Abyss Diana V2 Review (headphone)

[the_brunx]

I’ve seen so much abyss ads on headfi. Is abyss not a sponser at headfi?

 

[solderdude]

Actually Amir did have a good seal on the right channel and a slightly leaking seal on the left channel.
This can be seen on Jude's 54CA measurements where different seatings gave different results. The 5128 had good seal because of the HATS shape.

A question to Amir could be: was the shown high distortion measured on the L or R channel or did both channels measure the same ?
I often see different THD for L and R drivers and usually pick the worst or show both when they differ too much.

 

[mshenay]

 

[amirm]

Sigh. Here are his distortion measurements at 94 dBSPL with my annotations:

Notice the issues on the graph. First, he measures frequency response using 45CA but doesn't post the distortion. Why? He says this:

"NOTE: On the GRAS 45CA (RA0401) we measured at both 90 dBSPL (1 kHz) and 94 dBSPL (425 Hz), and the measured THD levels at 20 Hz were nearly identical to our Brüel & Kjær 5128 measurements (but there were slight differences through the midrange and treble). "

Oh? Why not let us see it?

Second, the moment your frequency response changes, your ratios change for the same distortion. So you can't make THD Percentage comparisons. He should have post absolute distortion as I have.

He also makes a mess of the comparison by using his scale with log which of course compressed the heck out of the graph vs linear one I use (same as what Klippel uses for speaker measurements).

Here is my quick attempt to match his vertical scale and only showing 94 dBSPL:

Notice how the same peaks exist in critical mid-band frequencies, rising up to 2% just the same. This is alone so damning of the performance of this headphone.

Look at the performance of Dan Clark Aeon RT I just measured:

The blue line which represents the 94 dBSPL hugs the zero line like nobody's business! There is no peaking of distortion at 5 kHz. Here it is using Jude's scale:

Distortion at 5 kHz is just 0.01% vs 2% on Abyss Diana V2. It is 20 times lower distortion! And the RT costs 1/6th of the Diana V2.

I am really disappointed in Jude not calling this out as a high distortion headphone. Nothing like this is acceptable. By papering over this, he is encouraging them to keep building non-performant devices. Jude did the same thing when we had disagreement with Schiit electronic products. Produced a bunch of graphs, confusing average person when reality was 100% on our side. I hope he doesn't continue to play this role and becomes an advocate for consumers, not help companies paper over flaws.

 

[mshenay]

Sigh. Here are his distortion measurements at 94 dBSPL with my annotations:

View attachment 106184

Notice the issues on the graph. First, he measures frequency response using 45CA but doesn't post the distortion. Why? He says this:

"NOTE: On the GRAS 45CA (RA0401) we measured at both 90 dBSPL (1 kHz) and 94 dBSPL (425 Hz), and the measured THD levels at 20 Hz were nearly identical to our Brüel & Kjær 5128 measurements (but there were slight differences through the midrange and treble). "

Oh? Why not let us see it?

Second, the moment your frequency response changes, your ratios change for the same distortion. So you can't make THD Percentage comparisons. He should have post absolute distortion as I have.

He also makes a mess of the comparison by using his scale with log which of course compressed the heck out of the graph vs linear one I use (same as what Klippel uses for speaker measurements).

Here is my quick attempt to match his vertical scale and only showing 94 dBSPL:
View attachment 106186

Notice how the same peaks exist in critical mid-band frequencies, rising up to 2% just the same. This is alone so damning of the performance of this headphone.

Look at the performance of Dan Clark Aeon RT I just measured:

The blue line which represents the 94 dBSPL hugs the zero line like nobody's business! There is no peaking of distortion at 5 kHz. Here it is using Jude's scale:

View attachment 106188

Distortion at 5 kHz is just 0.01% vs 2% on Abyss Diana V2. It is 20 times lower distortion! And the RT costs 1/6th of the Diana V2.

I am really disappointed in Jude not calling this out as a high distortion headphone. Nothing like this is acceptable. By papering over this, he is encouraging them to keep building non-performant devices. Jude did the same thing when we had disagreement with Schiit electronic products. Produced a bunch of graphs, confusing average person when reality was 100% on our side. I hope he doesn't continue to play this role and becomes an advocate for consumers, not help companies paper over flaws.

All things considered I appreciate your posting your findings using the same scale

 

[amirm]

He measured it on both B&K and GRAS rigs (the same one that you use). Please read his post fully.

I did read it fully. He had no business posting a bunch of graphs on BK5128 which has no preference curve or a way of validating its results. Of course you can get different measurements with a different HAT.

As I just post, he did NOT post distortion measurements using the GRAS but instead chose to use B&K. Why? He had the data. Why didn't he post it when it could be a more proper comparison?

As to frequency response, yes he used GRAS but so what? As I said, I can also put bungie cords on the headphones and get tons more bass. That doesn't tell you anything useful. What is useful is verification of measurements using controlled listening test and equalization. I did that, confirming that my measurements were correct. There is no such thing in Jude's.

He also changed all the scales, aspect ratios, etc. making comparisons so hard. All of his graphs are so wide which make every headphone response look good by stretching them horizontally. If your aim is to please every manufacture, that would be the way to go. But if you want to find problems, you need to use proper scales that make comparison easy and relevant. This is why standard such as CTA-2034 stipulate such.

 

[amirm]

I’ve seen so much abyss ads on headfi. Is abyss not a sponser at headfi?

Of course. Right on top:

 

[amirm]

I am going to go to sleep but let me quickly address his GRAS frequency response measurements. This is the second, supposedly more optimized seating:

Notice how he clearly shows a loss of low frequency energy starting at 50 Hz and drooping by whopping 13 dB by 20 Hz! There is no way this headphone produces sub-bass. I measured about 17 dB but that is not what I took to the bank. This is my equalization:

See the Band 1 boost? It is just 8 dB so way more conservative than even Jude's measurement.

Jude shows broad increase in energy around 1 kHz. I put a dip in there 1.2 kHz.

Jude shows a dip at 2.2 kHz, I put in a boost at 2.7 kHz.

And so on. He didn't show you the target curve so you can't figure out what his measurements mean. But now you do.

In other words, Jude's frequency response measurements highly correlate with my assessment. Indeed I could develop my EQ using his measurements and arrive somewhere close to where I am!

Really, it seemed like an aimless effort to spit out a bunch of chart hoping folks get confused and just jump to the punchline that I need more experience. No Jude. You need to verify what you are measuring using listening tests and do more measurements to know what is good, and what is not a good headphone.

 

[sweetchaos]

Of course. Right on top:

Kind of difficult to trust that post, after realizing the blatant conflict of interest.

 

[Dealux]

Sigh. Here are his distortion measurements at 94 dBSPL with my annotations:

View attachment 106184

Notice the issues on the graph. First, he measures frequency response using 45CA but doesn't post the distortion. Why? He says this:

"NOTE: On the GRAS 45CA (RA0401) we measured at both 90 dBSPL (1 kHz) and 94 dBSPL (425 Hz), and the measured THD levels at 20 Hz were nearly identical to our Brüel & Kjær 5128 measurements (but there were slight differences through the midrange and treble). "

Oh? Why not let us see it?

Second, the moment your frequency response changes, your ratios change for the same distortion. So you can't make THD Percentage comparisons. He should have post absolute distortion as I have.

He also makes a mess of the comparison by using his scale with log which of course compressed the heck out of the graph vs linear one I use (same as what Klippel uses for speaker measurements).

Here is my quick attempt to match his vertical scale and only showing 94 dBSPL:
View attachment 106186

Notice how the same peaks exist in critical mid-band frequencies, rising up to 2% just the same. This is alone so damning of the performance of this headphone.

Look at the performance of Dan Clark Aeon RT I just measured:

The blue line which represents the 94 dBSPL hugs the zero line like nobody's business! There is no peaking of distortion at 5 kHz. Here it is using Jude's scale:

View attachment 106188

Distortion at 5 kHz is just 0.01% vs 2% on Abyss Diana V2. It is 20 times lower distortion! And the RT costs 1/6th of the Diana V2.

I am really disappointed in Jude not calling this out as a high distortion headphone. Nothing like this is acceptable. By papering over this, he is encouraging them to keep building non-performant devices. Jude did the same thing when we had disagreement with Schiit electronic products. Produced a bunch of graphs, confusing average person when reality was 100% on our side. I hope he doesn't continue to play this role and becomes an advocate for consumers, not help companies paper over flaws.

His distortion measurements always seem like he's using some kind of weighting. He posted some Sennheiser THD graphs that showed way less distortion in the bass than the headphone actually has.

How did you miss this juicy bit in their reddit post?

How does body heat even interact with the driver lol?

 

[Maki]

His distortion measurements always seem like he's using some kind of weighting. He posted some Sennheiser THD graphs that showed way less distortion in the bass than the headphone actually has.

They also look smoothed somewhat. Would be helpful if he included what type of smoothing and/or the number of data points gathered. Either way, I don't think there's any way to deny this headphone has more distortion than expected for the price point. Then again, the target market is probably putting it on tube amps and god knows what so it probably doesn't matter to them.

 

[xup]

Seems like it. What setup does he use for measurements? Is it documented some place?

I couldn't find it, then again, not really familiar with their, work. Never watched any of their videos. Only thing I know the pads are comfier

Great analysis of Jude's measurements! (Too lazy to double quote on mobile)

 

[highpurityusbcable]

Doesn't look like people care much about what's really happening, it's mostly "WOW THANKS JUDE NOW I UNDERSTAND" and one disturbing guy who posts everywhere about Amir's arrogance while literally getting his elaborate replies about this subject for a few days. I guess that happens when you throw 3k into the garbage bin. Audiofleas are insane.

 

[gegebinazzi]

Am I wrong or are the graphs obtained by Jude on the Airpods Max comparable? (same equipment / method).

The only difference I see is that the Airpods were measured at higher volumes (110db vs 104db max), am I missing something or can we use the two graphs to get a reliable comparison?

Because if that were the case, the comparison seems awkward, perhaps worse than what Amirm measured: the y-axis at 1% in the airpods and at 10% in the 3k abyss is a joke, considered that airpods were set to 110 vs 104!

Am I missing something? I'll paste the two graphs for reference, thanks!

Edit: if you plot the two curves on the same graph with a linear y axis the comparison would be really harsh, basically a line that's the x axis corresponding to the airpods THD and then the Abyss curves going nuts in the graph

 

[MayaTlab]

Am I wrong

Probably not. You'd have to push the APM an extra 20dB at 1khz to reach the max THD value recorded for the Abyss at 90dB. And at a frequency that could be considered less problematic (300-400hz vs. 5000-6000).

 

[the_brunx]

Sigh. Here are his distortion measurements at 94 dBSPL with my annotations:

View attachment 106184

Notice the issues on the graph. First, he measures frequency response using 45CA but doesn't post the distortion. Why? He says this:

"NOTE: On the GRAS 45CA (RA0401) we measured at both 90 dBSPL (1 kHz) and 94 dBSPL (425 Hz), and the measured THD levels at 20 Hz were nearly identical to our Brüel & Kjær 5128 measurements (but there were slight differences through the midrange and treble). "

Oh? Why not let us see it?

Second, the moment your frequency response changes, your ratios change for the same distortion. So you can't make THD Percentage comparisons. He should have post absolute distortion as I have.

He also makes a mess of the comparison by using his scale with log which of course compressed the heck out of the graph vs linear one I use (same as what Klippel uses for speaker measurements).

Here is my quick attempt to match his vertical scale and only showing 94 dBSPL:
View attachment 106186

Notice how the same peaks exist in critical mid-band frequencies, rising up to 2% just the same. This is alone so damning of the performance of this headphone.

Look at the performance of Dan Clark Aeon RT I just measured:

The blue line which represents the 94 dBSPL hugs the zero line like nobody's business! There is no peaking of distortion at 5 kHz. Here it is using Jude's scale:

View attachment 106188

Distortion at 5 kHz is just 0.01% vs 2% on Abyss Diana V2. It is 20 times lower distortion! And the RT costs 1/6th of the Diana V2.

I am really disappointed in Jude not calling this out as a high distortion headphone. Nothing like this is acceptable. By papering over this, he is encouraging them to keep building non-performant devices. Jude did the same thing when we had disagreement with Schiit electronic products. Produced a bunch of graphs, confusing average person when reality was 100% on our side. I hope he doesn't continue to play this role and becomes an advocate for consumers, not help companies paper over flaws.

much prettier.

 

[danweast]

How exciting this is huh?

Company initially mocks the measurements of their product as amateurism,

Then says they're looking for a third party to measure said product for them to back them up.

Ends up asking a site they themselves sponsor to measure it.

Results come out, fine print basically says that the headphones aren't so bad after all.

Yet despite that, the results corroborate the original measurements and outcomes here, even despite possible conflict of interests and completely different measurement hardware.

Uh oh!

 

[Qna89]

His distortion measurements always seem like he's using some kind of weighting. He posted some Sennheiser THD graphs that showed way less distortion in the bass than the headphone actually has.

How did you miss this juicy bit in their reddit post?
View attachment 106201

How does body heat even interact with the driver lol?

it is true that Your body exchange heat with headphones. With every headphones. To be specific You exchange with every thing in Your nearest enviroment... But I can a assure You, If they Don’t even have rig to measure Thd of their own HP they didn’t check how the heat affect their drivers this is pure BS

 

[maxxevv]

Am I wrong or are the graphs obtained by Jude on the Airpods Max comparable? (same equipment / method).

The only difference I see is that the Airpods were measured at higher volumes (110db vs 104db max), am I missing something or can we use the two graphs to get a reliable comparison?

Because if that were the case, the comparison seems awkward, perhaps worse than what Amirm measured: the y-axis at 1% in the airpods and at 10% in the 3k abyss is a joke, considered that airpods were set to 110 vs 104!

Am I missing something? I'll paste the two graphs for reference, thanks!

Edit: if you plot the two curves on the same graph with a linear y axis the comparison would be really harsh, basically a line that's the x axis corresponding to the airpods THD and then the Abyss curves going nuts in the graph

Well, even if you are a conservative listener and listens at pretty low volumes, the 90db curves are pretty telling when you compare between the 2.

The max distortion of the DV2 is some 10 times that of the AP Max at 90dB ! 0.7% versus 0.07% !

And we all know how THD goes with higher output values.

I think if someone put it on a spreadsheet / Mathlab , the extrapolated for equal output curves will tell a very glaring story.

 

[sherm137]

Of course. Right on top:

View attachment 106190

Was just about to mention that Abyss is a major sponsor of HeadFi. This is especially troubling because Jude NEVER DISCLOSES THE SPONSORSHIP. It's unethical.