[stv014]

Originally Posted by castleofargh /img/forum/go_quote.gif
 
but look at Tyll's measurements and how much the distortions values can change from a test signal at 90db to a test signal at 100db.
 I take what I often see with ortho, that getting louder seems to make the distortions go down. if I get 0.3% at Xkhz at 90db from a headphone, then what lvl of distortion am I to expect from a music signal at 40db on a quiet part of a song(quiet not silent)?

 
If you see the distortion decreasing at higher SPL on the InnerFidelity graphs, it is often because most of the "distortion" at the lower level is in fact noise. The measurements show THD+N, rather than THD, and they include all microphone and ambient noise in the audio band with no weighting. Especially when the 100 dB SPL distortion is about 3 times (+10 dB = ~3.16x gain) lower than the 90 dB one, and the latter looks noisy/"fuzzy" on the graph, the THD+N is likely to be more noise than distortion. It is also unnatural for low (mostly 2nd) order distortion to increase at low levels, such effect is typically seen with crossover or quantization distortion, which produce high order harmonics.
 
It would be useful if InnerFidelity has shown THD vs. frequency instead of, or in addition to THD+N, and the actual distortion spectra (FFT) at one or more combinations of frequency and SPL.

 

[castleofargh]

 

Originally Posted by castleofargh /img/forum/go_quote.gif
 
but look at Tyll's measurements and how much the distortions values can change from a test signal at 90db to a test signal at 100db.
 I take what I often see with ortho, that getting louder seems to make the distortions go down. if I get 0.3% at Xkhz at 90db from a headphone, then what lvl of distortion am I to expect from a music signal at 40db on a quiet part of a song(quiet not silent)?

 
If you see the distortion decreasing at higher SPL on the InnerFidelity graphs, it is often because most of the "distortion" at the lower level is in fact noise. The measurements show THD+N, rather than THD, and they include all microphone and ambient noise in the audio band with no weighting. Especially when the 100 dB SPL distortion is about 3 times (+10 dB = ~3.16x gain) lower than the 90 dB one, and the latter looks noisy/"fuzzy" on the graph, the THD+N is likely to be more noise than distortion. It is also unnatural for low (mostly 2nd) order distortion to increase at low levels, such effect is typically seen with crossover or quantization distortion, which produce high order harmonics.
 
It would be useful if InnerFidelity has shown THD vs. frequency instead of, or in addition to THD+N, and the actual distortion spectra (FFT) at one or more combinations of frequency and SPL.

I thought about noise, but you come across both situations, distortion values getting lower with increased SPL, and for other headphones the opposite.
I seemed to notice that the ones getting lower with SPL are often orthos. but even that might just be me trying to explain something with the wrong anecdotal observations?

 

[stv014]

 
I thought about noise, but you come across both situations, distortion values getting lower with increased SPL, and for other headphones the opposite.
I seemed to notice that the ones getting lower with SPL are often orthos. but even that might just be me trying to explain something with the wrong anecdotal observations?

 
I guess high end orthodynamic and electrostatic headphones have fairly low distortion, and are thus more likely to produce a noise dominated THD+N measurement. When higher distortion is shown at 100 dB SPL, and at least the higher level trace is a smooth (not noisy) line, then it is probably really distortion. Ambient noise (e.g. traffic on the street) could vary over time, and headphones have different levels of isolation, which adds some variation to the measurements. I guess low frequency noise is particularly hard to isolate, that is why I noted the fact that THD+N is normally unweighted.
 
Another, unrelated issue to consider when reading the distortion graphs is that they show a THD percentage, and are thus affected by the non-flat frequency response. For example, if there is a narrow but deep notch at some frequency, it will appear as a large distortion spike, because the fundamental is attenuated, but the noise and harmonics (if they are produced by the driver, while the notch is an acoustic effect of the enclosure) are not. On the other hand, peaks on the frequency response can also increase the distortion due to the higher SPL.

 

[Grave]

  You've misplaced a decimal point there. The threshold of audibility is 1%, not .01%.

Nope.

 

[GrindingThud]

Source? Study's have shown it more like 2-4% with music material and .7+/- with pure tones.
http://www.audioholics.com/room-acoustics/human-hearing-distortion-audibility-part-3
http://www.aes.org/e-lib/browse.cfm?elib=2476

Nope.

 

[castleofargh]

 

 
I thought about noise, but you come across both situations, distortion values getting lower with increased SPL, and for other headphones the opposite.
I seemed to notice that the ones getting lower with SPL are often orthos. but even that might just be me trying to explain something with the wrong anecdotal observations?

 
I guess high end orthodynamic and electrostatic headphones have fairly low distortion, and are thus more likely to produce a noise dominated THD+N measurement. When higher distortion is shown at 100 dB SPL, and at least the higher level trace is a smooth (not noisy) line, then it is probably really distortion. Ambient noise (e.g. traffic on the street) could vary over time, and headphones have different levels of isolation, which adds some variation to the measurements. I guess low frequency noise is particularly hard to isolate, that is why I noted the fact that THD+N is normally unweighted.
 
Another, unrelated issue to consider when reading the distortion graphs is that they show a THD percentage, and are thus affected by the non-flat frequency response. For example, if there is a narrow but deep notch at some frequency, it will appear as a large distortion spike, because the fundamental is attenuated, but the noise and harmonics (if they are produced by the driver, while the notch is an acoustic effect of the enclosure) are not. On the other hand, peaks on the frequency response can also increase the distortion due to the higher SPL.

so for you seeing the distortion going down from 90 to 100db test signal could be seen as a cue that distortion values are really low(and we're really looking at noise), and not that they actually behave better louder?
because I did find that some electrostats needed a certain volume to sound great. it's a totally subjective feeling, I realize that, but being usually the kind to turn down the volume on anything people pass to me, I kind of though electrostats were simply losing it with lower voltages. if such a thing actually exists.
anyway thank you for answering, I don't really know where to learn about this without going back to school and take audio engineering.

 

[stv014]

  so for you seeing the distortion going down from 90 to 100db test signal could be seen as a cue that distortion values are really low(and we're really looking at noise), and not that they actually behave better louder?

 
Well, some of them might really have lower distortion at higher levels, it could only be found out for sure if InnerFidelity provided more information (i.e. THD (not +N) vs. frequency and/or FFT plots). But distortion that is higher percentage at low signal level (for example, crossover or quantization distortion) is usually "bad", because it does not benefit as much from masking (it is more like noise), and it tends to produce high order harmonics.
 

  because I did find that some electrostats needed a certain volume to sound great.

 
Louder generally sounds better subjectively, at least within reasonable limits.

 

[castleofargh]

 

  so for you seeing the distortion going down from 90 to 100db test signal could be seen as a cue that distortion values are really low(and we're really looking at noise), and not that they actually behave better louder?

 
Well, some of them might really have lower distortion at higher levels, it could only be found out for sure if InnerFidelity provided more information (i.e. THD (not +N) vs. frequency and/or FFT plots). But distortion that is higher percentage at low signal level (for example, crossover or quantization distortion) is usually "bad", because it does not benefit as much from masking (it is more like noise), and it tends to produce high order harmonics.

interresting, thanks.
 
 
Quote:

 

  because I did find that some electrostats needed a certain volume to sound great.

 
Louder generally sounds better subjectively, at least within reasonable limits.

sure but as I said I tend to listen very very quietly to music, I'm hyper-something or just find loud sounds annoying, I don't know. I stopped going to live performances because I get annoyed by the loudness and it ruins my experience. and my custom filters are "ok-ish" but it's just not the same and not really neutral either. also it seems to really mess with my sens of positioning ^_^ I'm almost better with real plugs. anyway, just to say that me thinking louder is better tends to go the other way around very fast outside of my quiet comfort zone.
but maybe you're right and it's simply that the electrostats I've tried where a lot more flat in signature than other headphones, so that the overall loudness didn't have annoying pics? arrrrgh damn you subjectivity!

 

[bigshot]

I just rewired some stuff in my theater after getting a nice new blu-ray player and I realize that the way I wired it, I can't diddle with my equalization settings any more. The AV receiver doesn't show up on the screen any more. I guess that's OK. Most of the time my diddling is just to change things a dB or less. Not much point. It's good to let it ride now.
 
The thing I really need to do is go around the room and put earthquake putty underneath all the knick knacks and take all the 78s out of the Victrola. Every time I put on a horror movie, the whole room rattles like a bull in a china shop.

 

[Strangelove424]

  I just rewired some stuff in my theater after getting a nice new blu-ray player and I realize that the way I wired it, I can't diddle with my equalization settings any more. The AV receiver doesn't show up on the screen any more. I guess that's OK. Most of the time my diddling is just to change things a dB or less. Not much point. It's good to let it ride now.

 
Hm, that seems to happen to me to when using HDMI signal in from an HD source then out to the TV. I can't speak for all receivers but don't think mine can layer its text over HD or process the video signal at that resolution. The only time I see an on screen display is when the source HD is off (totally impractical, I know) and signal itself looks like 4:3 SD. Usually if I need to make db adjustments I use the little tiny window on the receiver and squint my way through it.   
 
Edit: Actually, nevermind, I seem to be getting the OSD while my Blu Ray player is plugged in... I wonder when that started working. Honestly, these things are too complicated for me sometimes.
 

The thing I really need to do is go around the room and put earthquake putty underneath all the knick knacks and take all the 78s out of the Victrola. Every time I put on a horror movie, the whole room rattles like a bull in a china shop.

 
hehe, reminds me of the "you shall not pass!" scene from Lord of the Rings. Always feels like my floor is lifting up. 

 

[castleofargh]

 

The thing I really need to do is go around the room and put earthquake putty underneath all the knick knacks and take all the 78s out of the Victrola. Every time I put on a horror movie, the whole room rattles like a bull in a china shop.

 
hehe, reminds me of the "you shall not pass!" scene from Lord of the Rings. Always feels like my floor is lifting up. 

lol, I was wondering why the hell you had been impressed by the sound of that scene
 
^_^ then I read again.

 

[Strangelove424]

LMAO. That scene in Holy Grail is classic, but when you juxtapose it against a typical movie scene like LoR, it becomes even funnier. 

 

[bigshot]

Have you noticed the "name the amp" contest on the HeadFi home page. I want to name that amp, "Accident Waiting To Happen". Every time I see that picture, I imagine the log rolling over and smashing the little lighthouse decoration on top.

 

[castleofargh]

  Have you noticed the "name the amp" contest on the HeadFi home page. I want to name that amp, "Accident Waiting To Happen". Every time I see that picture, I imagine the log rolling over and smashing the little lighthouse decoration on top.

I know how I would decorate it if I get one:

 

[SilverEars]

 
I know how I would decorate it if I get one:

Would be useful to make it glow in sequence for the holidays.  But noooo, we can't have that. 

  You pay $5k for tube amp, and get tired of it after a week, you should be able to put it to decorative use atleast.
.