[SilverEars]

Has anybody seen graphs like that(headphone freq response scaling with different output voume levels)?

Please link them if it's out there. I'm wondering if the output scales proportionally or has variances in how different frequencies scales.

 

[castleofargh]

you mean a different frequency response at different levels? if so it doesn't happen on my gears(up to a point).
the lower limit for that is usually when I start measuring noise instead of the response because the output level is too low.
the upper limit consistently aligned with distortions becoming massive when the output is too high(clipping can look fun too).
in between was stable on the gears I tested back then. that's one of the first thing I wished to verify when I got my first coupler.

or do you just wish to see if voltage to dB follows a consistent variation? it does with the amps I've tried.

if you really want it I can make a FR graph for you with a few different voltages on IEM or headphone, but there won't be much to see.

 

[SilverEars]

I want to see snapshots of the sound level output across the spectrum as you change the volume level(turn the volume knob to change the volume level within range people typically listen to). If that causes proportional scaling across the spectrum(for the sound output level) or does it skew a bit?

Most importantly, this has to be measured from the output sound of the headphones.

 

[castleofargh]

ok worst job ever done in the universe of measurements. but I guess even that shows the main idea.
obviously the dB values are bollocks, REW is calibrated for my IEM measurements with the veritas and I didn't want to change it for this(sorry ^_^). here instead I used the IMM6 right into my laptop so that I had easy DC power and easy input boost. I measured one of the hd650 drivers poorly put against it. from voltage measured with a test tone, the loudest measurement was in reality around 92dB at 1khz(if that's of any interest to you, it sounded loud. louder than I would be comfortable with. all that to say, only the relative variations between each measurement are accurately giving the volume level changes at the amp.

the mess around 50hz is a parasite signal. only using the scarlett2i2 solves this for me, but for the couplers I own I need to add an external power(they're stuff for computers and cellphones so they need like 3V DC instead of 48V of phantom power). you can see the 50hz crap on almost all my headphone/IEM measurements, but I usually measure only at 90dB or 100dB, so it's usually just a little wiggling thing on the graph in proportion to the signal recorded. here it becomes a real monster when the input level goes down. and on the lowest levels you can clearly some low freq noises creeping in. if I measure 10 times I'd get 10 variations instead of the same graph, so it's definitly noise(the price of using cheap gears, the range I can use is more limited).
I could do the same with IEMs and make something cleaner if I put my mind to it, but the result would be the same. the signature doesn't change unless noises or distortions make it change.

and same thing but with the graphs stacked to better see changes or lack of.

 

[bigshot]

It's much more likely that your ears perceive sound differently at different volume levels.

 

[pinnahertz]

It's much more likely that your ears perceive sound differently at different volume levels.

Just a tad more than likely....

Equal Loudness Curves.


Why would anyone think that a transducer operating within its "linear" range would become nonlinear?

 

[RollsDownWindowsManually]

I did some measurements of a headphone I have at 100, 90, and 80 dB for fun. It doesn't look like it makes a difference in the frequency response, just like @castleofargh said. I didn't look at all the other parameters to see the effects.

 

[pinnahertz]

Think for a second about the kind of nonlinear mechanism that would be necessary to produce an audible change in FR vs level and not also produce horrific distortion levels. Such a device cannot be a mechanical transducer! It would have to be totally electronic and involve frequency selective gain-variable elements, with gain influenced by detected audio level, which also implies a time constant. No simple physical device is capable of that kind of dynamic signal processing. A related task is the true loudness meter, which involves critical band separation, level detection, squaring, summing, and square-rooting. Nothing mechanical can do even a part of that.

 

[bigshot]

A kazoo sounds different depending on how loud you toot into it.

 

[ev13wt]

A kazoo sounds different depending on how loud you toot into it.

So only toot at one volume.

 

[pinnahertz]

A kazoo sounds different depending on how loud you toot into it.

I never can really tell when someone is kidding, so, sorry if I don't have my humor plug-in working on this one.

Just as there's a difference between a kazoo and a headphone transducer (I can't believe I have to explain this!), there's a major difference between any transducer and an musical instrument/sound producer, starting with the simple concept that a perfect transducer doesn't distort the input signal, but a musical instrument pretty much always has some harmonic content, if not a lot, and deliberately. Every acoustic instrument varies its harmonic content with volume. A speaker/headphone/microphone/phono cart should not.

Transducer: a device that converts variations in a physical quantity, such as pressure or brightness, into an electrical signal, or vice versa.

The above means the musical instrument family, of which I suppose the kazoo is a member, is not a transducer.

 

[bigshot]

Always look in the humor in life. It makes it a lot more pleasant.

 

[pinnahertz]

Always look in the humor in life. It makes it a lot more pleasant.

Humor in life, yes. Humor in a potentially misleading post in Sound Science? Um....

 

[bigshot]

Well, you have to actually have a sense of humor to appreciate humor.

 

[pinnahertz]

Even with a sense of humor there has to actually be some there to appreciate.