Focal Pad Rolling - Experiments & Measurements

[boxster233]

Using any on the Focal lineup? I also have TH-X00.

[TylersEclectic]

I liked the Elite Hybrid on the TH-X00 Purplehearts…but now I have ZMF vegan pads on them…but they are slightly to big to fit properly on the rings, but they are soo comfortable…I should probably order proper sized ones.

@zach915m do you have any recommendations on pairings, not sure if you have tried using your pads with other HPs. I can’t wait to get my Aeolus in…insert squealing school girl reaction here

Edit: @boxster233 I just ordered the finestrated sheepskin models from the latest Dekoni drop, I ordered them before but was sent LCD 2 pads on accident, which Instead of being upset about just took a refund on the price deference and threw them on my modded M1060c open backs.

[Torq]

Focal Elex - Pad Rolling Results

Here are the frequency response plots for the Focal Elex equipped with their stock pads, as well as those of the Elear, Elegia, Clear, Stellia and Utopia:

The first set, above, is calibrated for 84dB @ 300Hz, raw plots, with mic-calibration only (no headphone compensation). The same plot, but vertically spaced, looks like this:

Finally, the same model but calibrated for 84dB @ 1kHz - again raw plots, with mic-calibration only (no headphone compensation).

Plots with experimental headphone compensation will follow once I’ve done all of the headphones this way.

---

Measurements performed on a miniDSP EARS with microphone calibration only - i.e. no headphone compensation in an isolating acoustic chamber. These results are only directly comparable with other measurements performed on the same rig. The miniDSP EARS exhibits a strong resonant peak around 4.5kHz, which you should take into account with these raw plots. Plots are an average of 4 passes for each of 5 different placements of the headphones on the measurement stand.

[Torq]

Focal Elear - Pad Rolling Results

Here are the frequency response plots for the Focal Elear equipped with their stock pads, as well as those of the Elex, Elegia, Clear, Stellia and Utopia:

The first set, above, is calibrated for 84dB @ 300Hz, raw plots, with mic-calibration only (no headphone compensation). The same plot, but vertically spaced, looks like this:

Finally, the same model but calibrated for 84dB @ 1kHz - again raw plots, with mic-calibration only (no headphone compensation).

Plots with experimental headphone compensation will follow once I’ve done all of the headphones this way.

Measurements performed on a miniDSP EARS with microphone calibration only - i.e. no headphone compensation in an isolating acoustic chamber. These results are only directly comparable with other measurements performed on the same rig. The miniDSP EARS exhibits a strong resonant peak around 4.5kHz, which you should take into account with these raw plots. Plots are an average of 4 passes for each of 5 different placements of the headphones on the measurement stand.

[Torq]

Focal Elegia - Pad Rolling Results

Here are the frequency response plots for the Focal Elegia equipped with their stock pads, as well as those of the Elex, Elear, Clear, Stellia and Utopia:

The first set, above, is calibrated for 84dB @ 300Hz, raw plots, with mic-calibration only (no headphone compensation). The same plot, but vertically spaced, looks like this:

Finally, the same model but calibrated for 84dB @ 1kHz - again raw plots, with mic-calibration only (no headphone compensation).

Plots with experimental headphone compensation will follow once I’ve done all of the headphones this way.

Measurements performed on a miniDSP EARS with microphone calibration only - i.e. no headphone compensation in an isolating acoustic chamber. These results are only directly comparable with other measurements performed on the same rig. The miniDSP EARS exhibits a strong resonant peak around 4.5kHz, which you should take into account with these raw plots. Plots are an average of 4 passes for each of 5 different placements of the headphones on the measurement stand.

[Torq]

Focal Clear - Pad Rolling Results

Here are the frequency response plots for the Focal Clear equipped with their stock pads, as well as those of the Elex, Elear, Elegia, Stellia and Utopia:

The first set, above, is calibrated for 84dB @ 300Hz, raw plots, with mic-calibration only (no headphone compensation). The same plot, but vertically spaced, looks like this:

Finally, the same model but calibrated for 84dB @ 1kHz - again raw plots, with mic-calibration only (no headphone compensation).

Plots with experimental headphone compensation will follow once I’ve done all of the headphones this way.

Measurements performed on a miniDSP EARS with microphone calibration only - i.e. no headphone compensation in an isolating acoustic chamber. These results are only directly comparable with other measurements performed on the same rig. The miniDSP EARS exhibits a strong resonant peak around 4.5kHz, which you should take into account with these raw plots. Plots are an average of 4 passes for each of 5 different placements of the headphones on the measurement stand.

[Torq]

Focal Stellia - Pad Rolling Results

Here are the frequency response plots for the Focal Stellia equipped with their stock pads, as well as those of the Elex, Elear, Elegia, Clear and Utopia:

The first set, above, is calibrated for 84dB @ 300Hz, raw plots, with mic-calibration only (no headphone compensation). The same plot, but vertically spaced, looks like this:

Finally, the same model but calibrated for 84dB @ 1kHz - again raw plots, with mic-calibration only (no headphone compensation).

Plots with experimental headphone compensation will follow once I’ve done all of the headphones this way.

Measurements performed on a miniDSP EARS with microphone calibration only - i.e. no headphone compensation in an isolating acoustic chamber. These results are only directly comparable with other measurements performed on the same rig. The miniDSP EARS exhibits a strong resonant peak around 4.5kHz, which you should take into account with these raw plots. Plots are an average of 4 passes for each of 5 different placements of the headphones on the measurement stand.

[Torq]

Focal Utopia - Pad Rolling Results

Here are the frequency response plots for the Focal Utopia equipped with their stock pads, as well as those of the Elex, Elear, Elegia, Clear and Stellia:

The first set, above, is calibrated for 84dB @ 300Hz, raw plots, with mic-calibration only (no headphone compensation). The same plot, but vertically spaced, looks like this:

Finally, the same model but calibrated for 84dB @ 1kHz - again raw plots, with mic-calibration only (no headphone compensation).

Plots with experimental headphone compensation will follow once I’ve done all of the headphones this way.

Measurements performed on a miniDSP EARS with microphone calibration only - i.e. no headphone compensation in an isolating acoustic chamber. These results are only directly comparable with other measurements performed on the same rig. The miniDSP EARS exhibits a strong resonant peak around 4.5kHz, which you should take into account with these raw plots. Plots are an average of 4 passes for each of 5 different placements of the headphones on the measurement stand.

[Luckbad]

Awesome! Is that 4.5kHz resonant peak also present around 9kHz?

I liked the Clears when I reviewed them, but something about them always caused me fatigue. Something in the upper mids or highs that I wanted to tame, and maybe one of those pads will help (Elear pads?).

[ufospls2]

Now that is an epic line up!

[Torq]

I don’t think that’s a resonance in the “canal” like the peak around 4.5kHz. It might be some other calibration artifact. I am still working on dialing in a headphone compensation model that I am happy with relative to what I hear (and in observance of my personal hearing profile).

Applying the SBAF compensations here result in those ~9kHz peaks going away and appearing as generally perceptually flat-to-downward trend.

---

Where you like bass … it’s probably the Stellia, in stock-form, that are your best-fit in the Focal line-up. But that’s not to say a bass-head should be looking at Focal in the first place …

[Torq]

Here’s a different way to look at the data …

These two plots (calibrated at 300Hz and 1khz, respectively) show a single set of pads as they interact with each of the Focal headphones. In other words, the pads are the same for all plots, it’s the headphone that changes (rather than the opposite, in the above plots).

[antdroid]

Great work @torq!

I may have to copy and paste my aftermarket pad measurements for the Elex and Elegia into here when I have time.

The Elex + Stellia pad combo has that dip in the upper mids and rise up to 4-5KHz that’s quite similar to how the HE560 behaves.

Update: Copy Pasta completed (next 2 posts)

[antdroid]

This is a comparison of the Focal Elegia stock pads vs the Brainwavz Microsuede pads .

The microsuede pads were attached using double-sided tape to the Elegia in this test.

Comfort:

These new pads are very comfortable to wear. They are softer to the touch than the scratchier feeling of the stock pads and use thicker memory foam inside. Besides that, they are relatively the same depth.

Listening:

Microsuede pads sounded like it elevated the bass a bit more, and sounded fuller and slightly boomier (though not distasteful). The soundstage was probably just a tad more closed in.

Measurements:

The first graph is raw Frequency Response and second graph is compensated using HEQ compensation which is going to be similar to the Harman Target compensation, but supplied by MiniDSP.

Generally they have very similar characteristics in the most of the spectrum except in the lower bass region, where the Brainwavz pads elevates it. There is a noticeable listening experience as I mentioned above.

I also decided to plot out the decay in the following charts. The first one is stock and the second one is the Brainwavz Pads. In general, they are pretty similar.

[antdroid]

Here are the measurements for the Focal Elex of the stock pads vs the Brainwavz Microsuede pads

Listening Impressions

Noticeably warmer sound signature with the Microsuede pads. A little bit more bass impact, tamer upper mids and treble. More laid back overall.

Is this what the Elear or Utopia sounds like? I havent listened to the Elear in a very long time and never listened to the Utopia. Forgot to listen to it when I was at The AV Source earlier this year. It sort of looks like a frequency response similar to either of those two from my apples to oranges comparison of graphs

Charts

Raw Frequency response

Compensated Frequency Response

Waterfall CSD - Stock (top) vs Microsuede (bottom)

I decided to remove the fabric screen/filter from the microsuede pads to see if it would elevate the treble a little more, and it did. It’s closer to neutral now.

[Anonimar]

@Torq I’m used to reading Innerfidelity’s measurements for various headphones, so seeing the huge treble peaks in your charts was throwing me off at first. I went back to Innerfidelity’s measurements and realized they are compensated, whereas yours aren’t.

Do you happen to have any resources on headphone compensation - googling isn’t getting me productive results. What’s it compensating for? Why’s it considered experimental?

[Torq]

There are two types of “compensation” at work in a typical headphone frequency response measurement (sometimes they get combined into one, depending on the software and hardware in use).

The first is Microphone Calibration. This is simply a way to deal with the facts that:

• Microphones are not necessarily fully linear - they might respond more readily to one frequency than another - those differences from a truly “neutral” response can be quite large.

• Two copies of the “same” microphone may respond differently to each other. You see this with the two microphones in the miniDSP EARS for example, where there are differences in the response between left and right ears, and each needs its own calibration data to normalize their apparent frequency response. These differences are generally smaller than the first type.

Microphone calibration might address JUST the microphone, or it can address the microphone and the test stand/head/model it is installed in. But in either case it is simply a way to avoid the non-linearities in the microphone/measurement devices adjusted to gain a neutral response.

The second is Headphone Compensation.

This is a set of data intended to account for the fact that how the ear (and head) shapes and responds to sound coming from, say, a pair of loudspeakers mounted in front of, and pointing back at, the listener vs. having two transducers that are stuck immediately against the head and may fire straight into the ear canal.

The pinnae and the ear canal, and the head itself, all have the effect of boosting certain frequencies and attenuating others, depending on if, where and how sound waves hit them. If a headphone was built that had a genuinely flat response it would sound all sorts of peaky when actually listened to due to these effects. So headphone designers use all kinds of techniques to boost or attenuate different parts of the frequency spectrum in attempt to get a specific sound at the listener’s ear in observance of the different contributions that parts of the ear make to sound waves that actually reach the ear-drum.

Thus, when you measure a headphone with a microphone all those peaks and dips show up. They’re there on purpose. But their purpose is to shape the sound so that what the listener hears is PERCEIVED as being neutral/flat (or whatever signature the designer is going for).

Headphone compensation is simply a way of, well, compensating for those effects so that the measured response of the headphone looks as close to flat (or the desired target curve - of which there are quite a few, the best-known probably being the “Harman Target Response Curve”) as possible.

The “HEQ” compensation profile with the miniDSP EARS is supposed to yield a harman-like response, for example.

And when measuring IEMs you need a different compensation, because you’re bypassing the pinnae and the first part of the ear canal.

---

Why do I say one of my compensation profiles is experimental? Because it’s a set of data I’m working on, progressively, to provide a closer-to-perceptual-neutral plot than the HEQ one. And at this point in time I’m still experimenting with the values in that compensation data to get it closer and more consistent than it is (adjusting it to get almost flat for one headphone will often throw off another one … which is partly due to limitations in the rig I am using).

As for references, the links I could share require an AES membership to read, but Googling for “Harman Target Curve”, “Sean Olive” are reasonable places to start. As are @pwjazz many posts on EQ (since a headphone compensation profile is essentially a reverse EQ).

Hope that helps!

[Torq]

Going forward, at least once I have finished fiddling with my own headphone compensation profile, I will be posting the FR plots for “raw”, HEQ and my own custom profile in my reviews or impressions.

At some point I may go back and add those to the existing front-page reviews to make them a bit easier to compare elsewhere. Though, as ever, measurements made on one setup are not reliably comparable to those done on a different setup*.

---

*With higher-end tools, a disciplined approach and a methodology that helps to mitigate differences (e.g. doing multiple measurements with multiple placements of the headphone), you can compare a lot more directly. Professional, lab-type, setups make this much more reliable.

But even minor environmental differences (temperature, humidity, a big truck rolling by when you are measuring) can skew things, so the more basic a rig and less-strict ones methodology the less useful such comparisons can become - which makes “at home” measurements harder to relate to each to a useful degree.

[antdroid]

I spent the last two days comparing my HEQ compensations with data from Oratory1990 who has measurements using GRAS since he works in the industry. In comparing a few headphones, I noticed that the HEQ compensation is actually really good up until perhaps the treble region (starting around 4-5KHz, where the MiniDSP has resonance issues). That said, the profile isnt terribly off either. One thing I had to note was his compensation was based off the 2018 Harman curve, while the HEQ one is based off of the 2013 research, as noted on MiniDSP’s website. I actually like the 2013 target more, so this works out well for me. If you read through the papers, or at least the powerpoint presentations, you can see the small differences between the two curves and I had to make mental notes when comparing my data with his.

[Anonimar]

@Torq

If a headphone was built that had a genuinely flat response it would sound all sorts of peaky when actually listened to due to these effects.

That bit is fascinating - never would have guessed headphone designers had to make them with skewed frequency responses, simply so our ears would be perceiving the frequencies as neutral/flat.

Thank you for the high-level overview, extremely informative.