Research Project: Infinity IL10 Speaker Review & Measurements

[restorer-john]

The fact of the matter is @amirm cannot "untrain" himself. Critical/trained listeners are realistically hyper-attuned to find faults, whereas audiophiles, even so-called golden eared ones, are pursuing sound that gives the most enjoyment or emotion. In short, they aren't listening to speakers to tear them to shreds, they are hoping that the next pair of speakers they audition bring them tears of joy.

That is essentially coming from completely different ends of the spectrum- just like subjectivists and objectivists. We like to think of ourselves as being able to separate the objective measurements and just enjoy the presentation, but that becomes impossible once you know or have characterized the product in an objective/trained manner.

The correlation in measurements between the Harman and the Klippel is however truly excellent and should put to bed any doubts of Klippel's accuracy.

 

[zermak]

It doesn't work that way. Lossy compression converts everything to frequency domain. A perceptual model is then applied to which frequencies can be truncated that are less audible. Then depending on how full the internal buffer is (i.e. how difficult past content has been to encode), the level of quantization (truncation) varies. The result is that distortion shoots up and down from one note to another. Even a full model of the encoder can't estimate the distortion correct as used in lossy encoders.

I see that and I kinda read how lossy formats work but I may be able to just listen to the delta (without putting the distortions/artifacts in percetages) and figure out what I was missing and what you was able to spot/listen, if you know what I mean.

 

[RayDunzl]

Critical/trained listeners are realistically hyper-attuned to find faults, whereas audiophiles, even so-called golden eared ones, are pursuing sound that gives the most enjoyment or emotion.

I can't really enjoy my little JBLs on music, whereas I find the big MartinLogans quite satisfactory.

They're fine for TV and Movies.

Though, I might fire up the big ones for a Feature Attraction.

I'm not golden eared.

 

[RayDunzl]

Lossy compression converts everything to frequency domain. A perceptual model is then applied to which frequencies can be truncated that are less audible. Then depending on how full the internal buffer is (i.e. how difficult past content has been to encode), the level of quantization (truncation) varies. The result is that distortion shoots up and down from one note to another. Even a full model of the encoder can't estimate the distortion correct as used in lossy encoders.

Could that be demonstrated with some tone test?

 

[amirm]

It seems odd (without looking at other speaker measures to be sure) that the tweeter output increases as the frequency goes down.

I should have been more clear. This is the near field measurement of the drivers but all of them are playing. So you eventually see the response of the woofer.

I forgot this speaker has dual binding so I could have tested each driver separately.

 

[amirm]

Could that be demonstrated with some tone test?

No. A test tone is trivial to compress since it has just one frequency. The enemy of compression is high frequency transients.

 

[amirm]

I see that and I kinda read how lossy formats work but I may be able to just listen to the delta (without putting the distortions/artifacts in percetages) and figure out what I was missing and what you was able to spot/listen, if you know what I mean.

Sure, do a difference file and then listen and it will tell you what is taken out. But note that a lot of that is masked so what is gone isn't all audible (most shouldn't be).

 

[HooStat]

I fired up the speaker and my "5 second impression" was good. This speaker sounded neutral. Female vocals were especially nice.

He also didn't say it was awful.

It does beg the question of what, if anything has changed in 20 years. It would be a nice blind test to compare the Infinity and the Revel. We could learn something.

 

[AudioTodd]

All models of reality are wrong; some are useful. We should aspire to more useful models.

I love this.

 

[AudioTodd]

Well, other testers don't have such training (to hear small distortions). The number of people in the world that have this training is very small so getting them to participate in this kind of testing is impossible.

So it comes down to me being tested somehow not others.

I have a question for @amirm: Do you think the rest of us should really WANT such training, or does it get in the way of enjoying music? You have devoted an unbelievable amount of time to testing audio equipment essentially for free, not to mention the money you have spent to get basically the best equipment available to pursue this passion. Is this driven, in some part, by what you now hear and I have probably learned to ignore (not really, as I love my Benchmark stuff for its non-dramatic dramatic superiority over everything before it even if I only THINK I hear the superiority...) and a desire to find components from which YOU DO NOT HEAR these now obvious distortions? Ignorance perhaps can be bliss as well as a torrent of wasted high end stereo money...

 

[aarons915]

Well, other testers don't have such training (to hear small distortions). The number of people in the world that have this training is very small so getting them to participate in this kind of testing is impossible.

So it comes down to me being tested somehow not others.

I wouldn't be so quick to assume it's distortion related, the M16 has higher than normal distortion in a few areas as well. My only point about getting other data points is to see if others feel the same way before we declare that something is wrong with the preference ratings based on 1 person's sighted impressions. Another explanation could be that the M16 has a smoother directivity index and has slightly wider dispersion just eyeballing the graphs.

 

[RayDunzl]

Do you think the rest of us should really WANT such training, or does it get in the way of enjoying music?

Music and sound are separate items for me.

Consider:

Same music, different "sound" and a bit different "quality".

The music remains the same.

I think I "like" the harp transcription the best, even though it was (probably) composed for piano. It might have to do with the visuals. And the quality of the recording no doubt plays a part. I'm inordinately fascinated with how she can trade hands on the "drone" and never blur a beat.

---

I just realized. The main body of the piece in in 4/4, the drone is eighth notes, and the fast parts are eighth triplets. I'm a sucker for odd meter.

4/4 and key of C. How pedestrian.

 

[Vladimir Filevski]

A factor remains that my hearing acuity for distortion/small detail is well above average ...

I am strongly supporting this. For some people (me included) distortion sensitivity is a curse and blessing, I have found that anything above 0.6% third harmonic in mid-high frequency band is becoming objectionable. Anyway, with today's technology, good and not-so expensive loudspeaker should not have trouble to better that.

 

[zermak]

Sure, do a difference file and then listen and it will tell you what is taken out. But note that a lot of that is masked so what is gone isn't all audible (most shouldn't be).

I indeed was doing testings with the Ogg Vorbis encoder/compression (to use with my portable player) and I concluded that I am not a good critical listener because, if I recall, even the -V6 profile was enough for a pleasant listening but I didn't pull out any ABX test on foobar2000 and maybe it's time I dug into it.

 

[ROOSKIE]

Well, other testers don't have such training (to hear small distortions). The number of people in the world that have this training is very small so getting them to participate in this kind of testing is impossible.

So it comes down to me being tested somehow not others.

Is there any particular online resource that offers at least part of the training you have? I'd be curious to see what it offers.

I don't doubt your hearing or the quality of your training that said some people do have excellent hearing naturally and others could train for a lifetime and never be able due to achieve a top tier due to their biology (like I won't ever beat Michael Jordan in 1 on 1, or even slam dunk once) I've been around people who are pretty clumsy listeners and a few others who surprised me a bit. I guess what I am saying is that I still really look forward to some well controlled blind testing with a group. I know you have mentioned possibly arranging some one day, I hope you keep this speaker as it makes for a very interesting component of such testing.

I did find this test one test online recently, it is a test for hearing harmonic distortion.
There are a few problems with it such as
-it unknown how much additional harmonic distortion the listeners system added to the tracks so no real meaningful threshold can be negotiated.
-the harmonic distortion amounts are not exactly the ones I would have picked myself (.oooooo2/.02/.3/3 %) as I think what was needed was at least one where the distortion is above 3%.
-the controller concluded that there was a meaningfull difference between the amounts folks picked up on but in my personal review I don't really see a strong correlation yet.
-this also doesn't tell us what happens when the harmonic distortion is only in a narrow range such as 1-2.5krz, vs a whole spectrum.

anyway here are the links

The test
https://archimago.blogspot.com/2020/01/internet-blind-test-is-high-harmonic.html

Results Part1
https://archimago.blogspot.com/2020/05/blind-test-results-part-i-is-high.html
Results part 2
https://archimago.blogspot.com/2020/06/blind-test-results-part-ii-is-high.html
Results part3
https://archimago.blogspot.com/2020...howComment=1591765199598#c1384889885114091245

 

[ROOSKIE]

I am strongly supporting this. For some people (me included) distortion sensitivity is a curse and blessing, I have found that anything above 0.6% third harmonic in mid-high frequency band is becoming objectionable. Anyway, with today's technology, good and not-so expensive loudspeaker should not have trouble to better that.

I understand that for classical type instruments and such harmonic distortion may stand out quite a bit.
What I am wondering is for Rock, electronic and really anything that uses synthetic or modified sound how capable are folks at picking up on the distortions? I mean essentially all pedals and amps and different things are adding massive amounts of harmonic distortion in on purpose to a achieve a really unknown sound on the recording, some guitar amps have stratospheric harmonic distortion. Even keyboards and such have mind altering properties. So much is altered in most contemporary music with a lot of 2nd and 3rd added at the recording level.
Essentially is .6% 3rd order harmonic distortion really only an issue with well recorded classical music and such?

 

[Vladimir Filevski]

Of course, I am referring to classical instruments/recordings only.

 

[GXAlan]

Is there any particular online resource that offers at least part of the training you have? I'd be curious to see what it offers.

http://harmanhowtolisten.blogspot.com/2011/01/welcome-to-how-to-listen.html?m=1

 

[Newman]

If Amir measured before listening, then cognitive bias becomes an issue -- especially if he has preconceptions around being sensitive to distortion.

Another issue is the unknown history of this 20+ year old sample of a speaker. It might be 'as new', or it might possess replaced (or even different) drive units, crossover mods, etc. Might have endured physical or electrical stress. Meaning that the logical lineage from this speaker to the Harman tests might be invalid.

cheers

 

[mhardy6647]

A factor remains that my hearing acuity for distortion/small detail is well above average due to extensive training. Harman listeners did not have such skills. So it is entirely possible that what bothers me doesn't bother hardly anyone else.

The current test was also performed on a used loudspeaker in (unless I missed something) "as is" condition, which might explain the objectionable audible grunge even if "immeasurable", I would posit.