Card carrying objectivists

[svart-hvitt]

On the use of nazi in modern era language:

 

[Blumlein 88]

@Blumlein 88 , flat in-room vs open field is a great distinction. I think that distinction often gets lost in translation.

On subjectivism vs objectivism: Subjectivism is important to guide audio research in the right direction, to make priorities. And here I see a sign that I take as unintelligent subjectivism masquerading as intelligent objectivism. On this site, the interest for measurements of DACs (and other boxes) is high and generates great traffic. But how intelligent is it to focus so much on a DACs ability to suppress noise and distortion below audibility?

An alternative, more intelligent approach could be to use subjectivism research in order to make a ranking order of priorities for research focus. So if acoustics and psychoacoustics are more important than say 120 dB SNR of a DAC, focus should be shifted away from measurements below audibility to what counts more; always based on empirical research of subjects’ preferences, of course.

Amir does a great job smoking out a few of the schitty producers of audio gear, but is focus too much on what can easily be measured, cfr. the drunkard searching for his keys below the street light because that’s where the light is?

This is written in good will. I think this site is great so this is just to provoke reflections on what good is objectivism without understanding the subjects’ preferences (and thus resulting priorities in intelligent audio research).

I think I see where you are coming from on this. First things first maybe.

My own idea on DACs, they are incredibly popular to talk about, so many are made so many different promises. Probably the least important part of the chain as so many are good. And yet surprisingly many have failed to live up to promises. Also like the recent info on how some headphone amps fall apart driving headphones is important. Shows you have to do your measurements mind the details etc. A good topic to get people interested in measuring gear though.

But past a certain point I simply see no way they can be audibly different. I mean a noise floor of -123 db instead of -108 db? If my speakers max out at 108 db SPL I'll never hear either one. i certainly don't think people want to pay extra money for less performance than cheaper gear. So if -123 db is the same price sure why not. But it isn't important anymore is it.

Now where I doubt much gear is so pure is amps drivings speakers and headphones. Worthwhile to do measures of those.

I do find myself fuzzy on the details of what you are proposing. I don't think anything other than amps a little and speakers/headphones making a difference. I see how it could go the other way. What is the minimum performance to sound perfect. Speakers and headphones are not close, and everything else might be far beyond sounding perfect.

As for DACs the Schiit gear may be the only one that sounds different than any of the others. It does so because it performs worse.

 

[Cosmik]

Some of this discussion seems to me to go like this:

"We have been looking into some facts, figures and measurements for various countries round the world. We found that the happiest country, as indicated in a rigorously objective survey of a sample of the population, is on most measures unhealthier, more dangerous and has shorter life expectancy than our own. It is clear that to make our country happier we need to introduce measures to reduce health and life expectancy, starting with a gently graduated cull of the old folk when they reach 76 - to duplicate the life expectancy statistics of the happiest country in the world".

The tail is wagging the dog. Dubious, selective measurements are being used as justification to 'correct' the outcome by brute force, rather than asking what, logically, from the ground up might lead to a population of happy listeners.

 

[Jakob1863]

I second DonH56´s concerns about using something like "Nazism" in forum disucssion about audio reproduction and imo its usuage should be restricted to real "Nazism" . There is a distinction between totalitarism, authoritatism and "Nazism" although there are of course shared ideas among these.

But as i already wrote in an answer to ThomasSavage (that part got lost in the thread splitting), if an card-carrying objectivists isn´t able to (or doesn´t want to) question his own (prior) beliefs then he is no objectivist anymore.
In another forum i learned about a state of mind that is called bulverism . It describes that people are so obsessed from the idea that they know about the TRUTH, that they can´t evaluate any arguments to the contrary (because they alread "know" that the contrary must be false) but instead only speculate about the hidden motivation of their opponents to maintain such "false" positions.

https://en.wikipedia.org/wiki/Bulverism

 

[Theo]

About the flatness of system response, this is the typical average spectrum of recorded music :

As far as flatness is concerned, it shows an attenuation of 20dB from 1 to 10kHz, and even more beyond 10kHz. So, it does not surprise me that some sloping is prefered. It would be interesting to know if preferences varie among people listening to different kind of music, as they would be used to different spectra...
About the "unintelligent subjectivism", does this refer to sighted vs blind tests? I would personally consider blind testing as "intelligent" subjectivism. Isn't subjectivism the domain where the listener brain comes into play?

 

[Blumlein 88]

The tail is wagging the dog. Dubious, selective measurements are being used as justification to 'correct' the outcome by brute force, rather than asking what, logically, from the ground up might lead to a population of happy listeners.

Better marketing combined with better drugs should do the trick. Have a better idea?

Schiit owners are not just happy but super loyal. Identify with the brand and love it. It ain't about the sound performance. I say mostly the story, idea, the marketing.

 

[Cosmik]

Better marketing combined with better drugs should do the trick. Have a better idea?

Schiit owners are not just happy but super loyal. Identify with the brand and love it. It ain't about the sound performance. I say mostly the story, idea, the marketing.

I'm specifically talking about the idea of thinking that 'objectivism' at the listener's ears is the answer.

What reaches the listener's ears is, like a real musical performance in a real venue, different from what was made by 'the source'. Attempting to 'correct' what reaches the listener's ears in some arbitrary room is like staging the concert in a completely different venue and then attempting (with bits of cotton wool, sellotape, rubber bands, blankets) to modify 'the sources' until one single measurement at the listener's ears matches that at the other concert.

Even if this were logical (it isn't), the single 'measurement' is just a load of buckets into which you catch 'frequency components' in any old order and then look to see how much you have collected at the end. If you change something like the time period over which it is collected, you won't even get the same answer!

 

[Cosmik]

Supposing you have a dry recording of a pipe organ. You think "This would sound good played in a stone church".

(This is not fundamentally different from saying "I have a recording of my favourite singer but I've only heard it in my car. I would like to hear it in a good room". It's just a question of degree).

You set up some speakers in a church and play it. It sounds heavenly.

You think "Aha, it's great, but it could sound better". You make some measurements. They're terrible! All over the shop, with huge resonances and nulls. Through judicious use of a graphic equaliser, you get a flat measurement - when you measure it a certain way. You convince yourself the system is objectively better and it therefore sounds better. Odd, but better. But different measurements give different results and this bothers you.

Gradually you use more and more sophisticated correction methods so that you get consistently flatter results no matter how you make the measurement. Now, when certain notes are held steady, where they previously hit a resonance that built up over time, they are now pre-emptively reduced in volume, so that those awful singing resonances are 'pulled'. Much better! Gradually you refine the system with more and more sophisticated DSP until eventually that glorious moment: you sit in a certain pew and hold very still and the recording sounds just like it does in an anechoic chamber. Success!

 

[Don Hills]

... As far as flatness is concerned, it shows an attenuation of 20dB from 1 to 10kHz, and even more beyond 10kHz. So, it does not surprise me that some sloping is prefered. ...

It surprises me.
Take a source of live music. Record it, in an anechoic environment. Now move the source into your room alongside a reproduction system. What frequency response should the reproduction system have, to exhibit a perceived similar frequency response to the live source?

 

[Theo]

Supposing you have a dry recording of a pipe organ. You think "This would sound good played in a stone church".

(This is not fundamentally different from saying "I have a recording of my favourite singer but I've only heard it in my car. I would like to hear it in a good room". It's just a question of degree).

You set up some speakers in a church and play it. It sounds heavenly.

You think "Aha, it's great, but it could sound better". You make some measurements. They're terrible! All over the shop, with huge resonances and nulls. Through judicious use of a graphic equaliser, you get a flat measurement - when you measure it a certain way. You convince yourself the system is objectively better and it therefore sounds better. Odd, but better. But different measurements give different results and this bothers you.

Gradually you use more and more sophisticated correction methods so that you get consistently flatter results no matter how you make the measurement. Now, when certain notes are held steady, where they previously hit a resonance that built up over time, they are now pre-emptively reduced in volume, so that those awful singing resonances are 'pulled'. Much better! Gradually you refine the system with more and more sophisticated DSP until eventually that glorious moment: you sit in a certain pew and hold very still and the recording sounds just like it does in an anechoic chamber. Success!

Very good! I like that! And, we're back to the room discussion... The anechoic dry recording may still not be what I would prefer, though

 

[svart-hvitt]

I think I see where you are coming from on this. First things first maybe.

My own idea on DACs, they are incredibly popular to talk about, so many are made so many different promises. Probably the least important part of the chain as so many are good. And yet surprisingly many have failed to live up to promises. Also like the recent info on how some headphone amps fall apart driving headphones is important. Shows you have to do your measurements mind the details etc. A good topic to get people interested in measuring gear though.

But past a certain point I simply see no way they can be audibly different. I mean a noise floor of -123 db instead of -108 db? If my speakers max out at 108 db SPL I'll never hear either one. i certainly don't think people want to pay extra money for less performance than cheaper gear. So if -123 db is the same price sure why not. But it isn't important anymore is it.

Now where I doubt much gear is so pure is amps drivings speakers and headphones. Worthwhile to do measures of those.

I do find myself fuzzy on the details of what you are proposing. I don't think anything other than amps a little and speakers/headphones making a difference. I see how it could go the other way. What is the minimum performance to sound perfect. Speakers and headphones are not close, and everything else might be far beyond sounding perfect.

As for DACs the Schiit gear may be the only one that sounds different than any of the others. It does so because it performs worse.

@Blumlein 88 ,

I think what I miss is a clearer ranking order of what counts and what does not (so much), based on research and science.

We’d need subjects’ preferences, I believe, to sort out which measurements are the most important and at what limits, thresholds, they become irrelevant.

Amir’s measurements are very important to point out that hifi needn’t cost much. He has also made us more cautious on buying blind - based on web forum gossip - things from the internet, things that may cause harm in a wider sense (bad sound, electric shock, burning down the house). However, our next step would be practical application of said and other research. Would «Audio Science Review Hifi Playback for Dummies» be a goal for us to produce? With practical information regarding price, quality and implementation in room?

 

[SIY]

It surprises me.
Take a source of live music. Record it, in an anechoic environment. Now move the source into your room alongside a reproduction system. What frequency response should the reproduction system have, to exhibit a perceived similar frequency response to the live source?

Frequency response measured how and where, exactly? What polar patterns for the speakers?

Once you get away from the simple matter of electrical signals, things get... not simple.

 

[Duckeenie]

Better marketing combined with better drugs should do the trick. Have a better idea?

Schiit owners are not just happy but super loyal. Identify with the brand and love it. It ain't about the sound performance. I say mostly the story, idea, the marketing.

I thought you were one of the objective guys?

 

[Arnold Krueger]

Frequency response measured how and where, exactly? What polar patterns for the speakers?

Once you get away from the simple matter of electrical signals, things get... not simple.

One of the things that years of widespread frustration with speaker measurements has led to is that there are indications of possible answers to exactly the questions you ask.

For example, IME we have gotten to the point where exactly how we measure something is less and less important.

If I measure THD with a computer using measurement software I can obtain similar results as I would get for the same tests performed with a classic HP 339, and ditto for an AP S1 test set, subject to their rather obvious limitations in terms of things like residuals, ease of use, etc. In terms of identifying which components have adequate performance for listening to music with high fidelity, they are equally useful, even if two of them are in many ways hopelessly obsolete.

Good measurement techniques now tend to converge, and the places they still diverge are well known to those who are willing to work a little to find such things out.

For example the two leading ways to measure speaker frequency response involve two very different test signals - namely pink noise and a chirp (swept sine). The pink noise is usually measured with a filter bank with long-term averaging. The chirp is typically measured with a tracking filter possibly with averaging over multiple sweeps.

The chirp method as commonly used (REW, Holme Impulse) tends to disregard the room's effects, while the pink noise method (REALRTA, etc) tends to include it.

You figure out what job you want done, and you pick the right tool for the job.

 

[SIY]

That's a matter of test signal (there's lots of ways to get impulse response), but it says nothing about what point(s) in space to set the mike and how to interpret the curves obtained at different points. Further, it says nothing about how the measurement is gated (or if it's gated), how bass and midrange/treble measurements are merged (if they are merged at all). A few minutes of thinking and I'm sure I could come up with a bunch more variables.

Basically, the 1-D aspect is trivial, the 3-D aspect, not so much.

 

[Arnold Krueger]

Supposing you have a dry recording of a pipe organ. You think "This would sound good played in a stone church".

(This is not fundamentally different from saying "I have a recording of my favourite singer but I've only heard it in my car. I would like to hear it in a good room". It's just a question of degree).

You set up some speakers in a church and play it. It sounds heavenly.

You think "Aha, it's great, but it could sound better". You make some measurements. They're terrible! All over the shop, with huge resonances and nulls. Through judicious use of a graphic equaliser, you get a flat measurement - when you measure it a certain way. You convince yourself the system is objectively better and it therefore sounds better. Odd, but better. But different measurements give different results and this bothers you.

Gradually you use more and more sophisticated correction methods so that you get consistently flatter results no matter how you make the measurement. Now, when certain notes are held steady, where they previously hit a resonance that built up over time, they are now pre-emptively reduced in volume, so that those awful singing resonances are 'pulled'. Much better! Gradually you refine the system with more and more sophisticated DSP until eventually that glorious moment: you sit in a certain pew and hold very still and the recording sounds just like it does in an anechoic chamber. Success!

I find this account to be speculative, limited in viewpooint and biased. I've actually done the experiment, only instead of recording a real pipe organ in an anechoic chamber which is difficult, I had a transcription of a modern high quality electronic organ's audio signal output which is easier.

What I found is that the transcription sounded different than the room's own real pipe organ. Pipe organs strongly tend to be vastly different in terms of their construction so no surprise there.

The transcription sounded like a different pipe organ, but a remarkably realistic sounding one had been magically transplanted into the room. I think that is what we would rationally expect.

The conventional wisdom is that you adjust room acoustics depending on what the mission of the room actually is. In audio, this means that the ideal listening room for certain genre's of music and dialog would be necessarily different.

Churches are acoustically tough because the same room is used for spoken word and music of various genres. It turns out that there are two flavors of spoken word, one for lectures and one for drama.

The ideal room acoustics for listening to these varied programs are different. Yet they may all be parts of the same church service. These are divergent rquirements, to say the least. One solution would be a room with adjustable acoustics, and while I've been in High School auditoriums, University recital halls and Opera house recital halls with adjustable acoustics, I've nver seen anybody roll out that technology in a church. Probably just a matter of time for the relevant parties to come up to speed.

 

[Arnold Krueger]

That's a matter of test signal (there's lots of ways to get impulse response), but it says nothing about what point(s) in space to set the mike and how to interpret the curves obtained at different points. Further, it says nothing about how the measurement is gated (or if it's gated), how bass and midrange/treble measurements are merged (if they are merged at all). A few minutes of thinking and I'm sure I could come up with a bunch more variables.

Basically, the 1-D aspect is trivial, the 3-D aspect, not so much.

I didn't get into those issues because I didn't think I was writing an omnibus treatise on measurements.

Mic placement seems to be a mystery to some, but IME people who do it a lot just do it. Basically, you put the mic where the thing you want to measure is present. This is similar to recording. Figure out what you want, use experience and experiments to find it, and there you are!

For example, if you want to measure on-axis frequency response in the direct field of a speaker in a room, you put the mic on axis and in the direct field of that speaker in that room.

 

[sergeauckland]

I think I see where you are coming from on this. First things first maybe.


As for DACs the Schiit gear may be the only one that sounds different than any of the others. It does so because it performs worse.

I think that's true for several products that have a reputation amongst 'subjectivists' for good sound. Take Boenicke W8 loudspeakers. They have one of the nastiest frequency responses of any loudspeaker of that sort of (expensive) price, yet have a devoted following amongst those for whom a decently flat response sounds just that, flat. (and for flat, read boring).

Similarly, the popularity of SET valve amplifiers, that have horrendous distortion at the frequency extremes, often offer flea-power that couldn't pull the skin off a rice pudding, and yet have a devoted following. Then there's the latest 'craze, that of using Reel-Reel tape recorders and recording a CD onto tape as it sounds 'better'. In other words, some people actually like treble crushing and bass woodles, and a little W&F doesn't come amiss either...from the same people that would shun a DAC that has nanoseconds of jitter for one with picoseconds!

I am a card-carrying objectivist (at least I would be, if I had a card) and I measure everything I buy, and as long as it's good enough for transparency, it's good enough.

I really don't understand how the Audio business has got itself into this state, where transparency is considered boring, and 'character' i.e. poor performance praised.
S.

 

[Cosmik]

I find this account to be speculative, limited in viewpooint and biased. I've actually done the experiment, only instead of recording a real pipe organ in an anechoic chamber which is difficult, I had a transcription of a modern high quality electronic organ's audio signal output which is easier.

What I found is that the transcription sounded different than the room's own real pipe organ. Pipe organs strongly tend to be vastly different in terms of their construction so no surprise there.

The transcription sounded like a different pipe organ, but a remarkably realistic sounding one had been magically transplanted into the room. I think that is what we would rationally expect.

The conventional wisdom is that you adjust room acoustics depending on what the mission of the room actually is. In audio, this means that the ideal listening room for certain genre's of music and dialog would be necessarily different.

Churches are acoustically tough because the same room is used for spoken word and music of various genres. It turns out that there are two flavors of spoken word, one for lectures and one for drama.

The ideal room acoustics for listening to these varied programs are different. Yet they may all be parts of the same church service. These are divergent rquirements, to say the least. One solution would be a room with adjustable acoustics, and while I've been in High School auditoriums, University recital halls and Opera house recital halls with adjustable acoustics, I've nver seen anybody roll out that technology in a church. Probably just a matter of time for the relevant parties to come up to speed.

It's a thought experiment, to highlight the confusion that I think applies to audiophile 'objectivity'. We could perhaps come up with three basic classes of card carrying 'objectivist':

1. The person who seeks perfect objective neutrality as far as the amplifier speaker terminals. This person may be very lax about speakers, regarding them as something of a mystery ("Perfect frequency response, but sounds different from other 'perfect' speakers. Oh well, that's part of the fun.").
2. The person who takes the quest for objective neutrality beyond the amplifier to the speaker itself (the same standards as they would apply to DACs and amplifiers). Someone like Bruno Putzeys perhaps; speakers are not a mystery to him. The main variable they allow is the dispersion angle of the speaker, but this is allowed within their definition of neutral as long as it is constant at all frequencies (as close as they can get).
3. The person who tries to take objective neutrality to the listener's ears including the room. They may be lax about the speakers themselves because they assume they are in the 'feedback loop'. This person thinks they want to listen to speakers in a room, but also wants to eliminate the room from the sound to varying degrees. The "varying degrees" is a subjective, arbitrary judgement and is limited by what is possible. Their criterion of excellence is how close the in-room measurements get to an anechoic chamber (perhaps "with a gentle slope downwards").

 

[svart-hvitt]

I think that's true for several products that have a reputation amongst 'subjectivists' for good sound. Take Boenicke W8 loudspeakers. They have one of the nastiest frequency responses of any loudspeaker of that sort of (expensive) price, yet have a devoted following amongst those for whom a decently flat response sounds just that, flat. (and for flat, read boring).

Similarly, the popularity of SET valve amplifiers, that have horrendous distortion at the frequency extremes, often offer flea-power that couldn't pull the skin off a rice pudding, and yet have a devoted following. Then there's the latest 'craze, that of using Reel-Reel tape recorders and recording a CD onto tape as it sounds 'better'. In other words, some people actually like treble crushing and bass woodles, and a little W&F doesn't come amiss either...from the same people that would shun a DAC that has nanoseconds of jitter for one with picoseconds!

I am a card-carrying objectivist (at least I would be, if I had a card) and I measure everything I buy, and as long as it's good enough for transparency, it's good enough.

I really don't understand how the Audio business has got itself into this state, where transparency is considered boring, and 'character' i.e. poor performance praised.
S.

On deviations from neutral and audio products: You are surprised to find that neutral is not of appeal to everyone. There is a plausible reason for this. The question is: Does neutrality match everyone’s preferences?

It is only in the case where you have match between neutrality and preferences that people prefer the neutral. Even if the average or median person’s preferences is that of neutrality, most - if not all - persons’ preferences could deviate from the neutral. In other words: One should not be surprised to find producers that practice some sort of preference matching. Because most people’s preferences may deviate somewhat from neutrality, one could argue that there is hardly a market for neutral products.

Adding to the problem of people’s deviations from neutrality, this devition could be a function of emotional state, time of day, week, seasons, age and more.

Admittedly, I may be stretching this thought experiment a bit far, but it gives me comfort in my own choice: I choose neutrality because then I know that my audio setup, on average over time, will sound pretty good. Not optimal every time for every emotional or physical condition, but on average I believe neutrality is a safe choice.