Audio Fidelity?
May 2, 2010 11:28 AM Subscribe
Is there an objective way to measure audio fidelity? If so, what is it? In other words, what quality of headphones and speakers makes the differences in sound quality?
A mintor correction to mikeand1: The best way to measure fidelity is via response curve. The best way to measure 'quality' is subjective listening.
posted by Jairus at 12:13 PM on May 2, 2010
posted by Jairus at 12:13 PM on May 2, 2010
Gosh. I am unsure if there is a way to truly objectively measure audio fidelity. To be sure, you can measure various parameters of both the equipment used to play audio on and the audio that is played, but there are many people, who, just can not "hear" the differences between same audio file at different qualities (bit-rates, method of reproduction, etc) and especially cannot hear the difference between the same audio played on different equipment of various types and "qualities". I certainly (within limits of course)hear differences in "audio fidelity" between equipment, methods of engineering (much more important to recognize now, post Oasis' first album, more people listening to music on the go and the race to the bottom with compression and other techniques used in engineering and mastering of music post recording), but I am a person who in their late thirties; By this I mean that I grew up in a musical household, where my parents took me to performances at the opera, jazz clubs and the like, where I played multiple musical instruments for years, and where I developed a desire to listen to music in a way that was as faithful to the live equivalent as possible. I teach middle schoolers, and I see people who have never experienced music outside of what they hear on their earbuds, grossly compressed so that they have nothing to fear from such bugbears as Dynamic Range; Something that exists in live music to the same degree that various notes in different keys do. This isn't to to say that growing up in such a musical (non-musical?) environment is an outright bad thing, I just say this to say that the ideas around "audio fidelity" are muddled, and fraught with opinions and outright delusions that make it almost as subjective a topic as "what is your favorite flavor of ice cream?"
Now, getting past all that, when looking at audio equipment personally I have always looked to equipment that removed as many barriers between myself and the music as possible. In my reference equipment, there aren't huge speakers or enormous amounts of power present. Rather, there is what is needed (in my opinion, mind you) to bring me as close to the music as possible. I won't get into X or Y product or how much you may or may not need to spend to achieve sonic nirvana *gulps when reminded about amount spent on home system, but it sounds lovely* but the only thing I can suggest is that you find music that you like (that you are extremely familiar with) that is recorded in a manner as faithful to the original recordings (sans heavy mixing and mastering) and take it to various higher end audio shops who (most importantly) are knowledgeable and not there for a quick sale, and listen to your music on lots of equipment, until you find the stuff that makes you smile the most, yet not send you to the poor house. Know also that there is a practical limit to how much better a thing will sound/vs cost. Often it is much lower than a person who has spent gobs of money will admit to. Good luck!
posted by chosemerveilleux at 12:21 PM on May 2, 2010 [1 favorite]
Now, getting past all that, when looking at audio equipment personally I have always looked to equipment that removed as many barriers between myself and the music as possible. In my reference equipment, there aren't huge speakers or enormous amounts of power present. Rather, there is what is needed (in my opinion, mind you) to bring me as close to the music as possible. I won't get into X or Y product or how much you may or may not need to spend to achieve sonic nirvana *gulps when reminded about amount spent on home system, but it sounds lovely* but the only thing I can suggest is that you find music that you like (that you are extremely familiar with) that is recorded in a manner as faithful to the original recordings (sans heavy mixing and mastering) and take it to various higher end audio shops who (most importantly) are knowledgeable and not there for a quick sale, and listen to your music on lots of equipment, until you find the stuff that makes you smile the most, yet not send you to the poor house. Know also that there is a practical limit to how much better a thing will sound/vs cost. Often it is much lower than a person who has spent gobs of money will admit to. Good luck!
posted by chosemerveilleux at 12:21 PM on May 2, 2010 [1 favorite]
Is there an objective way to measure audio fidelity?
Yes, if you define audio fidelity precisely enough. Coming up with such a definition that correlates with your subjective experience fidelity could be tricky though.
The response curve of a system is one measurement, but it certainly doesn't tell the whole story - transient response is another important measure that comes to mind.
posted by phrontist at 12:32 PM on May 2, 2010 [1 favorite]
Yes, if you define audio fidelity precisely enough. Coming up with such a definition that correlates with your subjective experience fidelity could be tricky though.
The response curve of a system is one measurement, but it certainly doesn't tell the whole story - transient response is another important measure that comes to mind.
posted by phrontist at 12:32 PM on May 2, 2010 [1 favorite]
If what you care about is fidelity - how close your output is to the original sound, than this is doable and totally measurable.
You compare the frequency response and harmonic distortion of a known setup, and compare that with another setup with the unknown element in the middle.
This tricky thing is that for a test of fidelity you are always testing 4 devices at once: an input transducer, a recording device, a playback device, and an output device.
A simple empirical way to "bootstrap" from a number of unknown devices to a good measurement system would be to test various combinations in recursive situations. For the recording and playback devices you go without input and output transducers and simply feed them into on another over and over - the pair which has the least change per iteration is the most accurate. You test the individual recorders and players by trying various pairings. Once you know the best recorder and player, you can use those to test a speaker / microphone pair - you do playback / recording iterations and the speaker / mic pair that shows the least change of signal in an acoustically dead room per iteration has the best fidelity. You try various pairings in order to evaluate the individual elements.
That is, of course, a thought experiment. In practice we have systems that we know are damn good, so we can compare an individual element swapped into that system. Realistically for top fidelity you would want a matched set of microphone, recorder, player, amp and speaker. Given the actual practical circumstances of recording and playback you match your playback system so it works with as much fidelity as possible with the recordings being made, and the people making the recordings match their systems so it sounds as good as possible on the sound systems people play on.
There are funny consequences of this arrangement: you end up with recordings where the producers expect you to play it as a cell phone ringtone or out of an ipod that will sound like crap on a system with good fidelity, and recordings where the producers expect it to be played on a system with low THD and a flat frequency response and it just sounds like crap when ripped to mp3 and played on an ipod. For pop music production really the goal is not "fidelity" to any original, but rather to sound as good as possible on the various low quality output devices that most consumers use.
posted by idiopath at 2:20 PM on May 2, 2010
You compare the frequency response and harmonic distortion of a known setup, and compare that with another setup with the unknown element in the middle.
This tricky thing is that for a test of fidelity you are always testing 4 devices at once: an input transducer, a recording device, a playback device, and an output device.
A simple empirical way to "bootstrap" from a number of unknown devices to a good measurement system would be to test various combinations in recursive situations. For the recording and playback devices you go without input and output transducers and simply feed them into on another over and over - the pair which has the least change per iteration is the most accurate. You test the individual recorders and players by trying various pairings. Once you know the best recorder and player, you can use those to test a speaker / microphone pair - you do playback / recording iterations and the speaker / mic pair that shows the least change of signal in an acoustically dead room per iteration has the best fidelity. You try various pairings in order to evaluate the individual elements.
That is, of course, a thought experiment. In practice we have systems that we know are damn good, so we can compare an individual element swapped into that system. Realistically for top fidelity you would want a matched set of microphone, recorder, player, amp and speaker. Given the actual practical circumstances of recording and playback you match your playback system so it works with as much fidelity as possible with the recordings being made, and the people making the recordings match their systems so it sounds as good as possible on the sound systems people play on.
There are funny consequences of this arrangement: you end up with recordings where the producers expect you to play it as a cell phone ringtone or out of an ipod that will sound like crap on a system with good fidelity, and recordings where the producers expect it to be played on a system with low THD and a flat frequency response and it just sounds like crap when ripped to mp3 and played on an ipod. For pop music production really the goal is not "fidelity" to any original, but rather to sound as good as possible on the various low quality output devices that most consumers use.
posted by idiopath at 2:20 PM on May 2, 2010
Stereophile does pretty rigorous analytical testing of many of the components they review. You can read and understand the ''measurement" section of some of the reviews on their website - that'll go a long way towards answering your question.
posted by kickingtheground at 3:21 PM on May 2, 2010
posted by kickingtheground at 3:21 PM on May 2, 2010
There are lots of objective measures, but the science of sound perception is crude at best. The available measures are very important and will get you most of the way there, but there remain items of equipment that test well and sound not up to par. You will need to actually listen to compare.
posted by caddis at 3:26 PM on May 2, 2010
posted by caddis at 3:26 PM on May 2, 2010
This thread is closed to new comments.
For speakers and headphones, you also want something approximating a flat frequency response curve for a wide range of frequencies (say 35 hz to 30,000 hz). But again, this will not guarantee that it sounds good.
Best way is subjective listening.
posted by mikeand1 at 11:39 AM on May 2, 2010