Compare headphones - best practice

[Mate77]

I want to compare two headphones, how can I make the level exactly the same on both so that I am not affected by the difference in volume?

 

[gimmeheadroom]

You can either use the same volume setting or do a calculation based on impedance and efficiency. Although it might seem the calculation is more accurate, impedance and efficiency are dynamic. Using the same volume setting is the safest bet.

If you have a measurement rig shaped like a human head, and expensive mics, and metering equipment, you could measure the output of both headphones. But again, it's dynamic. If you use a sine wave it just tells you the level is the same at that frequency. If you use real music, you'll need to have some fancy averaging and calculations to go with it.

I think the best way to compare headphones is with a recording or several recordings you know very well, out of the same chain. After listening for a while you will probably like one better. The other thing is comfort which can't be measured.

I think most people on this forum are not into measurements but rather sound and other issues (build quality, materials, comfort etc.) There is an audio scientology forum but I don't see a lot of overlap between members of the two forums.

 

[castleofargh]

I want to compare two headphones, how can I make the level exactly the same on both so that I am not affected by the difference in volume?

You can't. If they don't have the same signature, then matching the output at one frequency means it won't be matched at another for both headphones.
You could try and see if it feels like you're close in perceived loudness with white or pink noise, that would work better for subjective impression than an objective approach at one given frequency. But then it's a rather vague matching by ear, so maybe don't obsess over whatever impressions you'll get that way? IDK. I tend to get a pretty good analysis of the sound in the first seconds of using new headphones. And after that, everything is chaos until after about a month in, when my impressions finally start to settle down and the "new toy" effect loses its power.

 

[VNandor]

Is there anything wrong with putting a microphone "inside" the earcup (without any coupling at all) to record the RMS of some sound and expecting the results to be usable?

I tried it quickly out of curiosity with some music and the recorded levels looked consistent to me. If I reduced the playback gain by 3, 6, 9 dB, the recorded level dropped by the same amount. I dismantled and reassambled this extremely intricate measuring arrangement a couple of times but the RMS only changed by 0,2dB at most between the measurements. The playback levels and the recorded levels were always consistent. Maybe I just got lucky.

I know that the measured RMS level is not the same that would hit the eardrum once the headphones are on the head but I just assume the lack of seal would impact open back headphones similarly in terms of volume and for something like this, you only need the levels to be correct relatively to each other.

I don't have more than 1 open back headphones lying around, that would be crazy, so I can't actually try if this is at least as good as matching the volume by ears.

 

[gimmeheadroom]

I think it would be less accurate than listening. For one thing let us not forget the ears, and not meters or computers are the final judge of what we like. Headphones are a tricky business, they depend on some degree of sealing, of refraction, of balance, of fit, of timbre. There are zillions of variables. Just sticking a mic, even a good mic in the airspace of a headphone is not likely to give any helpful info at all.

 

[castleofargh]

Is there anything wrong with putting a microphone "inside" the earcup (without any coupling at all) to record the RMS of some sound and expecting the results to be usable?

I tried it quickly out of curiosity with some music and the recorded levels looked consistent to me. If I reduced the playback gain by 3, 6, 9 dB, the recorded level dropped by the same amount. I dismantled and reassambled this extremely intricate measuring arrangement a couple of times but the RMS only changed by 0,2dB at most between the measurements. The playback levels and the recorded levels were always consistent. Maybe I just got lucky.

I know that the measured RMS level is not the same that would hit the eardrum once the headphones are on the head but I just assume the lack of seal would impact open back headphones similarly in terms of volume and for something like this, you only need the levels to be correct relatively to each other.

I don't have more than 1 open back headphones lying around, that would be crazy, so I can't actually try if this is at least as good as matching the volume by ears.

RMS levels are better than some peak measurements to give a sense of loudness. But they don't solve the matter of having different frequency responses. As we don't perceive all freqs as being as loud at the same SPL, getting RMS readings from 2 different frequency responses will probably not solve our issue. Of course if the headphones are pretty close in signature then any mean of measurement will do. RMS or not.

 

[VNandor]

RMS levels are better than some peak measurements to give a sense of loudness.

You are right of course. What about measuring the LUFS of the recorded signal instead of the RMS? LUFS has been developed to keep the actual perceived loudness of music in check.

Headphones are a tricky business, they depend on some degree of sealing, of refraction, of balance, of fit, of timbre.

This is a problem that can't be solved easily. I just question the magnitude of the impact. Open headphones don't rely on the seal as much as closed back headphones to begin with. If most open back headphones respond the same way to the loss of the seal then you practically don't need the seal to measure the relative volume levels. I just think most open back headphones would lose some bass by breaking the seal, lose some resonace at high frequencies and I imagine large parts of the frequency response wouldn't be touched at all. So some parts wouldn't change at all and some other parts would change in kind of the same way so it wouldn't matter at the end but I can't really substantiate this assumption. All I can say is that 1-2dB change at one octave band or two (so narrow parts of the overall frequency response) wouldn't change the overall loudness that much if the music contains 10 octaves of evenly spread frequencies.

 

[gimmeheadroom]

What I meant to say is the headphones are designed to fit on a human head and designed to seal somewhat uniformly- to whatever degree intended. And there are damping mods that make a big difference even on open headphones. Just think about how much time and effort has been spent on making the HD 800 into an HD 800s.

Sticking a mic into the cup will not help much because it doesn't replicate how loud the headphones sound under normal usage.

The OP should check Tyll's videos and articles on headphone measurements. But I still say using the same chain and not touching the volume, and actually wearing the headphones for a while listening to known recordings will beat any non-human method.

 

[playmusic]

There is an audio scientology forum [...]

@gimmeheadroom Maybe you consider to correct it unless it is on purpose and not a typo?

Concerning the OP:
@castleofargh is of course correct about the shortcomings of level matching.

One other method is to always turn the volume completely down (i.e. -100 dB) when switching between headphones.
You then turn up the volume as you like.
This is at least some kind of standardization for comparison.

 

[gimmeheadroom]

@gimmeheadroom Maybe you consider to correct it unless it is on purpose and not a typo?

Are you contributing anything or just trying to stir up problems for yourself?

Maybe you should consider minding your own business because nobody was talking to you.

 

[sander99]

There is an audio scientology forum

@gimmeheadroom Maybe you consider to correct it unless it is on purpose and not a typo?

Are you contributing anything or just trying to stir up problems for yourself?

Maybe you should consider minding your own business because nobody was talking to you.

I was also wondering if you typed "scientology" by mistake. Could be as I assume you are not a native English speaker?
You know scientology is the name of a church/cult with a questionable reputation, and not really a normal English word?

 

[playmusic]

Are you contributing anything or just trying to stir up problems for yourself?
Maybe you should consider minding your own business because nobody was talking to you.

In this science subforum it is about the scientific arguments and not about discriminating between worthy and unworthy forum members.

In your post above you brought another forum (which focuses more on measurements) in connection with a specific sect/cult.
Clarity is one aspect of science, so I have an interest in knowing what you meant by this connection.
And at least one other forum member was also wondering about it.

 

[SoundAndMotion]

Clarity is one aspect of science, so I have an interest in knowing what you meant by this connection.

He answered, but his post was deleted. I assume his name-calling violated the rules.

He said he knows the meaning of the word he used, and stands by it as being an appropriate reflection of his opinion, i.e. that forum has some cult-like followers.

He also attacked you and @sander99 for criticizing his English language skills (clearly not understanding what you were both saying). A delicious irony!

 

[sander99]

for criticizing his English language skills

Actually I didn't even mean to criticize. It was just an honest question I asked because I didn't understand what forum he was talking about.

 

[SoundAndMotion]

I know. That’s why I said

clearly not understanding what you were both saying