How loud do you listen to your headphones?

[dc655321]

I use an SMSL iDEA primarily with my FLC8S in-ear monitors, and at home I now use a Topping NX4 DSD with a set of Massdrop HD6xx cans.
Prior to getting the NX4 a few weeks (months?) ago, the iDEA was used to power both earphones without complaint. IMO, the iDEA can drive the hd6xx louder than I like to listen.

I was and am still curious about how loudly I listen to music through my headphones. Am I consistent about volume as I switch between headphones and DACs?
Since I don't have a way to accurately measure the SPL I listen at (other than jamming my earphones into my iphone running an SPL app!), I thought perhaps I could glean some information by looking into the voltage ranges I set my sources to. Plus, I just got an Analog Discovery 2 to play with, so this gives me an excuse to kick the tires and maybe quantify some of the answers to my questions.

I wrote some software on top of the AD2's Python API to affect data acquisition. This was useful mainly to acquaint myself with some of the API's capabilities (and warts!). I used a 1 kHz tone into an open load through the iDEA and the NX4 in the following measurements. The AD2 was hanging off of my Thinkpad X1 (last year's model) running Linux, battery powered.

In the plots below, the dots represent dB as a function of volume position, while the lines are voltage as a function of volume position.

First up, is the SMSL iDEA. This has simple up/down buttons that discretely alter the volume in 20 steps:

I usually listen to the FLC8S at 5-6 clicks and the HD6xx at 14-16 clicks. In terms of voltages, that's about 0.020-0.028 Vrms for the FLC8S, and 0.260-0.450 Vrms for the HD6xx. Note the channel imbalance at minimal volumes here. Some of that is a 2-3 mV offset in the AD2. Clearly I need to do some calibration and try to eliminate that. Also note the different slopes (rate of volume change) before and after step #5.

Next up is the NX4 DSD in low gain setting. This unit has an analog pot for volume actuation.

The offset present in the AD2 cannot account for much of the channel imbalance visible here (and audible!) at low volumes. Note that Channel 1 and 2 appear almost "stuck" at some minimum value. Using the FLC8S with this DAC is not as seamless as I would like, as the channel imbalance does not resolve until well into my typical listening volumes. I can always turn down the volume in my player (Cmus - i'm a command-line junkie) to compensate for this, but of course I would rather not have to...
on
The issue is resolved just prior to 10 o'clock on the pot, but this is already at significant volume. Steps 7-9 for Channel 2 above correspond to 0.017-0.037 Vrms, bracketing the values I gravitate towards when using the iDEA.

I use a pot position of around 12 o'clock with the HD6xx on the NX4. Here, 12 o'clock on the pot is step #15, with a voltage of 0.450 Vrms. That compares well with my volume position on the iDEA. Of course, really great songs get cranked up a little bit more .

 

[RayDunzl]

Since I don't have a way to accurately measure the SPL

I set a level with the speakers in-room (measurable) and match that perceived level with the phones, just to get a good idea what the levels are.

In the plots below, the dots represent dB as a function of volume position, while the lines are voltage as a function of volume position.

Try using a log scale for the volts, too.

How loud do you listen to your headphones?

Not loud. Maybe 65-70dB.

 

[Dismayed]

My new amp goes to 11!

 

[amirm]

I usually listen to the FLC8S at 5-6 clicks and the HD6xx at 14-16 clicks. In terms of voltages, that's about 0.020-0.028 Vrms for the FLC8S, and 0.260-0.450 Vrms for the HD6xx.

Great measurements and timely data as I have been wondering what the average output voltage level people use for HD-650 type headphones. Like to focus more on performance in this region or at least mark it as such.

 

[svart-hvitt]

A health-related heads up from @Thomas Lund:

https://www.genelec.com/blog/dangerous-decibels

 

[dc655321]

Try using a log scale for the volts, too.

Shoulda, coulda, didn't. Thanks for the tip.
That would certainly make the low-volume voltage values easier to pick off the plots.
Perhaps I'll reformat the plots in an edited version.

 

[dc655321]

Cleaned up the plots.
Thanks, @RayDunzl for the suggestion.

Hmmm... I wonder what the output of my iphone se looks like?

 

[Timbo2]

One thing that just occurred to me is the usable volume is a consideration. I have albums from the 60s through the early 90s that I need a third more volume to listen to compared to albums created after the loudness wars.

I’d imagine your output voltage is roughly the same, but the “number of clicks” will be much higher on source material with a lower volume.

I use replaygain across albums generally, but not tracks so my volumes from album to album usually have significant differences.

Can I reiterate how much I hate what the loudness wars has done to music...

 

[JohnPM]

One potential concern for regular/prolonged headphone listening is a tendency to listen at levels 6 dB or more higher than with loudspeakers to achieve the same subjective loudness, especially for low frequencies. The original work on that was in the 30s, there is a more recent paper on it from Etymotic: https://www.etymotic.com/media/publications/erl-0137-1982.pdf

 

[tomelex]

Somewhat on topic, for me, its too damn easy to listen loudly to headphones, or say you hear one of your favorites so you turn up the volume but don't turn it down after that song.

There are two things I do. The best one IMO, is to start off by playing your headphones at a noticeable bit low of a volume, and your ears then seem to "increase" the volume on their own and you can be pretty secure in knowing you are not hurting your ears.

The other, is periodically lift the phones off your head and judge how loud they are that way.

The HD650 are a very sweet sounding headphone IMO, needing about a month of listening an hour or two a day to break in IIRC. I have never found them harsh and I suppose that is why they are an audiophile favorite. When I am doing a bit more analytical testing or design I use the hd 380 pro to give a more balanced frequency response (as much as you can say as far as headphones anyway)

 

[dc655321]

Somewhat on topic, for me, its too damn easy to listen loudly to headphones, or say you hear one of your favorites so you turn up the volume but don't turn it down after that song.

There are two things I do. The best one IMO, is to start off by playing your headphones at a noticeable bit low of a volume, and your ears then seem to "increase" the volume on their own and you can be pretty secure in knowing you are not hurting your ears.

Completely agree. Thanks for posting this.

In fact, one of the outcomes of this exercise has been that I've made a conscious effort to turn down my headphone listening.
On the iDEA with my FLC8S I'm down to 3-4 clicks, and I was just listening to my HD-6xx this afternoon at 12-13 clicks.
As you say, in no time at all I was acclimated to the new, lower volumes.

 

[dc655321]

I'm somewhat surprised, but my iPhone SE will output 1 Vrms:

There is not much headroom left in the HD-6xx range, but this is comparable to the iDEA and the NX4, above.

 

[Gus141]

I use an SMSL iDEA primarily with my FLC8S in-ear monitors, and at home I now use a Topping NX4 DSD with a set of Massdrop HD6xx cans.
Prior to getting the NX4 a few weeks (months?) ago, the iDEA was used to power both earphones without complaint. IMO, the iDEA can drive the hd6xx louder than I like to listen.

I was and am still curious about how loudly I listen to music through my headphones. Am I consistent about volume as I switch between headphones and DACs?
....
.

Your graphs say, “Volume Clicks” for the x-axis labels. I don’t have any physical “clicks” with the volume pot for the NX4. Do you? I was hoping I could just count clicks until 14 (where the channel imbalance disappears) and then add an iFi iEMatch set appropriately for the cans or IEMs in use. But no clicks on my NX4. You say, “I use a pot position of around 12 o'clock with the HD6xx on the NX4. Here, 12 o'clock on the pot is step #15...”. So that’s what I’m using (with good results): set 12 o’clock and use the ieMatch set appropriately for the cans or IEMs in use. thanks for your research and measurements.

 

[dc655321]

I don’t have any physical “clicks” with the volume pot for the NX4. Do you?

No, my nx4 does not have discrete "volume clicks". I was referring to the smsl iDEA which does have discrete volume steps.

set 12 o’clock and use the ieMatch set appropriately for the cans or IEMs in use.

I'm curious what headphones/iems you're using that might benefit from using the iEMatch with the nx4?

thanks for your research and measurements.

You're welcome!

 

[Gus141]

\
I'm curious what headphones/iems you're using that might benefit from using the iEMatch with the nx4?

\[

I’m a Sony-house-sound fan. So, I have Sony MDR-1A (24 Ohm), Sony MDR-1AM2 (16 Ohm), and Sony MDR-Z1R (64 Ohm). I use the iFi ieMatch with the 1A and 1AM2 Sony headphones; this results in a volume pot position (for my prefered listening level) of around 12 o’clock to 2 o’clock. However, for the Z1R, I don’t use the ieMatch. and can just set 12 o’clock or higher for my listening. As for IEMs, I have an ancient set of Shure’s (model number escapes me) that I bought in 2006, and I like them with the ieMatch.

Note: all ieMatch usage is with it set at “High” sensitivity. The “Ultra” sensitivity was too much attenuation for me.

Cheers

 

[soundwave76]

My advice, from a person, who got tinnitus from listening to headphones. Trust me, you do not want to get tinnitus.

1. Don't listen loud!
2. Take breaks from headphones listening. Having active speakers on the desktop is a good idea.

 

[Gus141]

My advice, from a person, who got tinnitus from listening to headphones. Trust me, you do not want to get tinnitus.

1. Don't listen loud!
2. Take breaks from headphones listening. Having active speakers on the desktop is a good idea.

Good advice. It’s why I like the iFi ieMatch (and Earbuddy) products.

 

[flipflop]

Take breaks from headphones listening. Having active speakers on the desktop is a good idea.

People listen to loudspeakers at higher volume than with headphones.
http://www.aes.org/e-lib/browse.cfm?elib=19252

 

[Dogen]

How does one determine how loudly one listens? It seems an app, if it knows what kind of headphones you’re using, could easily calculate that from the level of the music and the sensitivity of the cans.

SPL “feels” really different to me through speakers and headphones. I have a good sense for speaker levels but zero sense for headphones. I really would like to retain my remaining hearing.

 

[andreasmaaan]

In terms of voltages, that's about 0.020-0.028 Vrms for the FLC8S, and 0.260-0.450 Vrms for the HD6xx.

Great measurements and timely data as I have been wondering what the average output voltage level people use for HD-650 type headphones. Like to focus more on performance in this region or at least mark it as such.

Just to clarify @dc655321, wouldn't that volume setting correspond to 0.26-0.45V RMS for a 0dBfs sinewave? And would that not mean that most music signals would need significantly less output than that to get to your preferred listening levels? Perhaps I've misunderstood what you've done here.