so when you say tolerance you mean that advertising a spec is useless when its not in relation to a different spec?
i ask about sensitivity because of my dt 770s. before getting them i was advised to get an amp aswell because 250 ohm is high impedance and i wouldnt be able to drive them straight off my laptop. but then i got the cans and not only would they work fine with my laptop, they sounded fine out of my phone too, no amp. i asked about this and was told its to do with sensitivity...
Sensitivity is less problematic than frequency range because we know it's usually measured at 1 kHz with 1 V or 1 mW input.
Frequency range however, without a tolerance range, is meaningless. Marketing usually boasts with unrealistic numbers. Some manufacturers even make jokes about it (RP-21 fantasy response). If we fixed the range to +/-3 dB most headphones would have a range smaller than 40 Hz - 16 kHz.
Quote:
are we talking about a glitch in recording or headphones?
It's an imperfection in headphones, but as I said it's nothing to be concerned about. A/D and D/A converters usually use linear phase filters so they delay all frequencies equally.
Sensitivity is less problematic than frequency range because we know it's usually measured at 1 kHz with 1 V or 1 mW input.
Frequency range however, without a tolerance range, is meaningless. Marketing usually boasts with unrealistic numbers. Some manufacturers even make jokes about it (RP-21 fantasy response). If we fixed the range to +/-3 dB most headphones would have a range smaller than 40 Hz - 16 kHz.
It's an imperfection in headphones, but as I said it's nothing to be concerned about. A/D and D/A converters usually use linear phase filters so they delay all frequencies equally.
Yeah, there are 12 half steps, 12 keys in an octave.
One octave is half the frequency of the next, so (approximately, or actually it depends on exactly how it's tuned because pianos are generally not tuned equal temperament) it should be a ratio of 2^(1/12) = 1.059463 between keys. Between x keys: 2^(x/12).
17 kHz to 20 kHz would be a little less than 3 keys. Of course, pianos don't go nearly that high.
well i cant say i really understood, i think i got lost somewhere as to what i was even asking... thank you for your replies though. once again, i think these are subjects that must be studied properly, because simplified explanations on forum threads dont seem to explain. i need a pen and paper to do math, and iv never studied music.
wow. that doesnt look so good... i wasnt talking about that though, i doubt i can hear more than the average. infact, my hearing is probably worse for wear after all the concerts and partys... non the less, i guess i wanted to know how to calculate according to specs if i need an amp (i understood its to do with impedance and sensitivity) but:
a. theres this thread i havnt gotten down to reading yet. http://www.head-fi.org/t/607282/headphone-amp-impedance-matching-basics-you-need-to-know
b. i already own a cmoy and a fiio e11. i was asking out of interest because of the surprise i got with my newly arrived dt 770s (they sound fine out of my phone without an amp aswell). i just wanted to know what sensitivity means and how it comes into the picture. perhaps i didnt explain myself too well...
again, thanks for your patience, also, i like your signature, is there a story to it?
It's not "in relation to a different spec" - it's more of a "naked numbers" problem. They provide incomplete data. I'm trying to think of an example that would make sense, but nothing is coming to me.
i ask about sensitivity because of my dt 770s. before getting them i was advised to get an amp aswell because 250 ohm is high impedance and i wouldnt be able to drive them straight off my laptop. but then i got the cans and not only would they work fine with my laptop, they sounded fine out of my phone too, no amp. i asked about this and was told its to do with sensitivity...
Sensitivity is (as xnor explained and as stated previously) a ratio of input power (which we can assume to be @ rated Znom because you're testing a given pair of cans, so it isn't arbitrary), to output SPL. So basically to expand it in another way, if you have headphones that are rated 90 dB/mW (which we'll assume is at 1khz), a 1 mW input signal should correlate to a 90 dB SPL output. 10mW to 100 dB, or twice as loud. And so on.
Impedance is related in that higher impedance correlates to higher voltage requirement, and vice versa (Current is inversely related). Many portable devices can't drive into high Z because they can't deliver enough voltage, but if the cans are sensitive enough, it may work. There's also a big misunderstanding of impedance (and ignorance of sensitivity), which lead to the invention of the "ohm rating" argument, where people will assume that the "ohm rating" is some sort of standardized magic that explains how "hard to drive" a headphone is, with a "higher ohm rating" meaning harder to drive. That is pure fantasy.
Frequency range however, without a tolerance range, is meaningless. Marketing usually boasts with unrealistic numbers. Some manufacturers even make jokes about it (RP-21 fantasy response). If we fixed the range to +/-3 dB most headphones would have a range smaller than 40 Hz - 16 kHz.
Just a generic question, has Tyll ever stated where his no-gurantee range is, or can we assume his graphs are trustable from end to end? I just ask because I know a lot of measurement suites will advise a huge grain of salt with any results over 10khz and under 50-100hz unless you have a VERY well controlled acoustic environment. But otherwise +1.
Just a generic question, has Tyll ever stated where his no-gurantee range is, or can we assume his graphs are trustable from end to end? I just ask because I know a lot of measurement suites will advise a huge grain of salt with any results over 10khz and under 50-100hz unless you have a VERY well controlled acoustic environment. But otherwise +1.
I don't think he has, but he's done multiple measurements with the same headphones which align fairly well so there's consistency.
If you look at the LCD2 graph you can see it's pretty flat down to 10 Hz. If there was some roll-off from the mic or dummy head we'd see it there, because Audeze also provides FR graphs showing a flat line down to 10 Hz.
The higher end is much more problematic. But most headphones show at least a drop of 10 dB at 20 kHz.
Even if we use the -10 dB as reference point for 20 kHz the headphone mentioned above is still over 15 dB down from that. Btw, I chose that headphone completely randomly. Most if not all manufacturers provide those "fantasy responses".
holy cow... did they use the same measurement dummy? Or do all dummies have cylindrical ear canals that all behave the same way? And also pretty much the same compensation and normalization?
Sensitivity is (as xnor explained and as stated previously) a ratio of input power (which we can assume to be @ rated Znom because you're testing a given pair of cans, so it isn't arbitrary), to output SPL. So basically to expand it in another way, if you have headphones that are rated 90 dB/mW (which we'll assume is at 1khz), a 1 mW input signal should correlate to a 90 dB SPL output. 10mW to 100 dB, or twice as loud. And so on.
so sensitivity, in dummie words, could just translate into gain? if its measuring the output spl (which basically means volume right?) relative to input of power, then sensitive headphones would have a higher gain than not so sensitive headphones, or am i getting this entirely wrong?
"Gain" implies some kind of value being multiplied. Maybe "efficient" is a better word, as in efficiency in converting electrical input power to output sound pressure level. Given a certain power input, the higher the sensitivity, the louder the output sound (as you say).
I don't think he has, but he's done multiple measurements with the same headphones which align fairly well so there's consistency.
If you look at the LCD2 graph you can see it's pretty flat down to 10 Hz. If there was some roll-off from the mic or dummy head we'd see it there, because Audeze also provides FR graphs showing a flat line down to 10 Hz.
The higher end is much more problematic. But most headphones show at least a drop of 10 dB at 20 kHz.
Even if we use the -10 dB as reference point for 20 kHz the headphone mentioned above is still over 15 dB down from that. Btw, I chose that headphone completely randomly. Most if not all manufacturers provide those "fantasy responses".
Oh yeah, I wasn't trying to dispute their accuracy or the point that most cans are over a cliff by 15khz, just more general outloud wondering. I know, having owned/heard a few cans that he's measured "flat" or "flat-ish" down to 20hz or 10hz, it isn't consistently the same level of bass response (at least, perceptually).
holy cow... did they use the same measurement dummy? Or do all dummies have cylindrical ear canals that all behave the same way? And also pretty much the same compensation and normalization?
It depends on the HATS or coupler - not all measurement equipment is created equally. Tyll uses a HATS, some labs or people use couplers (which are less sophisticated, I've heard speculation that this is probably how Beyerdynamic tests the DT48A for verification, because their in-house measurements are quite different from the graphs Tyll put up; just for example).
Ideally if they're using the same compensation and same targets, with standardized measurement equipment, the end-result should be similar (in other words, the compensation should be appropriate to the mics and couplers they're using, to produce an FR that is at least similar to what a real human ear will hear; there's actually some ITU standards for how to do headphone measurements "correctly" - which I don't think anyone follows).
Standardization in measurement is a beautiful, but rare, thing.
so sensitivity, in dummie words, could just translate into gain? if its measuring the output spl (which basically means volume right?) relative to input of power, then sensitive headphones would have a higher gain than not so sensitive headphones, or am i getting this entirely wrong?
No. Gain and sensitivity are in no-way related. Gain is an unrelated topic. Output SPL = acoustic intensity, "volume" is too inaccurate a word to describe it (intensity is really the right word, but a lot of people use "loudness" as well (loudness refers to perception, which is not measured by dB SPL, it's measured with the Phon)). Don't try to conflate or over-simplify topics - it will often confuse you further.
Sensitivity is just the ratio of input power to output SPL. To put some relevance to that, THX specifies a "reference level" at 85 dB (which is what THX certified theaters are supposed to approximate, plus 20 dB of head-room (for peaks of up to 105 dB)), OSHA and CCOHS specify 85 dB as the onset intensity for hearing damage (NIHL), citing 8 hours of exposure as the maximum daily allowance. Generally I think most people are probably listening at something like 60-70 dB, at least I hope they are (85 dB is pretty loud, and does get fatiguing after a while).
Maybe these will help you make sense of it:
http://www.ccohs.ca/oshanswers/phys_agents/noise_auditory.html
http://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=standards&p_id=9735
http://www.thx.com/consumer/thx-technology/thx-reference-level/
This isn't quite the same thing as gain, which implies increase to the signal's intensity. Sensitivity isn't talking about increasing intensity, it's dealing with the relationship of input power to output intensity.
"Gain" implies some kind of value being multiplied. Maybe "efficient" is a better word, as in efficiency in converting electrical input power to output sound pressure level. Given a certain power input, the higher the sensitivity, the louder the output sound (as you say).
Efficient is a bad word too, because driver efficiency is not directly related to sensitivity. It deals with power transmission and loss - most dynamic drivers are *very* inefficient.
http://en.wikipedia.org/wiki/Loudspeaker#Efficiency_vs._sensitivity
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
