desirable distortion

[Krunok]

But I also know that curiosity is what made us fall into sin...

Not to forget what killed the cat, right?

 

[Thomas savage]

I put my mouth where my money is: Genelec speakers are often criticised for being boring; this is from a 6moons review of my speakers:

«Never once did I catch myself fantasizing about this venerable Finnish make making the leap into my own listening room. Why? Because the Genelec sound (something of an oxymoron, admittedly) often struck me as being unremittingly focused. Despite the total absence of cabinet signature, iron-fisted bass control and a midband free of any additive blemish, the overall effect rarely seemed to conjure up anything quite as nebulous as listener involvement. But of course you could still justifiably argue they were simply doing the job they were intended for».
Source: http://www.6moons.com/audioreviews2/genelec/1.html

So I am very much in the hifi camp, i.e. Ideally listening through the system and not to the system.

Still, I found the Katz article a bit intriguing. Maybe because it made me want to do the same listening exercise myself on my 8351s.

This is what made me curious:

«My Blender allows us to test this hypothesis and perhaps for the first time, quantify the amount that is needed».

But I also know that curiosity is what made us fall into sin...

Hey, don’t knock sin it’s some of the best times out there

( the new laughing emoji looks like a psychopath, it’s kinda disturbing lol)

 

[svart-hvitt]

TBH, I don't care much for the 6moon reviews as all they are offering is their subjective opinion.

The reason I quoted 6moons on Genelec was to illustrate what happens when a reviewer who is accustomed to tweaks and colouration listens to something which is correct, neutral:

«...the overall effect rarely seemed to conjure up anything quite as nebulous as listener involvement».

I used 6moons to illustrate that I prefer to listen to speakers that do not «conjure...listener involvement».

 

[Krunok]

The reason I quoted 6moons on Genelec was to illustrate what happens when a reviewer who is accustomed to tweaks and colouration listens to something which is correct, neutral:

«...the overall effect rarely seemed to conjure up anything quite as nebulous as listener involvement».

I used 6moons to illustrate that I prefer to listen to speakers that do not «conjure...listener involvement».

No worries M8, that's exactly how I got it.

 

[Krunok]

Hey, don’t knock sin it’s some of the best times out there

( the new laughing emoji looks like a psychopath, it’s kinda disturbing lol)

Yep, exactly as psychopath! Not to mention the winking emoji, with the chin on top of his head.

 

[Thomas savage]

Yep, exactly as psychopath! Not to mention the winking emoji, with the chin on top of his head.

Yes but i won’t wake in the night to find the winking emoji standing over me with a kitchen knife, that laughing emoji on the other hand ( shit my pants emoji lol)

 

[Krunok]

Yes but i won’t wake in the night to find the winking emoji standing over me with a kitchen knife, that laughing emoji on the other hand ( shit my pants emoji lol)

LOOOL

 

[Cosmik]

I don't know why people persist in the fantasy that their fuzzbox is a highly complex 'frequency domain harmonic analysis and synthesis processor' and why they plot out its response in terms of frequency response.

It would be far more informative if they plotted its response as the bent transfer function it is. Want to know what a transfer function with a clipped top does to a signal? Simply imagine the time domain signal going into a transfer function with a clipped top. Want to know what a transfer function that combines a pointy characteristic, a wiggle at the origin and a rounded off top does to the signal? Ditto.

 

[Wombat]

TBH, I don't care much for the 6moon reviews as all they are offering is their subjective opinion.

And prolific use of exaggerated superlatives.

 

[Wombat]

Bob Katz is one of the best-known and most highly regarded mastering engineers in the USA. In his book, «Mastering audio: the art and the science», he describes seven ear-training exercises for improving your hearing acuity in the context of recording or mastering music.

To compare Mr. Katz with an untrained listener (cleaning lady) is unfair.

In a paper published on blind listening tests a surprise finding was that participating recording engineers performed no better than the average 'Joe'. I am sure it was posted on this forum. I'll continue to look for it.

 

[Cosmik]

No distortion sounds better than a little bit of bad distortion, but moderate, well-distributed distortion sounds better, too!. I believe there's a middle amount where distortion can sound deadly. Why? Because in that middle area, where the overall distortion measures somewhat low, but not close enough to zero, the presence of some higher harmonics can psychoacoustically predominate over the important lower ones...
...why would anyone want to add more distortion to your audio reproduction? The answer is my reason number two: Masking. Just as noise can mask signal, distortion of a certain magnitude and frequency content can mask other distortion.

(Ignoring the huge amounts of 'expectation' in the whole idea)
Sounds like someone wanting to take an already-poor system out of 'the uncanny valley' and into the realms of the clearly artificial by adding a strong fuzzbox effect. Is this the height of the audiophile's ambition?

 

[Cosmik]

Distortion is additive. You, the consumer, cannot take it away,

Technically, is this true? If I feed a time domain signal into a slightly bent transfer function (without any 'points of inflection') I generate harmonic and intermodulation distortion.

Viewed in the frequency domain this looks like a mess that cannot be untangled. But if I feed this distorted output into the complementary bent transfer function, I remove all the distortion. In other words, viewing the distortion in the frequency domain obscures the triviality of what is going on.

Clearly real world distortion may be more complex, so the chances of you exactly counteracting existing distortion through the use of a diode, valve or transistor are slim; the thing to do is avoid audible distortion in the first place. As the article points out, this is already possible if you make a small effort. So why fiddle about with this stuff?

 

[Krunok]

As the article points out, this is already possible if you make a small effort. So why fiddle about with this stuff?

Obviously there are a certain number of folks who believe that things simply start to sound "better" if you introduce a little distortion, like playing things via the tubes. As I already said, I can't tell the difference between my tube preamp/amp combo and my SS preamp/amp combo, so I can't be the judge of that, but I would really like to know if Mr. Katz would be able to tell the difference between the tube preamp he was so euphoric about and some high quality SS preamp.

 

[Fitzcaraldo215]

It is intresting how knowing the pre Amp is there changed his perception of what he heard. I always fancied putting a Center speaker in a stereo setup and have people say what they think while having that Center channel unconnected.

I’d bet you would get a lot of “ oh, the voice is so centred and the image so much more stable “ etc..

I certainly know with my tv setup I have synced the visual with the sound from my two JBL’s, the image is amazing as voices come directly from the mouths of the people on the screen the illusion holds even when I’m far outside the effective stereo zone. The sound never collapsing into one speaker.

Before these JBL’s i had a sound bar, when I first connected it it was terrible as the sound was obviously divorced from the picture... then after awhile Is was as one..., maybe the sound bar ‘ broke in’ lol

We know, of course, that visual cues strongly affect our perception of what we think we hear. That is part of what is happening to you when listening to your phantom imaged JBLs with the TV. And, I am not doubting that some people would be fooled by a dummy center channel for the reason you cited. Our acuity for locating things in space is much greater via sight than by hearing alone. I think we are hard wired to trust what we see as overriding what we hear spatially.

Don't know what happened with your sound bar. Mine has different settings via remote for various spatial effects, and in the narrowest "close in" speech mode, it delivers decent consistency with the video. However, the wider modes sound increasingly dissociated from dialog on the screen with poorer dialog articulation. But, sound bars are sound bars, a necessary evil in some cases, like my lowly bedroom system. They are often adequate and convenient, but not much more. And, they are hardly a fair comparison to your JBLs.

But, if you are implying that phantom center imaging is somehow just as good as a center channel, I suggest you look at Floyd Toole's discussion of it in his latest book, with supporting data, of course. No one is saying that phantom center imaging is terrible by comparison or that a true center channel is day/night better. Phantom center can sound just fine, but without a comparison to the use of a true, properly matched center channel, you don't really know.

On my main Mch system, where I can experiment with 4.1 vs. 5.1, etc., I can hear the difference, apples to apples, and I much prefer the use of a properly matched center channel. For me with Mch music, even without centered soloists, that is consistently true. I have been able to identify just from listening to some music recordings unknown at the time to be 4.0 as lacking a recorded center channel on a friend's system, even with an unused center channel speaker sitting up front. I can get used to them and enjoy them, but most Quad-era 4.0 remasters have a perceptable "hole in the middle" now that I am used to 5.0/.1 recordings.

 

[Frank Dernie]

Technically, is this true? If I feed a time domain signal into a slightly bent transfer function (without any 'points of inflection') I generate harmonic and intermodulation distortion.

Viewed in the frequency domain this looks like a mess that cannot be untangled. But if I feed this distorted output into the complementary bent transfer function, I remove all the distortion. In other words, viewing the distortion in the frequency domain obscures the triviality of what is going on.

Clearly real world distortion may be more complex, so the chances of you exactly counteracting existing distortion through the use of a diode, valve or transistor are slim; the thing to do is avoid audible distortion in the first place. As the article points out, this is already possible if you make a small effort. So why fiddle about with this stuff?

This is pretty well what Devialet do in their SAM system. They measure the error between the speaker output and input, invert the transfer function add a thermal and mechanical overload limit and use it to multiply the output by.
For small speakers it is an amplification and phase correction and perhaps can be seen as a distortion of the frequency response in itself when it levels off the bass gain to avoid damage. For several big speakers it has to reduce the bass driver amplitude.
https://www.devialet.com/en-eu/expert-pro-sam-ready-speakers/harbeth/40.2/

 

[sergeauckland]

This is pretty well what Devialet do in their SAM system. They measure the error between the speaker output and input, invert the transfer function add a thermal and mechanical overload limit and use it to multiply the output by.
For small speakers it is an amplification and phase correction and perhaps can be seen as a distortion of the frequency response in itself when it levels off the bass gain to avoid damage. For several big speakers it has to reduce the bass driver amplitude.
https://www.devialet.com/en-eu/expert-pro-sam-ready-speakers/harbeth/40.2/

I wonder how Devialet measure the loudspeakers they model in SAM. Do they have (or rent) an anechoic chamber accurate to 20Hz? Also, do they equalise the loudspeakers anechoically upto 20kHz or is it just a bass function? It's a wonderful idea which should be used more, but a huge commitment (and expense) to get right.

S.

 

[sergeauckland]

As to distortion, desireable or not, my view is that any distortion added at the recording/mixing/mastering stage is a part of the 'art' being presented to us, the buying public. Consequently, if there is distortion on the CD, then it's (presumably!) what the artist/producer/mastering engineer intended.
On the reproduction side, any distortion is wrong, as it goes against what the recording intended. I se it as akin to going to a restaurant and putting ketchup on one's food. Not what the Chef intended.

There is an argument for no distortion on naturally-recorded acoustic music, as the art is the performance, not the recording, so the recording should be as 'natural' as possible. Fortunately, much classical music is recorded that way, even if the performance is made up of hundreds of individual edits. I applaud labels like Nimbus who use the edit (in their words) to save a performance, not to create one, and as such tend to record in long (at least one movement) takes.
S.

 

[DonH56]

FR can be fairly easy to correct, and if you compensate properly to include phase and not just magnitude, then time domain response should also improve. Not sure how much you can do if the system is not LTI, or least in some form of steady-state (which music does not normally have). Obviously you can go all-out to match and compensate transfer functions (once you know them) but that can be tough in the real world.

The usual example of "desirable distortion" is in a subwoofer where the harmonics are much more audible than the fundamental thus a sub with 50% THD playing a 20-40 Hz tone sound much "louder and fuller" than one with 1% THD playing the same tone at the same amplitude. Not desirable to me, but IME most folk who hear the two subs will pick the one with more distortion as "better".

 

[Cosmik]

Do they have (or rent) an anechoic chamber accurate to 20Hz?

Do you have to do that, or can these findings be relied upon?
http://www.xlrtechs.com/dbkeele.com/PDF/Keele (1974-04 AES Published) - Nearfield Paper.pdf

Low-Frequency Loudspeaker Assessment by Nearfield Sound-Pressure Measurement* D. B. KEELE

A loudspeaker test technique is described which depends on nearfield pressure measurements made in a nonanechoic environment. The technique allows extremely simple measurements to be made of frequency response, power response, distortion, and electroacoustical efficiency...

...The theory presented, along with supporting experimental measurements,shows that loudspeaker system piston-range characteristics can easily be measured by sampling the nearfield pressure with a test microphone held close to the acoustic radiator. Valid nearfield measurements may be taken in any reasonable environment. without the use of an anechoic chamber or large outdoor test site. Experimental measurements using the nearfield technique show excellent agreement with more traditional test methods.

 

[sergeauckland]

Do you have to do that, or can these findings be relied upon?
http://www.xlrtechs.com/dbkeele.com/PDF/Keele (1974-04 AES Published) - Nearfield Paper.pdf

Yes and no.

One can do pseudo-anechoic measurements using even free software like ARTA or REW but only above a critical frequency, depending on the room dimensions. For a typical room, that's somewhere around 200Hz.

Below that, one can do nearfield assessment, and then splice it into the pseudo-anechoic measurement, BUT:-

If the loudspeaker has multiple woofers, which one gets measured. Less of a problem if the loudspeaker has parallel bass units, but many are 2 1/2 way, where the bass units don't work directly in parallel.

If the loudspeaker has bass-reflex loading, how is the port output dealt with? Made worse if the port is at the back or underneath.

How is the baffle compensation dealt with?

At best, splicing in nearfield measurements with pseudo-anechoic is an approximation, so nowhere near good enough as 'proper' equalisation as can be achieved in a full-range anechoic chamber.

S.