Audio foolery 2.0 - The rise and fall of objectivism

[nimar]

Then why would you expect it to sound better?

Never claimed I thought it would "sound" better, simply that one would expect it to be better designed / executed as a complete package. I know full well that you can get the "sound" from a $99 DAC, or < $50 from a hifiberry, or a < $10 headphone dongle depending on power requirements.

 

[nimar]

This is an important point.

The limiting transducer in the replay chain acts as a gate. The highest noise floor is what we can hear. For those gentle readers closer to retirement than our first day - our hearing inexorably deteriorates.

It's still a tough call - higher performance for a small additional charge.

I liken this to owning a high performance car to haul groceries.

I thought a little about the car analogy. Sure from a numbers perspective brand A can go 500 Kph, that's grand. But the competition can go 400 Kph (still faster than legal / required) and it has air conditioning and a working sat nav.

Why would anyone promote a metal box that goes 500Kph that does everything else a car is meant to do poorly?

 

[Spkrdctr]

Nothing there suggests one sounds better than the other, they are just ranked by SINAD. Okay, so you have to know what that is and what it means and that's not explained on the chart, but just a little research would reveal it.

It's not a buyers guide, it doesn't claim to be. If someone interprets it that way, that's their mistake. I know we live in a world where packets of nuts have a warning on them saying 'May contain nuts' but does everything really have to be spoon-fed to us?

Sadly "YES".

 

[Spkrdctr]

It seems so, for some reason I thought a brands flagship model would be better executed. That there was some value in getting the best they offer over their $99 product. That I wasn't paying ~10X to have balanced outputs, turns out I was wrong.



I did the research and I _do_ understand the data. This is how I am here now, writing of this sad story of misadventure. As you, and if it wasn't clear I have said. The right information is very much inside these walls, but it takes quite a bit of looking to find. While at the same time there is very much an air that better number == better product.

Take the Eval-1 for instance, because of the review suggesting bypassing the buffer stage results in a 5db better number. I guarantee you that some not insignificant number of people went and pulled those jumpers out without a moments hesitation. Not considering that it also bypasses RF filtering and that their DAC doesn't have nearly enough power to provide the ~10V required to reach max output. Likely leading to a worse overall experience than just leaving the thing as is.

I appreciate the take it or leave it mentality; that a review is a balanced presentation of the numbers that can be easily measured but whether intended or otherwise that's not seeing the whole picture. If something gets a SINAD of 105 its "recommended" and 120 its "highly recommended" why?

The current state of things reminds me a lot of a story from Solzhenitsyn about soviet manufacturing. To start the factory overseers decided to quantify the productivity of a factory producing different types of nails based on the number of units produced. The managers at the factory, inspired by this bad incentive realised that if they only produced tiny nails that they would be able to triple the number of units produced and they would likely get promotions or better for a job well done. The overseers found the factory was producing only tiny nails that hardly anyone needed and decided instead that productivity would be measured by the weight of nails produced. Another bad incentive, this time the managers switched to only producing railway spikes, by the new measure production increased again but again the factory made something that hardly anyone needed.

Perhaps something akin to a preference score needs to be developed for DAC's as in speakers. Where features can provide a + / - number used in addition to raw measurements to come to an overall number. This way manufactures would at least have an incentive to design good products.

I agree. The raw data is nice, but ranking things according to performance is very helpful to those who can't put in the amount of work do learn all this stuff and then apply it. Maybe all they are looking for a a pair of good speakers? A nice AVR etc. A VERY useful idea is to rank say speakers for an example in dollar brackets. Such as best performing $500 a pair speakers, best $1000 a pair etc. The dollar amount can be anything. But when a speaker for our example, costs say $2000 a pair and performs poorly compared to a $1000 a pair, that is important news. Now if this is being done already, then I apologize for not knowing it. I'm just throwing out an idea or two.

 

[BDWoody]

Never claimed I thought it would "sound" better, simply that one would expect it to be better designed / executed as a complete package. I know full well that you can get the "sound" from a $99 DAC, or < $50 from a hifiberry, or a < $10 headphone dongle depending on power requirements.

Why I started buying more vintage gear.

Can't you explore functionality through the manual or other reviews? I'm not sure I'm fully understanding your issue. You bought based on reviews here, which are based on measuring fundamental engineering competence. They are fundamentally competent. Now you feel the reviews are incomplete because you don't enjoy them, for whatever reasons, as much as you had expected?

I'm back to wondering what you expected in the first place.

 

[Spkrdctr]

I thought a little about the car analogy. Sure from a numbers perspective brand A can go 500 Kph, that's grand. But the competition can go 400 Kph (still faster than legal / required) and it has air conditioning and a working sat nav.

Why would anyone promote a metal box that goes 500Kph that does everything else a car is meant to do poorly?

Yes, certain Italian performance cars!

 

[BDWoody]

The raw data is nice, but ranking things according to performance is very helpful to those who can't put in the amount of work do learn all this stuff and then apply it.

There are rankings here...

Learning is what we are about, not as much the spoonfeeding. Once you learn, you don't need anyone's rankings.

Performance beyond what is measured is very personal. The objective performance isn't personal, it just is.

 

[Frgirard]

The subjectiviste audiophile can not live without issues. He is ready to invent them even if they doesn't exist.
The room and the speakers It's all that matters.
Objectivly.

 

[BDWoody]

I think we're on well covered ground here...
We can let this thread pass into history without it being badly missed.

 

[BDWoody]

I think we're on well covered ground here...
We can let this thread pass into history without it being badly missed.

I was wrong! Reopened due to respectfully registered popular demand!!

 

[Maki]

This is more of a system philosophy question IMO. Personally I would never consider any of the DACs with roon streaming, EQ, or whatever - because I approach my audio system with the UNIX philosophy: I want each box to do one thing and do it well. All DSP and processing is done on the software side. Other functionality like switching, is built into other physical boxes. For the DAC I want it to have balanced outputs and maximum noise rejection and JUST convert the signal to analog. For my amplifiers I prefer them to do JUST amplification - that's it, no need for passthrough, volume control or anything else, though I have found the preamp functionality of the A90 to be very useful (although I'd still prefer to have a separate preamp, though I can't justify the price at the moment). If I want streaming I'll buy a dedicated streamer (probably rpi) and hook that in. If you subscribe to the philosophy of having everything in one box, then of course something like the RME makes a lot more sense than a Topping stack.

 

[restorer-john]

Never claimed I thought it would "sound" better, simply that one would expect it to be better designed / executed as a complete package. I know full well that you can get the "sound" from a $99 DAC, or < $50 from a hifiberry, or a < $10 headphone dongle depending on power requirements.

And the HiFi Berry/Dongle combo is a basically a kid's science project in a few plastic boxes on your bench. It's not HiFi for grown ups. Fine for a teenager's bedroom, along with the RGB LED light strips, subwoofers, an unmade bed and clothes all over the floor.

Beautifully made and executed HiFi gear is a joy to use and own. It stays with you for decades. No wiggly knobs, hard to read switches or menu buttons that perform a hundred functions by watching the number of times a microscopic LED flashes.

Topping has come a long way fast. They, of course play the numbers game with design, because they are in the business to make money and save costs. They will get there eventually in the quality stakes, but not for a while. A Chinese Accuphase they are not.

 

[JeffS7444]

• Poor UX from top to bottom
• Remote is silly, several buttons do nothing, have to be very particular when pointing it
• Pops when turning on / off
• Hardly any programmable features, can’t set default input, can’t rename inputs
• Doesn’t do volume control over USB, not even one way.
• No headphone out, requires separate unit
• Doesn’t adjust volume when switching between speakers / headphones
• Draws 2.5W of electricity even when mechanical switch on the front is set to off?

It would be great if these things were considered and figured into final test scores: Got an extra 40 hours in your week to help out?

 

[nimar]

Why I started buying more vintage gear.

Can't you explore functionality through the manual or other reviews? I'm not sure I'm fully understanding your issue. You bought based on reviews here, which are based on measuring fundamental engineering competence. They are fundamentally competent. Now you feel the reviews are incomplete because you don't enjoy them, for whatever reasons, as much as you had expected?

I'm back to wondering what you expected in the first place.

Thanks for reopening this. I was in the middle of trying to explain my original thinking.

My issue is not with the reviews on this site per se but with the current state of objectivist thinking. I want to be entirely clear that I am in no way suggesting or expecting one device to sound better than another (unless one is broken). Simply wondering aloud if we could expand on the numbers that are already measured to try and create a somewhat more holistic objective system for grading devices.

Also if it wasn't clear, I was the fool in the title. I jumped into a product range early in my reading of ASR's deep archives with some flawed assumptions. Namely that good engineering in one domain, doesn't necessarily translate into good engineering / design in other domains. But I made it out again to tell the tale, and I think it is overly simplistic to state that the information is there to be had and not to be spoon fed. Not everyone has the time or inclination to read the research papers, or even Toole's books. ASR is a community but also a repository of useful information, a lot of which (for better or worse) is commonly used to inform purchasing decisions. Shouldn't we attempt to provide the most useful information we can?

 

[nimar]

This is more of a system philosophy question IMO. Personally I would never consider any of the DACs with roon streaming, EQ, or whatever - because I approach my audio system with the UNIX philosophy: I want each box to do one thing and do it well. All DSP and processing is done on the software side. Other functionality like switching, is built into other physical boxes. For the DAC I want it to have balanced outputs and maximum noise rejection and JUST convert the signal to analog. For my amplifiers I prefer them to do JUST amplification - that's it, no need for passthrough, volume control or anything else, though I have found the preamp functionality of the A90 to be very useful (although I'd still prefer to have a separate preamp, though I can't justify the price at the moment). If I want streaming I'll buy a dedicated streamer (probably rpi) and hook that in. If you subscribe to the philosophy of having everything in one box, then of course something like the RME makes a lot more sense than a Topping stack.

I agree entirely in your summary of the unix approach. Though it isn't precisely the point I was trying to make. If a DAC box was all one wanted then one should purchase the D10. The D90 is marketed as a flagship model, and regularly touted as "a giant killer", suggesting you should get more than a D10 out of it. If all you are really paying for are balanced outputs then ok, but that seems like very bad economics when other companies can provide a complete package with excellent numbers for less money.

It would be great if these things were considered and figured into final test scores: Got an extra 40 hours in your week to help out?

Totally, there may be some additional effort involved, but that isn't in itself a reason not to do it. Perhaps the "official" reviews could be augmented by user contributions similar to how a helpful user often calculates the speaker preference score. The question at hand if what features would be good to standardise around and agree represent value.

There are some very low hanging fruit like including power consumption in use / in standby. Not only is this "green" but it likely directly correlates to the longevity of a device as the more time full power is being pushed through caps etc, the shorter their lives will be, leading to earlier degradation in sound quality.

 

[Atanasi]

Personally I would never consider any of the DACs with roon streaming, EQ, or whatever - because I approach my audio system with the UNIX philosophy: I want each box to do one thing and do it well. All DSP and processing is done on the software side.

In order to connect devices that cannot do their own DSP, there has to be a way to receive digital signals from those sources and do the DSP elsewhere. In consumer applications the interconnection would be SPDIF. If you receive SPDIF with a device distinct from the DAC, that introduces the problem of different word clocks.

RME ADI-2 DAC has the ability to receive SPDIF, and it can either do DSP internally or forward the signal to a host computer that can do more complicated DSP. The host computer would send back the processed signal, which the DAC would convert to analog. The process is synchronized to the word clock of SPDIF.

UNIX philosophy depends on the ability to compose parts, and in audio that often means the ability to synchronize word clocks. A simple DAC doesn't typically support external clocking.
If there is only a single host computer that does its own DSP and has a single DAC, that would not implement UNIX philosophy, because there would be no ability to compose.

 

[Maki]

In order to connect devices that cannot do their own DSP, there has to be a way to receive digital signals from those sources and do the DSP elsewhere. In consumer applications the interconnection would be SPDIF. If you receive SPDIF with a device distinct from the DAC, that introduces the problem of different word clocks.

RME ADI-2 DAC has the ability to receive SPDIF, and it can either do DSP internally or forward the signal to a host computer that can do more complicated DSP. The host computer would send back the processed signal, which the DAC would convert to analog. The process is synchronized to the word clock of SPDIF.

UNIX philosophy depends on the ability to compose parts, and in audio that often means the ability to synchronize word clocks. A simple DAC doesn't typically support external clocking.
If there is only a single host computer that does its own DSP and has a single DAC, that would not implement UNIX philosophy, because there would be no ability to compose.

I have no devices that cannot do DSP, nor am I interested in acquiring any - makes that problem easy. If I want to hook up another computer, a simple USB switch will suffice.

 

[tmtomh]

Everything in the OP makes sense except the thread title, which is click-bait nonsense.

TL;DR - the user experience is important, and even the most dedicated audio objectivist should - and based on the comments in this thread, does - agree that once you achieve a certain level of measured performance, other factors like usability come to the fore.

On a related but different note, I know I shouldn't take the bait, but I'm going to anyway: if we're going to talk about "hi-fi for adults," @restorer-john , then that should also include thinking twice before consuming lots of electricity running big, inefficient amps and promoting the environmental impact of producing and shipping big, heavy equipment all over the world when lighter weight, more energy efficient equipment exists that will perform just as well. Of course "big iron" that lasts for 40 years is more energy efficient than lightweight junk that has to be replaced every few years, and so I totally agree that repairing, restoring, and maintaining older, well-built equipment is a great thing to do. But the choice is not between "adult" big iron and "kids' toy" trash. There's plenty of well-built, well-engineered, user-friendly, high-performing equipment out there that you've implicitly tarred with your simplistic binarism.

 

[nimar]

Everything in the OP makes sense except the thread title, which is click-bait nonsense.

TL;DR - the user experience is important, and even the most dedicated audio objectivist should - and based on the comments in this thread, does - agree that once you achieve a certain level of measured performance, other factors like usability come to the fore.

Hah, thanks for the positive / negative criticism, one thing I like in life is balance. My intention with the title was to describe a journey and critique the end result of taking objectivism too far. I the fool in question, took a little objectivism too far, and came crashing out of the other side with the understanding that what is needed is not less objectivism but perhaps more subtlety in ones approach. But I doubt I am the only person that has fallen down the trap of "audiofoolery 2.0".

 

[ctakim]

Glad you find your favorite. The RME ADI-2 Pro is the best electronic device I've ever owned. Been keeping using it for years and still love it. My only complaint is the headphone amp is prone to burn out, and the encoder may fail, and the repair cost is high. Just keep those in mind.

ASR's measurements are written in a way to make you believe their benchmark is the "right" rubrics to grade a device, which is not necessary the case at all:

1. It never put usability and reliability at top --- Most of their top rated devices fucked up seriously and in a lot of cases the companies who made them refuse to take responsibility (Example: Topping DX3 Pro version 1). Most Topping DACs I have tested are not reliable in that it may stop playing in a middle of track. This never happens once with professional audio devices (RME, UA, Apogee) that I used. I have tested more than 10 top performing products from China in ASR and returned all of them within Amazon's return window. I simply cannot use them --- weird design choices, serious QC issues, and in many cases simply do not work.

2. The measurements are by no means complete. Luckily I have a ADI-2 Pro so I can do simple measurements by myself. Top rated devices usually only perform good under @amirm 's test, but fail in the most simple and straightforward test. Example: SMSL M500 performs really bad unless it is playing mono music. When it is playing stereo, the SINAD is well below 80. standard tests amir uses just don't tell that part of the story that is more practical to normal music listening. In addition, companies these days are super good at achieving good scores in amir's test in a way that may lead to problems or simply don't make sense (Example: Topping's HP Amp's input impedance. SMSL DAC's elimination of output buffers. those tricks can make them top performing, but lose the ability to work together with other products). And let's be practical, for music lovers, who cares about things that beyond audible? And why should I own a product that has the ability to output 6W (which is not a good thing) to my headphone and produce a lot of heat constantly?

3. Official description of many top performing products are just scams. Example: SMSL M500 claims they use OPA1612 but they are not. confirmed by my teardown and other's.

I would not be so quick to dismiss Liu's points as there are truths in this statement. The top ranking Topping L30 Headphone Amp fried a few users headphones leaving them permanently damaged. That is a huge flaw in my book and it underscores that SINAD performance is but one metric of engineering. It was an eye opening moment for me and I own Topping gear. Caveat emptor!