For those of you who haven’t seen, @Resolve and @Mad_Economist have written and filmed a great piece on SINAD.
It presents a bit of a deeper look at what it all means (while still hopefully being readable).
We also understand that this can be a hot topic so please keep the discussion civilized and happy reading!
Brb, just posting this on ASR
There’s more posted on ASR than just SINAD numbers for products, it even includes subjective listening tests. In fact, his recent review of the Hidizs S3 Pro made it on his list of recommended products even though it’s SINAD is measured at 105 one channel and 95 the other.
I don’t get why all the ASR hate, I think having more info available about products online for you to make your own informed decisions is better than having none at all. It’s like dismissing all the info Crinale has (unboxing, listening, etc.) and just going by the letters he puts ranking products, or antdroids, or others. shrugs
Couldn’t agree more - Amir’s speaker measurements have completely changed the game for people shopping for bookshelf monitors, for example. If you’d told me four years ago that we’d have not one but two people regularly posting independent spin-o-ramas, sometimes more than once a week, I’d have called you a crazy dreamer.
ASR is an interesting community - a lot of great folks post there, and there’s a vast wealth of knowledge available. You don’t see as many significant industry science figures…anywhere, honestly. It doesn’t deserve the hate it often gets. Of course, that doesn’t exempt it from criticism, either, and the over-focus on THD+N/SINAD is definitely an area where I think that criticism is merited.
Thank you and @Resolve for spending the time in sharing your knowledge and writing this up!
I’m glad you liked it! Hopefully there will be more to come, we’ve got a lot of ideas for topics!
You mean like how a single speaker is used to judge a speaker’s ability to image and stage?
That is… How do I put it… quite game changing alright, because it’s complete BS. You can’t get a stereo image and stage from a mono / single point. You can get general sound dispersion characteristics, to a point, sure. I’d love to hear someone argue how that’s the same though, because it’s absolutely not. It’s these practices and methods that call the validity of all of his reviews into question. Just like how he claims that he comfortably listens at over 100 decibels. Nobody comfortably listens to anything at 100+ decibels, but once he says it he can’t back down from it. This is another example of why I don’t believe anything he says - not because he made a ridiculous claim, but because he will deflect, lie, deride and attack people before he would ever admit he might have made a mistake.
You must not have been around for the start of it all then, when Amir’s smear campaign against Schiit started, and his completely inept review of the Yggy, his refusal to allow peer review of the methods and results then ended up with Amir backing himself into a corner, and trying to tell several lies that he got called out on as a result. The ongoing practice of banning people who question or challenge him in any meaningful way (Golden, Resolve, and so many more etc) is disgusting. Having several industry leaders like Rob Watts basically pointing out he’s a hack and why.
But despite all of that, that’s not what pisses me off the most. It’s the absolute closed-minded nature of that culture. The brainwashed masses who thinks he can do no wrong, who will gobble down every little piece of crap product that he shills (like the A90 and 789), and who have drank so much ASR Kool-Aid they are convinced that they are listening to the best music that modern gear can provide. But is it?
There are so many people on this forum who started off at ASR and were absolutely miserable with the music they were listening to, from gear that was highly recommended by ASR. Ask these people what’s wrong with ASR. They’ve been on both sides of it.
I’m actually not a fan of the monaural listening paradigm either - objectively, it’s better for differentiating FR quirks…but it’s not a realistic depiction of speaker behavior.
However, what I mean is his Klippel measurements - having a frequently-added-to database of CEA2034 measurements is incredibly valuable, and between him and Erin we’ve actually go two separate, completely independent sources now. That’s speaker measurement nirvana, as far as I’m concerned.
I’ve talked about this before long ago on Reddit, but Amir’s Yggdrasil review is much maligned and I still don’t understand why, when looking at the actual measurements under discussion.
I will also note that as of present, I’ve spent the last six hours arguing with Amir about this very article without being banned, and I don’t really expect to be. I don’t think he’s acquitted himself well in this argument, but the narrative that he silences all dissent isn’t reasonable either.
Banned or not, I think you’ve proven some of my other points. Or rather, Amir has. In that thread you’ve asked for specific information, and Amir deflects and refuses to provide any supporting evidence of his own point of view or methods, then excuses himself from the discussion before he hangs himself with the rope you provided.
I’ll second that take 
I’m not averse to criticizing Amir or ASR - I mean, I co-wrote an article about it! - I just want it to be limited to reasonable criticism.
I won’t speak for everybody, but for me it was the deceitful way in which he showed single ended only performance because it fit his narrative, and conveniently ignored the balanced output performance of it. On a DAC where even subjective listeners easily acknowledged that the balanced outputs were better, and this was an extremely well known characteristic of the product.
When questioned on this, he said nobody uses balanced anyway, RCA connection is the most commonly used in the industry, and XLR simply doesn’t matter. (Or something to that nature.) I think that’s extremely funny, because in the years to follow the ASR crowd latched on to balanced so hard, you would think RCA connections were a technology akin to 17th century phrenology.
When one of his own posted pictures of balanced output measurements we’re thrown back in his face, everything devolved into a festival of deflections and deceit, we’re at one point I believe he said the AP guys didn’t know what they were talking about and he knew more about audio measurements then they did, even though it was a AP product he was using to measure it…
At that point, the writing was on the wall in so many different ways. It clearly read: Don’t trust this guy. He acts like a cult leader. He is to be trusted, but not questioned.
I really don’t agree with this characterization - the performance of the Yggdrasil on both outputs will be dominated by the nonlinearities of the ADC itself, and indeed that’s what Jude’s measurements show:
Mind you, I think this is a criticism against the site, and I made that around the same time - the Yggdrasil, other than the strongly rising low-frequency distortion behavior that Amir and Bob documented, was performing in-spec, and yet somehow we’re meant to be scandalized that…a DAC meets its (not exactly cutting-edge) specifications? Nobody bought the Yggdrasil expecting four zeros after the decimal point on its distortion products, so I was frankly pretty bemused that that wasn’t the main criticism people had for Amir there: he got a DAC, measured parameters that were in-spec, and then complained about it; he didn’t even need to get the thing in to do that, he could have just screenshotted the specs page!
It doesn’t really help to debunk my point of view when I’m talking about amir’s measurements, and you post Jude’s. Never said I had a problem with Jude’s measurements. I seem to recall there being quite a difference between amir’s single ended and balanced measurements for the Yggy.
Amir’s THD+N measurements match Jude’s - which was the point of my Reddit post.
As far as I’m aware, this is the only measurement Amir posted of both the single-ended and differential behavior of the Yggdrasil
Brb, posting screenshots of this thread on ASR.
Very interesting reading @Resolve and @Mad_Economist, much appreciated.
I don’t work in the audio industry so I’m probably unaware of what deeply moves persons like Amir and similar.
I just don’t like and don’t care about that kind of persons and behaviours. And I avoid both as much as I can.
Paragraph.
We all have our personal things and likes and dislikes.
That given I don’t think some figures deserve so much time and attention.
Have a nice Sunday.
Long backstory: He has a history of making absolute pronouncements based on poor methods, and then engaging in confrontational online conversations. He has damaged the reputations of some products/vendors that never intended to meet his arbitrary measurement criteria. Merely mentioning his name can spark an angry argument in some sectors. Amir has also softened a lot in recent years.
And a vast range of self-serving, biased, naive pronouncements too. I very briefly participated on ASR’s forum (and think I exchanged a couple messages with you). I left and never returned upon realizing how many methodological problems affected the presented data (e.g., see @ProfFalkin’s various comments in this thread). Also, some followers place value on using recordings of tube vs. solid state amps to identify them. People do routinely fail blind amp comparisons, but the “listen to a recording” method is broken from the start.
First consider that amps are widely considered to have a modest impact versus other components, so valid testing requires extreme methodological control. You cannot expect to ID an amp when (1) its output is filtered through some arbitrary mic and ADC – each with its own performance potential and each less than 100% accurate, (2) the mic placement in a particular recording room (e.g., standing waves) may reinforce more significant speaker characteristics or cancel out genuine amp differences, (3) the playback of the recording by end users to ID the amp is fully constrained by the composite performance potentials of a separate playback amp, DAC, cables, and set of transducers – all less than 100% accurate.
Playback of recordings is a bloody broken “scientific” method from the start, and it reveals the wishful thinking of those who set up or accept the results of such tests. This conflates about a dozen relevant factors in advocating a desired conclusion: “there are no audible differences between tube and solid state amps with blind testing.” Many, many experienced audio reviewers don’t bother playing audio samples in their reviews because “playback only demonstrates the potential of a phone speaker or TV soundbar.” There is an art and science of recording drums alone, whereby moving the mic an inch here or there, setting the mic close versus farther away, etc. has a major impact on the end product. Mics are speakers in reverse…playing a recording of an amp through a speaker involves two additional drivers…confounding factors…useless data…no conclusion possible…science 101.
Sigh, yes. I bought the 789 in part due to its measurements. Upon comparing it to live performances (e.g., acoustic string quartets) and other amps, I suspect the measured performance is an artifact and that THX actually creates “subtractive distortion.” This speculation is subject to research and testing. What I think I’m hearing is that by attempting to aggressively remove amp noise, THX technology cannot distinguish between genuine, natural, and desirable recorded noise (e.g., the haze and air from playing an instrument or singing), versus the unwanted noise resulting from hardware flaws. In taking out everything that resembles noise to improve measurements, THX leaves only the etched hard edges that persist above the now deleted valid noise. Genuine testing requires a non-circular method: “Why is this amp better?” “It tests better, so it’s more accurate.” “How can that be, it sounds like ass, like a machine, and nothing like a live performance?” “It’s better because the test rig numbers are better.”
IMO, the “subjective” and “objective” debate follows from limited appreciation of perception science all around. Beyond electrical measurements, it’s essential to incorporate human biological characteristics and inherent biases for relevant testing and measurements. I’m making good progress in my write-up of a way forward – it’s meant to be a starting point for systematic efforts. [BTW, all of this is old hat science but not common in the audio community. @Resolve’s explanation in the video above goes in that direction with auditory masking, as the tip of the perception iceberg.]
Then what you are saying is there are no limits! He, he. (Yes, I’m joking.)
I think he’s very confident and convinced of his position and methods. His followers want to believe and they form the culture. What they don’t understand is that many “sometimes fuzzy but objective” human perception science methods go beyond APx555 test rigs. These methods are actually relevant to human biology with their nonlinear and arbitrary hearing characteristics. The necessary test methods are also very complex and tedious relative to plugging a device into a test rig and running a canned set of electrical tests. If you want superficial and easy but misleading “objective” answers, enjoy your ASR bliss and masochistically punished ears but walk away from full scope of relevant science. If you are happy with your current equipment, just turn up your favorite music and relax.
So there’s something with this whole discussion that I think needs to be clarified, and I’ve gone back and forth on posting this to ASR but figured I should at least post it here.
It’s true that Amir and I would likely have some core philosophical disagreements about certain things, and part of this article/video was aimed at expressing concerns over one aspect of his information publishing (specifically SINAD). But at the same time, I want to make it clear that the intention here wasn’t to try to discredit his work or attack his motivations - even though very obviously I do disagree with his commitment to SINAD. But the point is that the intention with this piece wasn’t to ‘go after’ the person measuring and publishing the data, nor the people who enjoy reading it or find it interesting. That’s one thing.
The main motivation for this was rather to deal with the more common perception that’s resulted from the publishing of SINAD rankings - something I don’t think core members of that community take all that seriously or perhaps even realize has occurred. Simply put, SINAD scores and rankings have created a kind of feedback loop, where people will make purchase decisions based on it, without putting much thought into anything else. Then in order to justify their purchases, they tout the SINAD score as a kind of validation for their decision over someone else’s. You can imagine the thought process goes something like “science people rate this higher on a ranking and therefore it must be better” - or even worse “it must sound better”, which as we’ve shown is not necessarily the case. This is what I end up having to deal with quite commonly, and you see it happening in many other communities as well - including this one.
I don’t think the more reasonable members of ASR (or other measurement focused communities) would find what we said to be all that controversial - in fact, when we were working on it I even mentioned to Blaine that it felt like what we were expressing was a bit of a platitude or low hanging fruit. And I think it kind of is. But it’s also something that needed to be clarified for that other group, or anyone who maybe hasn’t actually bothered to look into it beyond the ranking. As I mentioned in the video, none of this is a criticism of conducting or making use of measurements. We love that stuff too. And when it comes to the spirit of understanding more about sound, science and how all of this stuff works, I really think there are good intentions on both ‘sides’ here.
Now, as far as my take on that community, I think it really is a bit of a duality of those two groups. You have the generally reasonable discussions going on where people want to better understand sound/audio/equipment and so on, and then you have those who seem to be more interested in the tribal aspects of online communities, and that’s where I tend to see a lot of attempts to discredit people or throw ad hominems around. What’s funny is that in person, I imagine a lot of this stuff simply wouldn’t exist.
In any case, I’m under no illusions about what our recent SINAD piece did not accomplish, and it’s any grounding for subjective takes on source equipment. But it also wasn’t intended to be that. It would have been quite easy, and simultaneously a spectacular failure of conversation, to simply dismiss SINAD for subjective reasons. Nobody focused on the more objective side to this subject would be persuaded or even interested in what I’d have to say on that.
But on that front, we are working on doing ABX comparisons, which is something that has the potential to reveal a lot about subjective evaluations - including the limitations to that and where we might get it wrong. In fact, if it turns out that all the subjective findings are wrong, then that on its own is interesting and worth publishing.
Hi Resolve!
Hope the kickback isn’t dampening your morale any; regardless of what the SINAD piece did or did not accomplish, this conversation has needed to happen for a while.
Could you elaborate some on your upcoming ABX work? It would be interesting to test a larger population of enthusiasts/professionals to try and feel out the audible limits of harmonic distortion. Outside of strictly SINAD-relevant discussion I can think of about a hundred other excellent ABX tests you could run. Exciting stuff 
His first 2 HBO specials were comedy gold and I will meet you in the parking lot if you disagree
