Understanding Frequency Response Measurements

An explanation of how to use frequency response graphs on this channel and other channels / sites too, including the limitations and how graphs can be helpful.

00:00 Introduction
01:26 Headphone measurement devices
02:37 Measurements and explanations
09:36 Measurement software
09:55 The importance of scale
11:20 Smoothing vs unsmoothed
14:25 Compensation curves
18:19 Headphone vs IEM measurements
20:14 Summary and conclusions

SUPPORT THE CHANNEL
Or put one of these links in your bookmarks for all of your future purchases of ANYTHING:
🇺🇸 USA & 🇨🇦 Canada
🇦🇺 Australia
🇩🇪 Germany
🇬🇧 UK
🌏 Global

PASSION FOR SOUND RECOMMENDS
(Affiliate links to Amazon and Apos Audio help to support the channel - thank you!)

Playback software

DACs

USB cables

Interconnects

Active Speakers:

Amps

Headphones

Headphone Stand

Portable

Earphones

Wireless 'phones

Reviews filmed using:

[dasninjastix]

Great video, I understand why people like FR response graphs but now with the experience I've accumulated I've yet to find them useful in making a purchase. I have used them when I'm trying to set an EQ and I know there's a certain frequency range I want to address but not precisely. But evaluating a device based on a FR is nonsense to me and I know, I know most reviewers go to lengths to caution against doing so.. only to then spend 10-12 minutes of a review discussing the ins-and-outs of the FR 😒.

I know you're not in that camp so I appreciate the video showing what you might/may be able to discern from a graph. I think people are very confused by the difference between a given frequency response and THE QUALITY of that frequency response. And there in lie the problem for this audioninja. A headphone or earphone is a complete package. Even a driver engineered for minimal distortion can sound whack af if it isn't tuned right or isn't dampened sufficiently or doesn't vent properly or doesn't seal properly. There seems to be too many much mucho Physics getting in the way of my audio enjoyment than the blasphemy of a V-tune. I'm sure the acolytes of ASR and King Crin will have my head for such heretical postings but.. there is no "correct" frequency response. It's a sum-of-its-parts proposition for audio quality in every damn headphone or IEM I've tried.

Rant over. Thanks for the video 🤘

[xwqr]

I like how wild graphs of different headphones and IEMs can get. It's like food, you get the same seasonings but with different chefs you get different amounts of seasonings, different styles of cooking and ingredients. I'm probably just hungry cuz it's midnight and i just finished work so sorry for talking about food all of a sudden, should probably sleep soon...

[RbNetEngr]

Another excellent and educational video. Hopefully the viewers will pay attention and absorb all the concepts that you explained. Or maybe they’ll watch it several times, so the concepts take hold.

Here are a few questions that I have:

1) What do you use as your audio source for the FR tests? White noise? Pink noise? Any special parameters applied?
2) What audio level (in dBm) is your source signal? Is that the level applied to the input of the headphones? Or the input to the DAC? Do the FR responses reflect audio level losses caused by the efficiency level of the headphones? And would that be of any value to show how impedance affects the signal level?
3) Have you considered building a small anechoic chamber or enclosure for your “ears” setup, to block out stray audio from the surroundings?
4) Your vertical scale on the FR curves is linear, from 60 - 96 dB, right? Are the lines spaces at 2 dB each? It’s hard to read from the video.
5) Are there “ear systems” (sorry, I didn’t catch the official term for the system) that are optimized for testing IEMs? And, if so, would such a system offer additional value in being able to also test the effect of different tips on the FR? Or are there just too many variables with IEMs that the test results wouldn’t translate well to user experience due to the variety of shapes and sizes of people’s ears?

Bottom line, I was very impressed with your video. As I watched it, I was already composing a list of questions to ask. But as the video progressed, you were answering them, and the answers went to great depth - more than I expected in most cases. As an engineer, who is a stickler for technical details, I think you did a great job of presenting this topic in enough detail for most viewers, and hopefully also revealed topics and terms that interested viewers can use to continue their research and learning.

[tynanmedia]

Loved this one mate! Very educational and easy to understand.

[robertnoblitt]

Thanks, Lachlan, for this enlightening and insightful video!

[heyguyslolGAMING]

Tyvm for the vid, I'm very much interested in the future course.

[MatteoCeccarini]

Thank you very much for this instructive video.
How should one be concerned about the width of the frequency response of headphones? For example, the HE400se are 20-20k, while the HE400i 2020 are 20-35k. Is that extra 20k to 35k range any meaningful? If I remember correctly you prefer the 400se among the two.
Thanks!

[RbNetEngr]

Can you explain what a waterfall graph is? I am familiar with frequency response graphs (being an engineer, and also having been in the audiophile world in my earlier years (BK - Before Kids).

[kyron]

Would be interesting to see the waterfall graphs in the reviews.

[MrBNCLE]

Edit: Yes, I know I can't compare measurements from different people, but the mini dsp is really off when it comes to IEM bass to the point where it doesn't even resemble reality

I've always wanted to mention this about your IEM measurements: the bass is always unnaturally rolled-off by like a good 8dB. I always thought it was your target curve, but it seems like the mini dsp is simply incapable of measuring IEM bass.

The raw graph of the OH10 for example, should have an 8dB bass shelf, yet it measures flat there. You did mention that it is bassy, which it is. But at this point, I think all the measurement does is just confuse people as to what they actually sound like, especially if they're familiar with how to read graphs.

I suggest simply just describing the sound signatures for IEM reviews (you're already plenty good at that w/o graphs), as the graph is simply misleading.

[canalfamiliar]

How is the compensation curve made? video please

[Koopagamelouda]

I don't know if this is the best place to suggest this but I recommend Mona Lisa Overdrive by Juno Reactor for a test track. It's a very suspenseful track with both hard electronic beats and an orchestra. It's a good test for a headphone to render both convincingly i.e acoustic instruments with depth and beats with slam... And Matrix Reloaded is a fine film.

[TheMrZnyper]

Is it just me or is the picture in picture very choppy?

[kyron]

What's the difference between a wiggle, a squiggle and a jiggle?

[STSS_]

is it my phone or the video have low fps? 😅

[RbNetEngr]

I think that trying to record sound from headphones to do a comparison is fraught with too many variables to make it valuable. First there’s the setup, and making sure it’s consistent for the headphone(s) in the comparison. Then there are losses and variations when recording that would likely make the results very different than what actual human ears would hear. Third, the quality of your recording then needs to be done using some Hi Res Lossless setup to insure that the recording is as good as possible. And finally, the consumers of your comparative audios will listen to them on a variety of headphones, speakers, IEMs, DACs, etc., so the consumers will hear different results. To me, the effort to produce comparative audio recordings is probably more work than the value it would provide for a very small subset of your subscribers.

[urbenkam]

Thank you for explaining it. Btw if you like to review he6se v2 do let me know :$

[kyron]

It would be much better to display the frequency response charts by themselves rather than overlying them on the video.