Headphone measurement stuff

The place to sound off on all things related to audio, from sound cards to speakers.

Moderator: Captain Ned

Headphone measurement stuff

Postposted on Mon Jul 21, 2014 10:30 pm

(This post got me thinking about this again - it was already a bit of a tangent in that thread, so I won't derail that one.)

superjawes wrote:If you want some interesting reading on the subject of personal preference, Harman International is doing research on what constitutes a "good" sound for headphones. I've been reading this through Inner Fidelity (most recent writeup linked).

It's interesting because there is already a pair of headphones (NAD VISO HP50) that was designed based on this research, but they also found that the right frequency balance for "a good speaker in a good room" isn't necessarily a good response for "head-mounted speakers".


It's clear that a flat frequency response isn't the ideal, but what is? The referenced research was supposed to figure that out (and it looks like they did a pretty good job), but the ideal changes depending on who's doing the listening and in what situation. I tried to come up with a way to create an FR band using my own ears as the mics, so I had data that worked well for me, and I think it turned out very nicely.

I used Audacity to generate sine waves (there are better suited tools, but this was what I had handy). I started at 125 Hz (I forget why I didn't test bass, this was months ago) and an amplitude of 1, and then adjusted a 175 Hz signal's amplitude until it sounded the same volume to my ears as the 125. When that's done, compare the 175 to a 250, and on down the line. Recording all of the resulting amplitudes and applying some corrections gives an FR band. This is what I got for my heavily modded Grado SR60i's:

Image

... and the SR60i graph from the Inner Fidelity link, for comparison:

Image

Mine is tilted CCW a tad, has more bass at a slightly lower resonant frequency with higher Q (though it's not really visible here), is very nice 2.5K to 10K, clearly shows the drop at 12.8K, and my hearing doesn't quite make it to 16K (other testing shows my ears are good to 14 or 15K). Correcting for scale, the 2K peak is roughly the same height on each graph.

The Harman correction and my own graph agree much better than most corrected FR graphs I've seen. Most of the differences between them can be explained by my mods, and I think the only difference that's perceptual is the CCW tilt. The tricky bit is the 2K resonance - my mods reduced it greatly, but it shows up about the same.

Thing is, peaks on a FR graph are often also points where transients are bad - these Grados take a lot longer to start or stop making a sound at 2K than they do at other frequencies. A quick scan, such as that done for most FR graph generation, doesn't give the resonance time to reach anywhere near full volume. Switching on and off a 2K sound with other frequencies nearby for comparison makes it very clear there's a problem, in a way that sweep-ish testing doesn't.

Overall, I learned a lot more about the behavior of my headphones from this process than I would have any other way. I only wish I'd thought of this before the mod process; seeing the effects of mods in detail would have been great fun.
deepblueq
Gerbil
 
Posts: 36
Joined: Sat May 17, 2014 9:32 pm
Location: the middle of nowhere

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 5:50 am

It's a shame I can't thumbs-up a forum post.

I tried something similar using sine waves to setup my EQ at home, so that my living room's sound was less picky about what music it played well, and what music it made a mess of. It was surprisingly easy to do, but I also guess that it's different for each person.

I assume that my ears are damaged with age/abuse more at some frequencies than others based on my life experience. Other people with different age/experience may have better or worse hearing than me at any given frequency.
<insert large, flashing, epileptic-fit-inducing signature (based on the latest internet-meme) here>
Chrispy_
Gerbil Jedi
Gold subscriber
 
 
Posts: 1885
Joined: Fri Apr 09, 2004 3:49 pm

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 6:11 am

Chrispy_ wrote:It's a shame I can't thumbs-up a forum post.

I tried something similar using sine waves to setup my EQ at home, so that my living room's sound was less picky about what music it played well, and what music it made a mess of. It was surprisingly easy to do, but I also guess that it's different for each person.

Yes, this way at least you can start with a blank palette, and tweak from there.

Chrispy_ wrote:I assume that my ears are damaged with age/abuse more at some frequencies than others based on my life experience. Other people with different age/experience may have better or worse hearing than me at any given frequency.

I don't just *assume* that, I *know*. The combination of middle age and too many loud rock concerts has taken its toll. I stopped being able to hear "CRT whine" before we got rid of our last CRT-based TV. I also find myself tweaking the high treble up ever further as the years go by, to get music to sound "right". (I also imagine the extreme treble boost may be doing further damage, whenever I listen at more than moderate volume...)
(this space intentionally left blank)
just brew it!
Administrator
Gold subscriber
 
 
Posts: 37737
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 7:42 am

Hey! I recognize that post :o
Damage wrote:Don't try to game the requirements by posting everywhere, guys, or I'll nuke you from space.

-Probably the best Damage quote ever.
superjawes
Graphmaster Gerbil
Gold subscriber
 
 
Posts: 1090
Joined: Thu May 28, 2009 9:49 am

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 7:44 am

The CD of "Sgt. Pepper" recreates the original LP lead-out groove, which is a 15KHz tone. I can still hear it.
Life is hard; but it's harder if you're stupid. Big Al.
Captain Ned
Global Moderator
Gold subscriber
 
 
Posts: 20311
Joined: Wed Jan 16, 2002 7:00 pm
Location: Vermont, USA

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 8:46 am

My effective cutoff appears to be just below 15,000 Hz. I can still hear a 14,500 Hz tone OK, but 15,000 is a no-go unless I crank the level way up.
(this space intentionally left blank)
just brew it!
Administrator
Gold subscriber
 
 
Posts: 37737
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 10:17 am

just brew it! wrote:My effective cutoff appears to be just below 15,000 Hz. I can still hear a 14,500 Hz tone OK, but 15,000 is a no-go unless I crank the level way up.


I still hear 15kHz, but it drops off the cliff at 16kHz. This is why I resample my MP3s at 32kHz to save space
NeelyCam
Gerbil First Class
Gold subscriber
 
 
Posts: 134
Joined: Fri Oct 16, 2009 12:25 pm

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 10:25 am

NeelyCam wrote:I still hear 15kHz, but it drops off the cliff at 16kHz. This is why I resample my MP3s at 32kHz to save space

You may actually be getting some audible artifacts below 16 kHz due to the sample rate conversion though. You'd probably be better off tuning the encoder to roll off the high-frequency content (assuming yours has a setting for this...) while maintaining the original sampling rate. This will still result in some reduction in file size, but potentially with less damage to the fidelity (at least, as far as *your* ears are concerned).

Disk and flash memory have also gotten so cheap that saving a few MB here and there on music files by lowering the bitrate doesn't make as much sense as it used to.
(this space intentionally left blank)
just brew it!
Administrator
Gold subscriber
 
 
Posts: 37737
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 3:33 pm

A more in depth explanation of all of this would be handy for those of us who are just moving toward bettering our music listening experience.
Fastfreak39: I feel like they should change the phrase "jumping on the band wagon" to "sailing on the pirate ship"
Philldoe
Minister of Gerbil Affairs
 
Posts: 2791
Joined: Thu May 13, 2004 12:12 pm
Location: Turn around...

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 4:11 pm

In a nutshell, figure out where the peaks and valleys in the frequency response of your system are (where "system" includes the listening room if you're dealing with speakers), then use a graphic or parametric multi-band EQ to compensate.

Edit: Also, a thought on the original post. Your ears aren't *expected* to have flat frequency response, even if you're young and your hearing hasn't been damaged by years of loud music, mowing the lawn, job-related noise exposure, and so on. I'm not sure what your original goal was here, but if it was to flatten the response curve of the combined system of your headphones *plus* your own ears, I think maybe you've missed the point.
(this space intentionally left blank)
just brew it!
Administrator
Gold subscriber
 
 
Posts: 37737
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 4:37 pm

I recently found a pluggin for Foobar(mathaudio room eq) which uses a a microphone and your speakers to correct for your room acoustics. I have a entry level usb condenser mic I used (fairly flat accorded to their data) and found out I have a huge hole at about 100hz at my listening position. Was actually surprised because I moved my desk because I noted the huge lack of low end with my desk perpendicular to the long part of my room and with my setup now my left speaker is close to a corner. According to my results my right speaker has less of a dip there but more of one below there in the real low stuff. I have it setup for a partial correction, otherwise the bass becomes overbearing and the highs become recessed, might be either my mic or my unscientific method of just holding my mic as close as I can to ear level.
I use M-audio BX8 studio monitors btw, just on my desk, too cheap for speaker stands.
Ps I find my Sennheiser 485s to be lacking in high end as well as having a hole in the mids to midrange in comparision to my speakers
For the curious: Image
Edit- everything above the white line gets corrected, everything above gets ignored, that's what the slider is for, how corrected you want your response to be.
Mikael33
Gerbil In Training
 
Posts: 8
Joined: Tue Mar 18, 2008 5:13 pm

Re: Headphone measurement stuff

Postposted on Tue Jul 22, 2014 4:55 pm

The dip at 100 Hz could very well be room cancellation effects. Hanging some sort of sound absorbing material on the far wall might help.
(this space intentionally left blank)
just brew it!
Administrator
Gold subscriber
 
 
Posts: 37737
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Headphone measurement stuff

Postposted on Wed Jul 23, 2014 11:08 am

I notice you've got a strong signal at 50Hz but then dips at the first, second, and third harmonics (100, 200, and 400 Hz). Seems indicative of a standing wave that is out of phase with the source.

Out of curiosity, what are the approximate dimensions of your room?
He who laughs last, laughs first next time.
ludi
Gerbil Elder
 
Posts: 5441
Joined: Fri Jun 21, 2002 10:47 pm
Location: Sunny Colorado front range

Re: Headphone measurement stuff

Postposted on Wed Jul 23, 2014 1:42 pm

superjawes wrote:Hey! I recognize that post :o


Sorry, continuity without derailing can be a bit tricky with linear threads.

Philldoe wrote:A more in depth explanation of all of this would be handy for those of us who are just moving toward bettering our music listening experience.


(I'm not an expert, and my explanation is likely to contain inaccuracies. I'd love to learn more though, if anyone sees something that doesn't scan.)

Speakers and headphones differ from ideal in a variety of ways, the most notable being that different frequencies have different volume levels (frequency response), they can't start or stop making a sound instantly (transients), and they add some small amount of distortion (negligible unless you're using very high volume levels or very cheap equipment). There's also some stuff with phase effects, but I'm not qualified to say much on that (my impression is that if you've got frequency response and transients sorted, you've also got these sorted). Generally, if you've got good frequency response (FR) and transients, you've got a very good sounding system.

Frequency response is usually measured by applying constant volume at different frequencies to the speakers or headphones being tested, and measuring how much of each frequency makes it to a calibrated microphone. My way was basically the same thing, just using my ears as the microphones (and calibration methods better suited to that). The graph created just shows volume at each frequency.

Transients are also different for each frequency. They're best represented in a 3D graph (or waterfall plot), where frequency is the x-axis, time since the sound was started is the y-axis, and volume is the z-axis. As you might imagine, creating these is a bit of a cumbersome process. We can make it a lot simpler by just taking two slices along the y-axis, instead of trying to fill the graph in 3D. Measuring each of these two slices works the same as measuring an FR band; the difference between the two is the time between the initialization of a sound and the measurement of the sound. The further apart the two graphs are, the worse transients are at that point.

My graph effectively gives me some transient data as well as FR data, both because the graphs I have to compare it to were made with much shorter test tones (so the differences between mine and the rest are noteworthy), and because when testing a point with bad transients (2 KHz, for Grados), the ramp-up / ramp-down time is clearly audible.

Raw (microphone-measured) FR bands shouldn't actually be flat, and the corrections to flat are a bit different for everyone. By the time the measurement is done using your own ears (like I did), you could argue that flat is ideal (but there are also reasons you might not want it to be). Personally, I like bass and treble to be a bit higher, because it more closely matches the human ear's sensitivity at lower volume levels, so I get the same detail with less stress on my ears (look up the loudness curve for more on that).

Transients should be fast. I'm not clear on the limit behavior as things get ridiculously fast (subwoofers with Q < 0.5 in sealed boxes don't sound quite right - is that the same thing?), but almost all of the time, you'd rather they were faster. In general, I'd rather have headphones with fast transients and a bad FR band than the other way around, because it's a lot easier to correct an FR band with an equalizer.

Equalizers work pretty well, all told. Transients should be correctable to some extent (boost volume on noise initialization to get it started quicker, apply a negative to stop the noise quicker), but last I knew, Eastern Acoustic Works held a patent on it and was only using it for auditorium-scale stuff. I wanted to mess with that tech a bit, but there's not much point now.

just brew it! wrote:Edit: Also, a thought on the original post. Your ears aren't *expected* to have flat frequency response, even if you're young and your hearing hasn't been damaged by years of loud music, mowing the lawn, job-related noise exposure, and so on. I'm not sure what your original goal was here, but if it was to flatten the response curve of the combined system of your headphones *plus* your own ears, I think maybe you've missed the point.


The original goal was just to do some fun stuff with sound. My experience didn't match the FR bands I'd seen very well, so I thought I'd see what I could come up with on my own. Incidentally, tuning so that this graph is flat actually sounds very good, but I don't use it anyway (I like the uncorrected FR band for its loudness curve effect). I mainly use my phone as a source, and there doesn't seem to be such a thing as a parametric EQ for Android (not that I can tell why - the underlying API is all parametric), so I rarely use EQ at all with the Grados.
deepblueq
Gerbil
 
Posts: 36
Joined: Sat May 17, 2014 9:32 pm
Location: the middle of nowhere

Re: Headphone measurement stuff

Postposted on Wed Jul 23, 2014 2:31 pm

If you want to compact a few things or another take...

Frequency response plots are showing power. This could be electrical or audible, and it also changes depending on where we are in the communication stream. For music, we're starting with audible power. The sound being picked up is going to have a frequency response with power being dropped in around each instrument or voice's range. The specific blend will depend on your choice of music.

But then we switch to electrical devices, and any signal that goes through a device with a variable frequency response is going to have amplification and/or attenuation (amplification means more power was added and attenuation means power was lost). Yes, every electrical device has a variable frequency response. Generally we talk about usable range, and we don't really care if an audio device attenuates a signal that is outside human hearing. This thread is generally focused on the last step: equalization for human ears. That's basically just adjusting the final content so everything sounds "right," as some frequencies have probably been attenuated by the system or speakers, and adding a little more power brings them back to where they "should" be. [i](Note: the "right" equalization might not recreate the original signal. A listener might prefer more power in the bass, recessed mids, etc.) If you have a speaker, your sound waves will reflect and interact with newer waves, further changing how things sound and effectively adding another component to your listening system.

Transient responses are really important because we're talking about a physical system. There are general frequency elements, yes, but most speakers and headphones use dynamic drivers, which are spring-mass-dampener systems. If you had a giant dynamic driver and kicked it, it would oscillate for some time before settling. If you applied a DC current, the same thing would happen, but the system would try to settle somewhere off the resting point since you're adding a magnetic force. These effects still happen with complex signals, but measuring the responses gives you an idea of how much of the audible output is the music and how much is being added by a non-ideal system. Just knowing how long your speaker wants to "ring" at any given frequency is immensely useful.

Now generally we talk about flat frequency responses in electronics, and that's mainly because we don't want to lose any of the information. Even if all of the information makes it to the speaker, with a non-flat response, some frequencies will get more power than others, resulting in an imbalance of the information. Your ideal transients are nonexistent. That is, you don't want any transient behavior because it adds new, undesired information to the signal (and since no ideal systems exist, you will always have some transient behavior).

EDIT: I didn't condense things too much...oh well. Hopefully this was useful :lol:
Damage wrote:Don't try to game the requirements by posting everywhere, guys, or I'll nuke you from space.

-Probably the best Damage quote ever.
superjawes
Graphmaster Gerbil
Gold subscriber
 
 
Posts: 1090
Joined: Thu May 28, 2009 9:49 am

Re: Headphone measurement stuff

Postposted on Wed Jul 23, 2014 3:05 pm

This sounds like a neat little project for me. I picked up some Sony MDR-10 RBT's recently. The sound quality on them far surpassed what I had before and they really bring out a lot in my music that I never heard before. I'd love to know exactly how you guys go about making these adjustments so I can see if I can squeeze more out of these headphones.
Fastfreak39: I feel like they should change the phrase "jumping on the band wagon" to "sailing on the pirate ship"
Philldoe
Minister of Gerbil Affairs
 
Posts: 2791
Joined: Thu May 13, 2004 12:12 pm
Location: Turn around...

Re: Headphone measurement stuff

Postposted on Wed Jul 23, 2014 3:19 pm

ludi wrote:I notice you've got a strong signal at 50Hz but then dips at the first, second, and third harmonics (100, 200, and 400 Hz). Seems indicative of a standing wave that is out of phase with the source.

Out of curiosity, what are the approximate dimensions of your room?

I realized that too after looking at it better heh. My room is about 11x14, my desk is setup to my speakers face long ways, slightly to the left. I definitely need some acoustic treatment for my room, especially since I record and mix music in there but I've been to cheap and lazy to.
Mikael33
Gerbil In Training
 
Posts: 8
Joined: Tue Mar 18, 2008 5:13 pm

Re: Headphone measurement stuff

Postposted on Thu Jul 24, 2014 8:51 am

Mikael33 wrote:
ludi wrote:I notice you've got a strong signal at 50Hz but then dips at the first, second, and third harmonics (100, 200, and 400 Hz). Seems indicative of a standing wave that is out of phase with the source.

Out of curiosity, what are the approximate dimensions of your room?

I realized that too after looking at it better heh. My room is about 11x14, my desk is setup to my speakers face long ways, slightly to the left. I definitely need some acoustic treatment for my room, especially since I record and mix music in there but I've been to cheap and lazy to.


Growing up, my father and I were into speaker building and acoustics, and he has some friends that were over the top into it. Some of the contraptions hanging from ceilings to dampen standing waves were interesting to say the least, and may have caused more than one divorce. Probably the best ones were free-standing super-sized cylindrical "cat scratching posts" in the corners. You could cleverly make them into actual cat scratching posts and climbing towers. Appropriately placed near corners of a room, they do a fantastic job of dampening standing waves. The ones I saw were created by using 1/2" steel fencing as a frame, filled with stuffed animal stuffing, and wrapped in carpet to match room decor.

Of course there is always the heavy-duty window drapes with insulation. And they don't take up room space much. Spouses have been known to prefer this option, though 3 year old boys think they make great things to rip down during games.

A side note, properly reproducing transients is related to dampening of the driver which couples electrical energy into mechanical energy of the driver. A speaker for example with have absorptive materials inside to assist with this, plus the amplifiers typically driving them have a very low impedance well below of the speaker impedance which also absorbs regenerated power created by mechanical ringing of the driver/voice coil/magnet generator. A good setup between speaker and amplifier will not resonate or ring at any specific frequency. No amount of equalization will help with this, but speaker design, wiring, and amplifier selection will.

Headphones are in a different order. Usually they are in free air, requiring a baffle behind the driver to provide the damping since there usually isn't the luxury of volume to add enough absorptive materials to make a difference. The baffle and driver have to work together in conjunction with the coupling to the ear to get a reasonable transient response without ringing, yet still provide a fairly level response. It is actually very difficult to make good headphones, and why they can all sound so different from each other. It becomes a black magic since there are so many variables that go into the mix. Least of which are the amplifiers that drive them, which are NOT low impedance in some cases, exasperating the difference between headphones.

Then you have the phenomenon of you ears, just like your eyes, equalizing the sound all by themselves. Another knob you just don't have control over. Your brain keeps turning it on you while you are trying to calibrate "perfection".
liquidsquid
Minister of Gerbil Affairs
 
Posts: 2447
Joined: Wed May 29, 2002 10:49 am
Location: New York

Re: Headphone measurement stuff

Postposted on Thu Jul 24, 2014 9:58 am

just brew it! wrote:I don't just *assume* that, I *know*. The combination of middle age and too many loud rock concerts has taken its toll. I stopped being able to hear "CRT whine" before we got rid of our last CRT-based TV. I also find myself tweaking the high treble up ever further as the years go by, to get music to sound "right". (I also imagine the extreme treble boost may be doing further damage, whenever I listen at more than moderate volume...)


Yeah, that's the smallest stereocilia that get damaged/die-off as you age, and that affects the cuttoff frequency pretty at pretty much the same rate for everyone, give or take.

I was thinking about perhaps my ears have been subject to abuse at, say, 1KHz and 8KHz so a flat response to me would be volume spikes to someone else at those frequencies, but you have been exposed to large volumes at 3KHz and 6KHz so your "flat response" would seem to have volume peaks at 3Khz and 6Kz but be too quiet at 1KHz and 8KHz because that's where my own hearing is worn out

In other words, ignoring the peak frequency degeneration with age/abuse, is there also no such thing as a "flat response" because everyone's perception of the response curve is completely different?
<insert large, flashing, epileptic-fit-inducing signature (based on the latest internet-meme) here>
Chrispy_
Gerbil Jedi
Gold subscriber
 
 
Posts: 1885
Joined: Fri Apr 09, 2004 3:49 pm

Re: Headphone measurement stuff

Postposted on Thu Jul 24, 2014 12:56 pm

I don't know that I'd say everyone loses their high-frequency hearing at the same rate. While I'm sure there's always some loss with age, genetics and lifetime environmental exposure are factors that would be hard to take into account when testing. I've seen some older folks take those hearing tests (one was a coworker in their upper 50's) and be able to hear the "mosquito sounds" which are supposed to only be perceptible to teenagers (I can still hear them in my lower 30's, while some in their 20's could not). Keep in mind that you wouldn't say everyone's eyesight degrades at the same rate.
i5 2500k - P67 - GTX660 - 840 Pro 256GB - Xonar Essence STX - Senn HD595's
The Egg
Gerbil XP
 
Posts: 478
Joined: Sun Apr 06, 2008 4:46 pm

Re: Headphone measurement stuff

Postposted on Thu Jul 24, 2014 2:45 pm

Mikael33 wrote:
ludi wrote:What are the approximate dimensions of your room?

I realized that too after looking at it better heh. My room is about 11x14


Yup. 11 feet is ballpark for a 100Hz wavelength, hence resonance. And apparently you're getting a couple harmonics out of it as well.
He who laughs last, laughs first next time.
ludi
Gerbil Elder
 
Posts: 5441
Joined: Fri Jun 21, 2002 10:47 pm
Location: Sunny Colorado front range


Return to Echo Vale

Who is online

Users browsing this forum: No registered users and 3 guests