Determining safe level of headphone volume with umik-1

[Wendigo79]

I started to get worried about sound volume when I am using headphones. I tried to determine the safe level for music volume from headphones by using the Umik-1 microphone which is calibrated. Is this an OK way to do it? I just inserted the mic in to the ear cup of my HD650. My logic says it's the right way. What do you think about my settings on REW?

The first image shows my typical "loud" listening volume. For normal listening I got just above 60db. I am going to keep this laptop nearby and try sometimes later when I have been listening to music for longer period.

 

[Wendigo79]

Just realized, if I cover the cup with my hand so that the microphone is in the cup, I get much higher readings. I guess it simulates the situation when the phones are over ears

 

[abdo123]

I think you should just use the sensitivity value of the headphones.

 

[Aerith Gainsborough]

Just realized, if I cover the cup with my hand so that the microphone is in the cup, I get much higher readings. I guess it simulates the situation when the phones are over ears

Poke the head of the umik through a CD and lay it flat on top of the pads. I call this system "M.A.L.P. on a stick". :<

You get fairly repeatable readings in the bass area that way. Everything beyond 1KHz is nonsense though, so keep that in mind.

Pick a track with very loud bass and start measuring. It won't be super accurate but it'll give you a good impression of what you are subjecting your ears to. I calibrated my system to 90dB Z-weighted peak.

Average level is considerably lower so I remain in a safe area for long time listening.
Often I turn it down another 5-10dB when background listening.

I think you should just use the sensitivity value of the headphones.

Would be more accurate, yes but you'd need a voltmeter and a way to measure while playing. Not every amp tells you the current wattage via nice V/U meters. :'D

 

[abdo123]

Would be more accurate, yes but you'd need a voltmeter and a way to measure while playing. Not every amp tells you the current wattage via nice V/U meters. :'D

Another option is to check Amir's review and how much max power a device can provide at a particular impedance. you can work your way down from there using simple math.

My amplifier shows me current wattage / voltage so i didn't consider someone being completely in the blind.

 

[Aerith Gainsborough]

Another option is to check Amir's review and how much max power a device can provide at a particular impedance. you can work your way down from there using simple math.

My amplifier shows me current wattage / voltage so i didn't consider someone being completely in the blind.

You're assuming that the volume control of the DUT is linear, if you want to extrapolate that via math. Not sure if they are. Even then, from the pic it looks like some generic laptop, so fat chance that Amir tested that.

I think it's enough to poke the mic in there to get a guesstimate of the SPL. After all, completely exact figures are not needed in order to get the general picture. Whether it's +-3dB or +-1dB doesn't matter in this context.

 

[solderdude]

@Robbo99999 once posted a much easier way to determine how loud one listens.
This is more reliable than using a mic.

You just need to know which headphone, the gain of the amp and output voltage of your DAC and be able to use digital volume control calibrated in dB's

Robbo99999 can point you to his thread.

 

[Robbo99999]

Hello folks, @Wendigo79 , you could try to work it out mathematically by combining the sensitivity specs of your headphone and your max subjective listening levels.....I described how I did it at the following post, you could try it that way if you like:

*NEW* AKG K702 has been measured by Oratory!

Hi, like the title says, Oratory measured my K702 I sent him, I've got two pairs and I sent him my old pair from year 2015 (made in China, not Austria). Yes, so Oratory published the result today: https://www.dropbox.com/s/033kxyf8a0o3iv6/AKG K702.pdf?dl=0...

www.audiosciencereview.com

It should be more accurate than your UMIK experiments I think, and you don't need to rely on any kit to actually measure your headphones.

 

[DVDdoug]

If you have a multimeter, I'd make a "simple calibration" from your SPL measurements and the headphone specs. Except, all I found on the Sennheiser website was "SPL = 103dB". I think that's at 1V you'd need to confirm that. You can create a 1kHz test-tone with Audacity and then as long as you can get 1V you can simply subtract from your umik/REW reading to get the error/correction factor. (It's just addition/subtraction but technically it's a "factor" because decibels are logarithmic.)

You should measure the voltage with the headphones connected. A cheap multimeter is OK as long as it can read 1V reasonably-accurately. (I have a meter that's OK for low DC voltages but no-good for AC measurements below around 1V.)

 

[Aerith Gainsborough]

You just need to know which headphone, the gain of the amp and output voltage of your DAC and be able to use digital volume control calibrated in dB's

So let's see if I got it:
ADI-2 DAC Manual states +7dBu @ 0dBFS setting (Low Gain) = 1.73V (RMS)

Confirmed this site's first column.

dB dBu dBFS dBV to volts audio conversion digital - calculator volt to dBu and dBV dB mW SPL dB decibels 0 dBFS - convert dB volt normal decibels relatioship relation explanation analog audio absolute level true rms convertor converter decibel to dbf

dB dBu dBFS digital audio conversion dBV from dBu to volt, dBV to volt, volt to dBu and dBV dB decibels calculation SPL 0 dBFS normal decibels relationship relation explanation analog audio absolute level true rms convertor converter decibel to dbfs converter attenuation ratio reference convert...

www.sengpielaudio.com

At -18dB volume setting (my "calibrated" volume for 90dB Z Peaks), that would then be an output of -11dBu and 0.218V (RMS) per the above calculator.

My DT-880 has a "Nominal SPL" (whatever that's supposed to mean) of 96dB (at 1V? 1mW? Guessing game here, but lets roll with the 1mW)
Since I don't know the formula for putting this together myself >.<, I'll do the lamer gamer method and use the site below:

Headphone Power Calculator - Headphone & Earphone Audio & Reviews

How much power do your headphones need? Does your amp have enough power? Find out here with our handy calculator. Learn about voltage, current and power.

www.hear.audio

Which gives me 89dB SPL @ 0.22V RMS.
So I'd say the UMIK variant is pretty good, for a "seat of the pants" assessment.

 

[Wendigo79]

I will do try this method when I get back to it. Amp is JDS labs atom. And dac is topping E30 in dac mode, so always max volume on its end. So windows volume to max and then take only amp and phones in to the calculation.

 

[Wendigo79]

Great info about the topic, by the way!

 

[deafenears]

@Robbo99999 once posted a much easier way to determine how loud one listens.
This is more reliable than using a mic.

You just need to know which headphone, the gain of the amp and output voltage of your DAC and be able to use digital volume control calibrated in dB's

Robbo99999 can point you to his thread.

Is that more *reliable* though? Especially when we have measured HPs not matching spec, e.g. impedance of the HE6SEv2 measured here to be 64Ohm where the published spec is 50Ohm. Then also sensitivity numbers.

 

[Robbo99999]

Is that more *reliable* though? Especially when we have measured HPs not matching spec, e.g. impedance of the HE6SEv2 measured here to be 64Ohm where the published spec is 50Ohm. Then also sensitivity numbers.

Well, it might depend on how far the manufacturers specs are from reality, but that's the trade-off.....you're swapping the variable of bodging a UMIK measurement vs the accuracy of the manufacturer specs - take your pick as to which is the most accurate......try to research how much certain models of headphones vary vs manufacturer spec by seeing if there's any reliable resources on the net who have measured the headphone for sensitivity then see how that stacks up to the headphone model in question that you want to apply the calculations to. I'd use the calculation method because it's easier, mind you I have a miniDSP EARS now so I could measure it on there instead. Or do both the calculation method & the measurement method and see how close they are - then maybe pick the higher number if you're worried about how loud your music is and if that could damage your ears.

 

[Aerith Gainsborough]

mind you I have a miniDSP EARS now so I could measure it on there instead.

Have you ever thought of comparing an EARS measurement with a UMIK measurement?

 

[solderdude]

Is that more *reliable* though? Especially when we have measured HPs not matching spec, e.g. impedance of the HE6SEv2 measured here to be 64Ohm where the published spec is 50Ohm. Then also sensitivity numbers.

Manufacturers often publish incorrect specs. Best to use specs as published by independent measurebators.
Using a mic incorrectly can lead to much bigger errors though. The USB mic is calibrated but incorrect usage (given the attempts or incorrect usage of software) is a bigger potential problem.
Of course one can also be bothered by unintentional software gain settings.

At least, when done right, you don't need a meter nor microphone.

Downside is you only know what potential peak levels are. Determining average levels requires more info about the used music.

 

[Robbo99999]

Have you ever thought of comparing an EARS measurement with a UMIK measurement?

Not really, I have no reason or motivation to bother. Theoretical calculation is fine for me, and perhaps I could compare it against an EARS measurement if I'm interested (and I'd have to work out to be sure that I'm comparing like for like in terms of is this peak or average), but I don't want to go to the lengths of rigging up a UMIK bodge too.

 

[Aerith Gainsborough]

Downside is you only know what potential peak levels are. Determining average levels requires more info about the used music.

Could one use the crest factor for that?
If my music has a crest factor of 9.1 dB, I would still subject my ears to 80.9 dB on average. At that point I need to take exposure time into account in order to determine whether I am endangering my hearing.

Or should one use RMS directly?
Example track: -15.2 dB so 74.8 dB on average.

 

[solderdude]

Yes, you can use those kind of tools or use DR database when one doesn't know how to use such tools. One also needs to know how high peak levels are. Not all recordings reach 0dBFS.

A general rule (for people with rather 'normal' hearing) is that when you get the urge to dial down the volume after a few minutes, you are listening too loud.
Exposure time is important so listening to a song or part of it loudly now and then is not so problematic.
Aging ears will always only get worse. Even when one would be listening to very soft sounds all their life you can't preserve the hearing till old age anyway.
Abuse is never good.

 

[Robbo99999]

Some of you people might find this interesting or useful, I've measured a typical loud level that I listen at on my miniDSP EARS over a range of a few different songs (using my K702 with Harman EQ hugging Harman Curve down to about 35Hz), and I think it's louder than I expected I listen. I don't really know dB(A), dB(Z), but I chose dB(Z) because it seemed to be showing the highest numbers out of the available recording/logging options in REW. @solderdude I think you have a good understanding of the different dB types, what are we looking at in the following graphs & differences between dB(A) dB(Z), etc, could you help put some perspective to it above & beyond what I think I've worked out below?

Here are graphs showing how the dB(Z) levels changed through a few various tracks I would listen to. First one I'm gonna show you is one of my loudest tracks subjectively with lots of bass but also lots of everything else, it's Muse, Supermassive Blackhole:

LzPeak, the red line is representing the bass in the track I'm pretty sure - maxing out at 103dB(Z), that's quite a lot it seems, I always thought I didn't put this much energy into my headphones (hello distortion measurements??).

Now following is a graph of a track that is not a loud track, but it's got ungodly amounts of bass in it, and so far is the track that I have to make sure has the highest negative preamp to avoid digital clipping (bass EQ), this is Massive Attack, Angel:

Here you can see that the LzPeak red line is at a maximum of 104dB, so that's in the bass, whereas the rest of the music is not as loud as the Muse track (first graph).

And following just one more track for the hell of it, Massive Attack, Inertia Creeps:



I think I listen louder than I worked out theoretically, when I calculated it last year - either I listen louder nowadays or I've misinterpreted the information.

EDIT: or perhaps the LzPeak red line is not the bass and it might just be representing the 4kHz peak that the miniDSP EARS is renowned for in how it measures & displays it's frequency responses, see red circled in following graph? That peak in the measured frequency response is 17dB higher, so LzPeak might just be an artifact of that right rather than actually measuring the bass, I'm not sure? But whilst watching real time development of LzPeak it seemed to coincide with bass notes, so I'm still leaning towards LzPeak representing the bass during my measurements.