Would u care to record Temperature, Humidity and calibrate your multimeter?

Would you care to record Temperature, Humidity and calibrate your multimeter?



It is awesome that there have been tons of testing reports that could be a very helpful library on this planet for hi-end headphones.


However, I noticed a very disturbing fact that most of the reports have the same problem:

Room Temperature, Humidity and Atmospheric pressure are absent in your reports.


Since the speed of sound, performance of the headphones, performance of your microphones are all likely to be affected a lot by these factors.


I still remember my college reports got scored "0%" and "F" because of omitting the room temperature or forgetting to calibrate the multimeter.


I am wondering how come you could skip Room Temperature, Humidity, Atmospheric pressure and believe the existing reports are trustworthy??

Sorry to bother you with my blunt questions.



Please shed some light here.


Thank you very much.


ref:

http://www.montana.edu/rmaher/publications/maher_aac_0704.pdf

 

If you take a look at ebay, you might see that measurement microphones always come with a report bearing
Temperature:

 

This isn't an organic chemistry class or government work where data is taken for data's sake, headphone measurements help us understand how a headphone will sound but aren't the final word. Negligible factors such as the ones you mentioned aren't listed on any reports throughout the audio industry not out of laziness or oversight, but rather practicality. All measurements carry the caveat that measurements rigs / microphones vary unit to unit, so no rigs measure exactly the same anyway.

Feel free to do your own measurements that report the weather, position of the moon, and time of most recent bowel movement (for TP mods).

 

Room temperatures changes are mostly moderate. The difference in speed of sound between 25°C and 20°C is less than 1%. It is possible that we could see the effects of this in measurements, mainly room modes shifting in frequency in speaker measurements.
Humidity changes the speed of sound even less and air pressure doesn't change it at all.

Microphones operate very linearly at a higher temperature range than humans could ever live in. I think sensitivity changes less than 1db from like -50°C to +300°C with most microphones.


However, something that is of absolute importance is the microphone sensitivity calibration if we want to measure distortion. Ideally this should not be done on a headphone coupler, as coupler design changes the measurements way too much. Best would be to use an SPL meter and play back a tone with a loudspeaker at a fixed distance in an open space (to avoid ripples in the FR) and calibrate the sensitivity in software, so that they both show the same volume.
EDIT: Or just calculate it using the microphone sensitivity.

 

You meant by using an anechoic chamber and a calibrated standard noise box and a multimeter to calibrate the microphone ?

 

I meant holding a microphone in the near field of a speaker driver. Use a cheap SPL meter and then calibrate the software (adjust mic sensitivity in software) to get the same reading as the SPL meter. This may not be the most accurate but it should work ok.

Another way is to use the 0dbfs ADC input voltage and microphone sensitivity (given that there is no gain in the signal path). This should work quite well if the microphone sensitivity is given.

None of this will give you absolute accuracy, but I believe that achieving absolute accuracy simply isn't possible for most people (noone has access to an anechoic chamer and a calibrated standard noise box).

The more I think about it, the more we should discuss what's the best way to get an accurate db calibration. This is super important for distortion measurements and I suspect that most measurements taken at X db really aren't taken at that level, simply due to inaccurate calibration. The SNR simply varies too much with what are mostly the same mic capsules.

 

As you might have noticed:

http://www.hometheatershack.com/for...ction-values-all-new-cal-files-published.html


, capsules in/from noise level meters are only reliable from 100~5KHz.

 

I think some of the points DANIU was trying to touch could actually be valid. Not sure about all types of drivers but at least planar magnetic ones are quite temperature sensitive. In fact it is fairly easy to check, measure planar headphone at normal room temperature, then warm it up a bit with a hair dryer and remeasure (expect to see a couple of db's difference in a couple of places) or just give it a listen.

A good friend of mine who makes exceptionally good loudspeakers has measured thousands of ribbon-planar drivers throughout the years and would never settle on a twitter until it gets measured during different times of day within a few days period to minimize temperature effect.

Could these factors be important for a DIY measurement rig? Probably not but a proper headphone placement on a properly designed coupler using properly calibrated microphone in a consistently quiet environment etc all would be in the order of magnitude more important for sure.

 

I'd actually be kinda curious to see if temperature plays any non-insignificant role in the tension of the diaphragm.

 

This actually looks pretty good, as far as one can tell on a 8db/division scale that goes from 10Hz to 120Hz... After all, we're not trying to measure the FR with one of these capsules. You could just use a sine at a frequency that the SPL meter is accurate.
The most important thing is that the sensitivity matches exactly. Looking at the digital meters graphs it does look as if the sensitivity is off by up to 3db, which would be pretty bad. As long as they match within ± 0.5db, that would be pretty good.
EDIT: No, wait. Sensitivity does look pretty similar. Might buy one of these.

 

If memory serves me right, Tyll's burn-in measurements did indicate some changes in FR due to environmental conditions. All in all - for most headphones positioning will change FR more than reasonable temperature changes.

 

Hmm, doing some very very very gross generalizations...

using mylar coefficient of thermal expansion = 1.7e-5
using mylar modulus of elasticity = ~500 GPa
assume temperature change dT = 10°C (kinda reasonable between ambient and body temperature?)
assume diaphragm thickness of 2 micron, width and length of 10cm
resonant frequency is proportional to the squareroot of the tension... which I am going to totally bork with quasi one dimension mathemagic

I calculate a... 13% change in frequency?? Ugh that's off... maybe?

Assuming I kludged a factor of a hundred somewhere, rooting that out would still pull me into a 1% ish range.


Meh I dunno, just making crap numbers up as I go. I make no claims as to audibility or other assorted voodoo. Someone else go do some empirical testing with a guitar, a tuning app, and hair dryer.

 

Despite the caption to the OP's second illustration, and the illustration itself, I have no clue what effect changes in temperature are supposed to have on sound as perceived. If they just change the speed of sound in air, what does that do? change the frequency? I hear different notes in December than I do in July?

My innumerate take on the figures might have triggered an attack of audiophillia nervosa. I was beginning to shiver a little, but I made myself a hot drink and reassured myself that this was not speakers in a huge room that had to be thought about, but headphones, and the distance from the headphone driver to my eardrum is...

And no, even though it does get pretty damned hot and humid in that small space, I am not going to start researching tropical headphones.

But wait... I am but a recovering audiophile... all it takes is that one small seed of doubt.

Pass the DeOxit!

 

It might be easier to heat up the room by using a $15 heater fan from Home Depot instead of cooling down the air using an air-con.
You might also just open the door & windows and let in some cold air to save on power-grid / hydro bill while changing the room temp.

 

Tyll did a great job in providing the most needed datasheets for amateurs globally, kudos to him.

And he did not reveal record of Temperature and Humidity as far as people could see from the datasheets.
Regarding calibration, his AP had been serving him without calibration for a while as per his own articles.

With all due respect, I am seeking enlightenment here.

If we are looking for entertainment, why are we measuring?

If the basic rules and principles are negotiable and negligible then why shall we bother to measure?

Yes, we are living in the 21st century.

I hope we are not rendering examples of confirmation bias.


There have been too much cases of Confirmation bias in our human history.

fyi: https://en.wikipedia.org/wiki/Gregor_Mendel
in 1998 amounting to an accusation of experimental fraud by removing outliers, "tidying" datasets and repeating experiments.[47] Fisher asserted, that "the data of most, if not all, of the experiments have been falsified so as to agree closely with Mendel's expectations,"[44]and he called Mendel's result "abominable", "shocking",[48] and "cooked".[49]

Fisher accused Mendel's experiments as "biased strongly in the direction of agreement with expectation... to give the theory the benefit of doubt".[44] This has been often cited as an example of confirmation bias.[50]

https://en.wikipedia.org/wiki/Confirmation_bias

 

OK, so perhaps I should not be so quick to rubbish the idea. Perhaps it makes a difference. But I am disinclined to believe that it makes a difference where over the very small distances from headphone to eardrum.

If it makes a difference when listening to speakers, what would that difference be? A difference of pitch? Would that even matter? It doesn't, except to the very small number of people who have or suffer (as some would say) perfect pitch.

I am pretty sure that I don't care if A is 440Hz or not --- but A at 1760 (two ocatves up) is now already in my diminished hearing so I can talk about this stuff, but I'm not a good test subject,

 

That is just forcing the principle, expecting the students to work in government or something.

I calibrate multimeters (among other electrical/electronics equipment) for living and (50 +/- 20)% RH or 20 (+/- 5) deg C has never resulted in meaningful error of measurement in my experience, as other variables are much more intrusive, like even air movement or el/mag fields.
Multimeter from good brands like Fluke, Gossen, HP(Agilent, Keysight) etc are very, very unlikely to be out of spec in 10y of use.
If measurements here in SBAF had some significant legal meaning, sure a calibrated and certified bla-bla-bla system should be used for the data to be used as some sort of evidence.

As for mechanics and material properties, the temperature or humidity could have real impact. Many polymers soften up in room temperature range (think the usual chinese PVC power cord) or become brittle below certain point in temperature (polypropylene below few deg C). Some materials are porous, suck in moisture etc.

If speed of sound changes it changes more or less uniformly in all audible frequencies, I for one could not remember how middle c sounded in 20 deg C vs 10 deg C outside, maybe some pro musician could differentiate, haven't asked. Luckily our senses don't have very good reference for absolutes.

 

Shouldn't the frequency stay the same for reproduced sounds, as in with a headphone or speaker? Of course changes in speed of sound will change, say the pitch of a flute (which depends on the speed of sound), or other instruments (strings on guitars change tension with temperature, these changes will likely be much bigger than the change due to the different speeds of sound alone).