a multitude of reasons. The ability to provide great measurements when driving very demanding transducers, build quality, versatility, customer service, etc. The trick is to find the right amplifier for your needs. I find it's generally best to keep conclusions to yourself and avoid generalisations etc, as to prevent causing damage to people or companies where you didn't mean to (although I understand what you were getting at).

post #44 of 395

But I also believe that sharing your conclusions with others helps. I mean many companies do seem to feed off placebo effects of others...

post #45 of 395

But how can you be sure when a company is feeding off placebo effects, rather than making amps to the best of their ability?

 

Anyway, it's just not a discussion I really want to get into. It's not personal ofcourse - but it's something that might deserve another topic :).

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Learning more about the science of sound