Easy (but rough) frequency response measurements!

In a continued effort to be more quantitative in this hobby, I’ve started to use this super easy and free method of taking relative frequency response measurements of my headphones and mods. I’m sure a few of you have done this or something similar before, but I’d thought I’d share because I thought it was neat :slight_smile: . If you mod your headphones and don’t want to drop the 2-hundo for the minidsp EARS, but you want to see roughly how your mods are changing the frequency response of your headphones, this is a great method. Note - this can’t be used for accurately for EQ’ing your headphones, as it doesn’t utilize the ear shape of other measurement setups and might not seal against the pad perfectly.

I’m using the app ‘spectroid’ for android, I’m sure there is an iOS equivalent.
This app FFTs the signal from your phones microphone in real time, and plots the maximum amplitude of the FFT for you. I’ll list what I’m doing in order below, then show some pictures of results.

  1. Download spectroid, or equivalent iOS app
  2. Download a high-def frequency sweep from audiocheck.net - https://www.audiocheck.net/testtones_highdefinitionaudio.php
  3. Stick your phone (microphone end) into your headphone
  4. Seal your cup against your phone with something - I used a densely rolled up towel, then wrapped it around my phone and pressed against the pad to make a seal. Be sure to pull your phone up a bit from the driver to approximate where your ear would be.
  5. Reset the max on the spectroid app and play the frequency sweep!

You might have to play around with the settings in the app to achieve optimal results. For example, if you use spectroid, make sure to set the decimations to 1 to avoid sharp changes in the spectra at the decimation edges.

I’m sure there’s a better way to make a seal and achieve more accurate results than what I did, but if all you’re using this for is relative measurements, then you can just measure a handful of times to make sure your results are consistent.
Here are some of my measurements I made today just after I started! So far I’ve measured:

  • 400i’s with stock and angled brainwavz pads
  • DT770s with stock and angled brainwavs pads
  • 6XX’s with stock pads
  • All through a khadas tone board and THX AAA 789

The red is the max value and the yellow is real time. With a better seal you’ll have more accurate results, especially below ~100 Hz. And yes, I should have changed the y-axis range on these plots… Next time :wink:

400i’s with stock pads

400i’s with angled brainwavz pads

DT770s with stock pads

DT770s with angled brainwavs pads

6XX’s with stock pads

Hope you all enjoy! Going to compare some mods on my Thieaudio Phantoms that are on the way with this method - I’ll be sure to post to the appropriate thread!


did you forget the image of the 6XX with alternative pads?

Naw, the senns are more of a pain to pad swap so I left em out.
Mostly just trying to show off the method here rather than the individual measurements.

It never occurred to me measure headphone this way lol. But if your doing enough headphone modding to warrant measurements, $200 seems well worth it to get the EARS. I’ve been tempted to get it for assisting with doing EQ

I would take the ears measurements with a grain of salt tbh. IMO its only valuable if you compare it to other ears measurements or realize the shortcomings of the ears (both minidsp and human lol)

Yeah I’ve looked at the reviews and don’t expect it to perform well compared to the high end stuff. I wish there was a mid range measurement rig. Anyway, I would use my HD600 and 6XX as frames of reference. Then hopefully the relative differences between measurements will match the relative differences from what I hear.

It depends on what you’re using the measurement setup for of course… For me, $200 is a lot if I’m not doing any EQ’ing and I’m just comparing mods.

For example, if you pick up the same features and relative amplitudes in your frequency spectrum from measurement to measurement, you can start to get an idea of what your mods are doing if say, you consistently see an enhancement through the mid-range with a pad swap (like I see in the 770’s).

That said, the phone method doesn’t take into account the phone/microphone’s transfer function, so you have to be careful. You can start to identify non-physical features by running test tones in the app or by finding identical features b/w different sets of headphones. Certainly not great for an ‘absolute’ measurement of a headphone.

The AutoEQ site has over 2000 headphones measured on the high end gear. With maybe a 1000 data points for each plot. Of course each unit of a given model of headphone is not going to be identical to the unit(s) that were measured. But one can listen for any variance using sine sweeps.

I totally get wanting to go mano y mano with a measurement system. But for someone non-electro-technical like me, and after reading about the struggles of the guys on SBAF with mini-DSP calibration, I feel really good that I can make do with the alternate approach.

Not to be a wet blanket or anything.

1 Like

I think the motivation to do our own measurements is partly just for the fun of it. AutoEQ is certainly a good option if you’re not into that though.

Can also be a useful troubleshooting or research feature. Don’t think yours sounds similar to a friends? Want to see how pad wear can effect response? Measuring might tell you something that you couldn’t find elsewhere (but still take your results with a grain of salt)

Absolutely. I certainly feel the allure myself, but I know I don’t have the right technical background. So I have that sour grapes rationale in my previous post ready to trot out to keep myself talked out of it.

Yet part of me is nevertheless hoping Marcgli or someone presents an irresistible counter-argument, so I too can talk myself into buying a new toy. :wink:

1 Like

I mean if you want to do it there’s nothing wrong with getting something like that. Its mainly time it takes more then technical background. You can easily teach yourself how to measure accurately and have fun with it. Just don’t expect high level pro results because of the lack of higher end equipment and treatment and other stuff you need.

Sadly, I probably couldn’t live with that reduction in accuracy you mention. But on the up-side I discovered a new toy, yet one that was under my nose all along. I’ve been using the rather Mickey Mouse spreadsheet that comes bundled with my MacBook for a while now to handle my EQ number crunching needs. Then a few days ago a light bulb turned on – I realized I could graph those same numbers to look like real frequency response charts, replete with pretty colours.

When Marzipan and you were discussing the Sundara vs the Ananda’s bass interactions with various amps, I wanted to see how close they were natively. So I quickly cooked up this graph:

Answer: pretty darned close!

When I was shopping for a new headphone last winter and wanted a rough idea what the various headphones sounded like, I used a headphone I owned to simulate their frequency responses with an EQ like this:

When I wanted to try a new candidate for a neutral compensation curve, I used an EQ curve like this:

This is the sort of thing a deadly boring person like me does to amuse himself in his dotage.

And any time you want your thread back, s_beck, just say the word. :wink:


There’s some software like toneboosters morphit or others that I can think of, but I don’t know how accurate they are.

But yeah the new hifiman stuff isn’t that amp reactive

But those look like pretty nice graphs. Definitely easy to read and understand

1 Like

It’s all good! Nice looking graphs. I might be weary of such few samples though, as I’m sure the frequency response changes by multiple dB much faster than the space b/w those data points in some places.

Separately, if you’re into manual EQ’ing, (maybe some people already use this), I like to use this online tone generator with a slider. Super handy if you think you can here a bump or a dip over a particular part of the spectrum, but you have to be careful about aliasing from your browser - https://www.szynalski.com/tone-generator/

1 Like

Whoa! Super cool! I may well be using this instead of sine sweeps and Audacity tones from now on.

Amen to that. What system-wide tool do you use that has finer control? On a Mac, the only options I have are either 10-band or 31-band GEQ. That said, working my way across the frequency spectrum with that slider app you suggested, I’m hearing buttery smooth and gradual rises and falls just where I want them.

I recently started playing with EqualizerAPO+Peace on my Win box. There must be some trick to PEQ I haven’t hit on yet, but so far I’m not seeing getting as much control.

Quick check shows it to be a DAW plug-in. Don’t have any DAWs on my MacBook. Then there’s Sonarworks/True-Fi. 'Fraid to install anything like that. The EQ software I’m using is flaky; I’d be afraid to render it unusable.


You could always drop a couple hundo and pick up a miniDSP :slight_smile:
Then you’d have all the control you want…

Regarding EQ APO, lack of control shouldn’t be an issue.

Here’s a guide.

1 Like

EQ APO is a very powerful tool. I wouldn’t underestimate what it can do