Oh shit! SINAD can suck my ****s

I registered here just because of this thread. Back in 2011 or so I used to browse head-fi quite a bit and now, after some years out of the audio loop, I landed on ASR (thanks Google). While I am not 100% on board with everything they do, especially when it comes to the headphone measurements, I do find the forum has had an overall positive impact on the industry, as its measurements have demystified a few things. However, I do not believe the measurements being done at the time are the end-game, as seen on the M500/SU-9 saga. That's why I found this thread interesting, as I thought it may bring something new to the table...but does it?

Amir actually posted proof that he had measured this same phenomenon back in 2018, but stopped after realizing the measurements were faulty/didn't achieve what they were supposed to https://www.audiosciencereview.com/...e-amplifier-review.15226/page-242#post-837357

Several people here and over at ASR have hinted at a faulty test setup and inaccurate interpretation of the results. Maybe this isn't as groundbreaking as initially thought?

 

The finding of "issues" is not groundbreaking. However, working through these things, making sense of them and, finding solutions to the problems/questions they pose can shed light on previously unknown or not well understood attributes of amplification (for example). Perhaps people like Nelson Pass and others know exactly what's going on. But that knowledge doesn't filter down to the general public. Or at least not in an easy to understand way. I noted Nelson specifically because he shares tons of knowledge via the DIY audio forum and various publications.

How cool would it be to gain an understanding of how/what measurements correlate to subjective experiences that people have with gear?

I'm super pumped by what Marv's doing not because it's easy but because it's hard. When's the last time Amir or anyone at ASR bothered? There's no curiosity over there. There is no quest to learn or uncover. Just some incompetent "measurements" of a few things that anyone who spent 10 minutes doing research can tell you are next to meaningless.

As a bonus, find and read the ASR Review thread on this site for some illumination.

 

@Marvey Maybe grabbing something like this might make things easier for future testing? (And remove much of the variation caused by the headphone itself from run to run) https://neurochrome.com/products/headphone-dummy-load

True reactive load box with selectable impedance and parallel capacitance characteristics. Also controllable by APx so can be automated too.
I've ordered one as I need a headphone dummy load and having the ability to do reactive load testing in a consistent manner is great.

 

You could read the linked pages in the post I posted (woah redudancy) and see that several headphones were tested, as well as several headphone amps. If anything, I'd argue you have less data points, as you only compared two amps with different outputs (SE vs Balanced) and one headphone before proclaiming this was huge and SINAD was dead (though afterwards you did add more devices).

I did read this. I read this whole thread, yeah. And I agree more testing is needed. However, both JohnYang's and your latest test showed a different behaviour, didn't they? This after modifying the test setup a little bit. This pretty much aligns with solderdude's explanation over at ASR, which he provided before JY posted his measurement.

I'm thinking for myself. It's why I'm posting both here and over at ASR. My field of expertise is unfortunately not within this realm, so I don't have the technical know-how or tools to do the testing.

Also, I really don't appreciate you calling me a troll. My first post was very honest and sincere. I gotta say I did debate with myself for around a week before registering, because I feared something like this may come up, because I do like some snark here and there.

I don't usually do that, but sure, just for you.

Well if it hadn't been for ASR, you probably wouldn't own an AP, would you? Neither would you be doing this kind of testing. Back in 2010, when I used to post on head-fi, I really don't recall there being much technical talk. It was mostly verbouse descriptions of sound. I'd argue the impact has been big and mostly positive.

Absolutely. Finding issues is the easiest part, finding a cause and eventually solving them is the hardest step. I do this on a daily basis, I should know.

I'm all for people finding issues and holding companies accountable, just as the banned ASR member liu did on his thread about the SMSL DACs. But, from what I've seen so far, I don't think this is really as big as it was prooclaimed to be in the first post. People both here and on ASR have pointed out the issues, which to my non-electrical-engineer eyes, seem very plausible. Moreover, modified testing methodology as proposed by JohnYang at ASR seems to change the outcome significantly.

 

It's forum rule and etiquette. No one is asking for your biography and personal details. Read others' intros for examples.

 

No worries. I posted already

 

There are six mentions of another forum in your third post in this thread. But alright.

 

Why on earth would asr have anything to do with whether other people are/are not testing stuff?

 

I was posting technical measurements on both head-fi and changstar long before ASR existed with both dScope and Klippel analyzers along with an assortment of other analyzing equipment. I would also point out audio analyzer results are from stationary measurements. This limits the ability to infer dynamic performance. A high resolution vertical axis oscilloscope observing step response provides some insight into component dynamic behavior. I provide those but I don't see this measurement from others.

 

That looks to be a great device for the lab. However there are 42 combinations of R, C and no L. Consider a standard suite of 20 measurements. That would allow for 840 possible measurements. How much effort to post such a quantity of data and would anyone read it?

 

Maybe the distortion can be approximated with nonlinear parts. Like cheapo inductor with an iron core so small that would blow Dannie Richie's shirt off. Tantalum/ceramic capacitors might also work. It would be hard to match the same distortion level to an actual headphone though.

 

Yeah I'm not sure what the exact impedance curve it offers is, but can find out once it arrives (Tom from neurochrome may already have charts so I'll ask him).

And yeah, it wouldn't be done with the full set of combinations. I'd likely pick a low and a high capacitance option for 32, 300 and 600 ohm.

Maybe set a full sequence to do as many as possible which would probably take all day. But hey, with the automation the APx offers and the fact that this box is controllable from the APx software, honestly that's not unfeasable. If the sequence is done properly I could literally set it up, hit go, go to bed and the full report will be waiting in the morning.

How useful that is or whether anyone would read it I don't know.
I'm still trying to work out what 'normal' test sequence I want to do for stuff I get in. Because as @atomicbob hinted at, there is a balance between more information, and having it actually be readable.

I want to have thorough tests of everything I get in and certainly not boil everything down to a silly SINAD number. But also want to make sure that it is at least somewhat accessible to the average reader even if with a little help