Burst Response! HD800, SR-207, HD650

Give it time. You've have time to study only the old traces. It's like CSDs. We are just getting started.

 

I don't.

This is mostly coincidence.

The SR1a is subjectively "faster" than all these headphones you have mentioned yet has no 1kHz emphasis. Yet it's decay characteristics is faster than all the others in most ways, and its attack characteristics are the most perfect with a slight bit of rounded leading edge at the higher frequencies. Coincidence it's this way? Maybe, maybe not.



The Grados are fast too, yet they have no 1kHz emphasis in their FRs. This is RS2 with TTVJ pads below.

Ding ding ding! This is where I am starting to see some stuff, e.g. Grado above. And again, don't forget decay.

 

@purr1n With the latest visualization you're playing around with, one thing that might help less familiar readers is a reference chart that shows what a perfect response would look like. I could see that being especially helpful for any visualizations that don't necessarily have an immediately apparent waveform sort of look to them. That's not to say readers will immediately understand it because of that, but it might aid over time as more data points roll in.

Yeah, I know, it's simple stuff, but folks inevitably ask for that sort of stuff. Alternatively, in the day and age with a million compensation curves for something as (theoretically) straightforward as frequency response measurements, many have started asking for a reference curve, chart, or target simply because you never know if flat really means flat. So, it might help prevent similar questions in the future.

To expand on these new presentations, it might be interesting to create an animation that shows the waveform being manipulated and transformed into these different visualizations. I don't think I currently possess the skills or knowledge to create something like that, but I have to imagine it's doable.

One other thing, @Serious got me thinking about how much the recording hardware and process might affect results as well. In theory, it should be basically nothing, at least on a basic AD/DA level, but for anyone like myself using a little DIY PSU for their mic, it may be enough of a factor to warrant checking. That is, I shouldn't assume that what a WAV file looks like in Audacity, for example, will match closely enough that file when recorded through my whole measurement chain in a loopback scenario (i.e. no headphones).

And @CEE TEE was wondering if amps might affect results. It would be super fascinating if they did, meaning a possible clue into why certain headphones pair best with certain amps. But, on the flip side, it would also make for a fun headache of a variable to consider.

 

• I'd rather not provide references because then we start playing the HF square-wave game. The best way is to provide a load of data, a la Chang in the early days, and note relative differences and discover patterns that may tell us something. I've already kind of hinted that the SR1a is a good reference.
• No time for animations. Do you want to see these sooner (keeping in mind this is my hobby job) do you want to see them another five years for now because I spent time doing animations to how I arrived at them?
• @Serious will get you thinking to paralysis and doing absolutely nothing. Too much doubt for little gain but only seeing the trees and not the forest. He's the quintessential German, designing their Panzer tanks to be perfect, so complexly perfect that they were barely serviceable in the field. Compared to Russian tanks, that probably had shit for tolerances, but just kept working.
• Expanding on the above, we already know, different rigs = different results.
• Don't expect PhD rigor type work. These are just ideas and quick-and-dirty working solutions. They aren't mean to be in a thesis with equations, proof, peer review, etc.
• I bet amps could affect the results for sure, but I'm betting the differences will be small. Maybe even too small to overcome measurement error factors.

 

Are you surprised at this? I actually feel HD660, HD600, and Elex have similar speed in the lows and mids IMO. Most headphones will be very similar with the attacks from 120Hz to 1500Hz as these are not very challenging signals to deal with.

Also, the Elex being the entry level Focal isn't particularly fast. Maybe less distortion in the mid and upper bass compared to the Senns. More mid-treble. And definitely more edge to the attack in the highs. The HD660S is the most grainy, pricky, heck even grating of the bunch.

On this topic, could @Hands not like this aspect?


And to a lesser extent (the comments on grain)


or maybe this too, but slightly different, as I used "jumpy". Note sharper initial attack, but less energy to the others in by the second-third wave?

 

For the 8" model, the lowest line Voxactiv is 2db more efficient than the TOTL Lowther. Likely the Voxactiv achieves this with the combination of lower mass, stronger magnet, and less damped suspension - higher Qms. We really don't know because TS parameters are not published for these drivers.

 

Exactly. Many factors, interrelated, and no one single factor. Already said this and will say it again. Anyone who thinks that there is one single measurement that will tell us everything or 90% of it should be shot in the head.

300-500Hz FR dip is a good example. Brings about clarity by depressing the "cloud" region. Rolling off the bass may result in less lower quality bass since distorted bass gets in the way of us hearing transients. Adding a treble peak helps with the edges. All these may also increase perception of "fast" transients to varying extent.

 

Need more data. But 'stats tend to do this at many frequencies. The large stretched membrane tends to act like a drum, like those drums heard in the distance in those Jumanji movies. I posit that the fluidity and lack of grain is because of this. There's no hard stop, but rather transients are interconnected with ongoing vibrations, albeit at a very low level. Some of this has been seen in 'stat CSD measurements at Changstar. I'm no longer in the stat game and general ambivalence on my part (mostly the costs and need for specialized amplification) means that I don't have units to analyze. Note that orthos do this too, but in different ways depending upon the ortho.

I also wouldn't describe the ESP95X as having super articulate bass either.

 

Says someone with fast high-efficiency Voxactiv wide-banders and big paper woofers instead of Accutons.

 

Hmm, I am not sure. Hard to say. Would need more data points. The RS2e's results look funky at 4.8KHz too, yet they are a fair bit dark sounding to me with the TTVJ pads and don't seem to share common issues perhaps those other three headphones exhibit. But it's possible any issues it might have in this regard are more covered up by it sounding tonally dark.

 

RS2 is actually different from the others if you look closely. The others have emphasis over steady state at later waves 2-3. The RS-2 has a sharp leading edge at wave 1 and by the time it hits 3, it's leveled off.

I would posit that the later energy rise at waves 2-3 for the 4.8kHz signal could be a sign of this grain, prickly, grating sound.

A spike at the 1st wave could signify the delineation effect heard on various fast planars and stats.

Would be interesting if there's another headphone you can think of and we can get the results of. What did you think about the Australis in this regards? I actually thought the Australis was slightly more evident in this regards to the Borealis.

 

I dunno, the Borealis and Ether are pretty mild looking in that regard. The HD660S, yeah, I see what you mean about how it has more energy for the first few cycles, but I also found that the least problematic of those three headphones in general. 4.8KHz may of course just not be capturing the specific pain points I had with each.

The RS2e still looks kinda gnarly at the start, so maybe it's the droop after telling us more in this case. It could be that the sharp start lends itself to some of that fast attack sort of sound, further "exacerbated" by less energy after a few waves, if that somehow would make it easier for the driver to "shift gears" to something else.

Still, going back to the first 2 or 3 waves in general, in some ways they all have similarities, in some way notable differences, especially looking at the full graph, and I'm not sure I'm totally able to connect the dots compared to what I heard on each. At least, not for the higher frequency graph. (Again, may just be a limitation of not capturing a wide, somewhat granular, and sizable range of frequencies, which would be a real headache to do.)

The Australis...um, I might recommend you read my impressions and measurements. But if you'd rather not just yet, the treble was so, so shelved for me, and it had so much cup reverb, that I couldn't really hear much beyond that.

Do you have anything like the HE-1000 around? Maybe a PM-2 (stock)? MA900?

I also went back and noticed the Elex, which I imagine should sound pretty close to the Clear, right? Its 4.8KHz results are a real head scratcher for me. Trying to think of how I could translate that into what I heard on the Clear. I'd say maybe it fits the guess I tossed out about the RS2e results at that frequency, but I would have guessed the Focals to not look relatively subdued in comparison. I might just be looking at and thinking about these in the wrong way.

 

Processed Elex here. Elex is a bit different then Clear. Clear is muted in the upper mids, but they do share attack characteristics (subjectively). I would not be surprised it the Clear had similar leading edge spike at position 1. Now what might be really interesting to see would be Utopia.

Elex does have similar burst characteristics to Grado RS2 for 4.8kHz, but to a much lesser extent. FWIW, I do find Elex can be slightly difficult with its subjective attack characteristics and really preferred something like the Night Blues Mini to pair it with.

In the case of the RS2 with the TTVJ pads, despite the screwiness, it's what you said, probably the dark FR tilt and all hide this. Even then, there's still that Grado flair with the guitar crunch and whammy bar effect. The mids and highs are still unmistakable Grado.



I think it's important to not see these as the one and final explanation. It's simply more data and complementary presentations to FR, distortion, CSDs, etc.

 

Jumping ahead a bit, I found a tradeoff situation for the DCA Aeon Flow closed. The thickest white filter had the best response up top, however it also killed dynamics and attack - killed as in sucked the soul and life out of music. The best compromise for my tastes was the medium black felt. Highs were hotter, even slightly twitchy, but dynamics and attack were much less crushed. Can you guess which is which?

Measurement A: thick white felt or medium black felt?


Measurement B: thick white felt or medium black felt?