PS Audio sent Erin their speaker??!!

[napilopez]

Of course - my original posts said I love the Klippel data Erin provides

Great for our community

His subjective impressions, I don't believe he can hear a 2kHz scoop listening to music in a normal room, with room acoustics issues - all before seeing the simulated anechoic data

Matters of truthfulness aside, remember a wide, shallow deviation from flat as in the LS50 Meta is generally easier to hear than deeper, narrower one. This is shared often, so you may have seen it, but the following bumps are equally detectable in listening tests. (From Toole's book)

Note that it's actually easier to detect these at higher frequencies, presumably up to a point.

(Edited to include a more illustrative image)

 

[CleanSound]

but I definitely agree about listening in mono.

I respectfully and vehemently reject this method of mono only listening.

Let me explain why: I have a pair of Neumann KH 120, for those who have heard them will tell you about the 3D imaging, it is uncanny. Those speakers go on standby and will auto power on based on input signals. Sometimes, when they turn on, they don't always power on at the same exact time, one speaker will come on before the other. And my God, each time that happens, I am reminded the grandness of the imaging.

How in the world can you listen to such impeccable imaging in mono?

I believe during subjective impression, the individual should also conduct listening in stereo in addition to whatever mono evaluation is being done.

I also believe there is value in subjective impression (either before or after the measurements, as I can see possible bias going either way) as there are some characteristics that the measurements alone cannot tell you, to use the example of 3D imaging again. I am not saying to trust the subjective impression 100%, on the contrary, take it as a grain of salt. And know that you can fast forward if you are hearing too much subjective diahera.

 

[napilopez]

I respectfully and vehemently reject this method of mono only listening.

Let me explain why: I have a pair of Neumann KH 120, for those who have heard them will tell you about the 3D imaging, it is uncanny. Those speakers go on standby and will auto power on based on input signals. Sometimes, when they turn on, they don't always power on at the same exact time, one speaker will come on before the other. And my God, each time that happens, I am reminded the grandness of the imaging.

How in the world can you listen to such impeccable imaging in mono?

I believe during subjective impression, the individual should also conduct listening in stereo in addition to whatever mono evaluation is being done.

I also believe there is value in subjective impression (either before or after the measurements, as I can see possible bias going either way) as there are some characteristics that the measurements alone cannot tell you, to use the example of 3D imaging again. I am not saying to trust the subjective impression 100%, on the contrary, take it as a grain of salt. And know that you can fast forward if you are hearing too much subjective diahera.

I personally don't listen only in mono when testing a speaker, but I do think without a shadow of a doubt it makes it wayyyy easier to detect flaws in a speaker reliably. I agree that it would be better to do both given enough time, but to me it makes total sense that Amir, given limited time to listen to a speaker and with a primary focus on measurements, would choose to do things this way.

The thing is wider speakers sound wider in mono too, given sufficient sidewall reflections in the room. This has become increasingly apparent to me recently as I search for a good center channel...

There is some complication when the music has a lot of encoded spatial effects that might lessen this gap. But overall, I'm 100% in on the usefulness of mono listening.

 

[JSmith]

Why do ASR members run to post the latest link to Erin's reviews on another person's forum

That's why I just ripped screenshots of the measurements in his video, Klippel data only, as that is of interest here.


JSmith

 

[Music1969]

Note that it's actually easier to detect these at higher frequencies, presumably up to a point.

Those listening tests are done in very good acoustical rooms right?

Not in normal living room ?

 

[moonlight rainbow dream]

Pleas educate me.

Genelec advises to raise their monitors from the floor:
"Monitors should always be aimed towards the listening position. The higher the monitor is from the floor, the lower is the reflection induced frequency response irregularities. However, half room height placement should be avoided, as at low frequencies the ceiling is typically also a reflective surface."

That statement seems diametrically oppose to the idea of floor bounce reduction due to low placement of the woofer.
Which one is true?

The answer has practical implications to me: if I place my Genelec mains vertically, the distance to the floor is mere 16cm. Horizontal placement increases that distance to 53cm, but introduces other complications in my setup.

Don't think about it in terms of floor bounce "reduction" or "increase". That can only be done via absorption. What you're doing when you move the driver higher or lower is changing the frequency at which the first and strongest cancellation occurs. When you move the driver further from the floor the frequency goes down; when you move it closer the frequency rises. The strategy is then to carefully place the drivers (bass elements and midrange) such that the floor bounce frequency is pushed out of the driver's passband. I would guess that Genelec's recommendation is based on the understanding that frequency response aberrations are less audible the farther you push them out of the critical midrange where human hearing is most sensitive.

You can try playing with this calculator:

Floor/Ceiling Reflection Calculator

I would be skeptical of claims of hearing this phenomenon or absence of without a controlled experiment.

 

[napilopez]

Those listening tests are done in very good acoustical rooms right?

Not in normal living room ?

"good" not really haha, but you are totally correct that this not from a study in a living room, these are from anechoic listening tests.

However, that doesn't mean they are less audible in a standard living room. When it comes to resonances in particular, it's likely the opposite. All from 4.6.2 in Toole's book:

"Continuous sounds are more revealing of resonances than isolated transient sounds. The repetitions (reflections) necessary to lower thresholds (increase our sensitivity to resonances) can either be in the recorded sounds themselves or contributed by the listening room. The conclusion reported there was that we appear not to be sensitive to the ringing in the time domain (at frequencies above about 200 Hz at least), but to the spectral feature—the peak."

"for demonstrations intended to impress, use simple sounds—solo voice, guitar, small combos and so on, especially with little reverberation—and, if possible, use a relatively dead room. To look for problems, listen to complex instrumental combinations with wide bandwidth and reverberation—and listen in a room with some reflections.

"Finally, all of these threshold determinations were done in anechoic listening conditions. As mentioned, these thresholds may be even lower when listening in reflective rooms. However, if the thresholds were determined using signals incorporating significant repetitions (e.g., reverberation), the effect of the listening environment is minimal."

Not looking to argue the specifics dB and Q at which something is audible, mind you, just pointing out that small deviations have been tested to be audible, are possibly more audible in a reflective room, and that wide q deviations are generally more audible at lower dB amplitude than narrow ones.

 

[CleanSound]

There is some complication when the music has a lot of encoded spatial effects that might lessen this gap. But overall, I'm 100% in on the usefulness of mono listening.

If a reviewer were to provide me their subjective feedback only in mono, I would be missing many key points of evaluation.

I'm actually not sure if Amir's subjective feedback portion is only mono, I believe so? For instance, in Amir's Neumann KH 120 review, there was no mention of the imaging and that is by far one of the most awesome characteristics of these speakers. It's mind bending awesome, but yet, not one word mentioned about it, presumably it wasn't detected as a result of mono only listening?

 

[amirm]

Let me explain why: I have a pair of Neumann KH 120, for those who have heard them will tell you about the 3D imaging, it is uncanny. Those speakers go on standby and will auto power on based on input signals. Sometimes, when they turn on, they don't always power on at the same exact time, one speaker will come on before the other. And my God, each time that happens, I am reminded the grandness of the imaging.

That is the reason you do NOT want to evaluate a speaker in stereo! Instead of properly focusing on tonality, your mind shifts to admiration of spatial qualities of stereo recording. It is for this reason that surround sound with its more captivating and "real" version of this, further neuters listener ability to tell the good speaker from bad.

We have a direct analogy in video. Professional monitors have a switch to turn off Chroma (color) so that they can scrutinize the luma (Black and White) signal. This showed a lot of problems in the source video that were not visible in full color images. You could argue no one watches in black and white but if we are critical and want to find real problems, then we better do our best to find them. After all, what may be hidden now (with color or stereo in audio), may come right out in other scenarios we have not tested.

And remember, according to Dr. Toole, no speaker that has done well in mono, has lost in stereo. Goodness in mono directly translates to stereo.

 

[kemmler3D]

Those listening tests are done in very good acoustical rooms right?

Not in normal living room ?

I tend to agree that a dB or two here or there is detectable if you're very familiar with your listening material.

When I used to demo speakers for my job, I had (still have) a playlist of tracks, mostly with flawed mixes. Especially when certain frequency ranges are over-or under-emphasized, you can quickly tell where the peaks and valleys of an FR are.

I don't do critical listening much (at all) anymore so I probably couldn't do this on demand anymore, but I used to be able to find (say) a 2dB dip around (say) 200hz (or wherever) by ear. Not that I was 100% on this, but my listening notes were usually borne out by measurements later. So it's not impossible to do by any means, especially for someone that does as much critical listening to good speakers as Erin does.

I know this is just an anecdote, but I wanted to explain why it's not hard for me to believe he can pick out a 1 or 2 dB dip by ear.

 

[CleanSound]

That is the reason you do NOT want to evaluate a speaker in stereo! Instead of properly focusing on tonality, your mind shifts to admiration of spatial qualities of stereo recording. It is for this reason that surround sound with its more captivating and "real" version of this, further neuters listener ability to tell the good speaker from bad.

We have a direct analogy in video. Professional monitors have a switch to turn off Chroma (color) so that they can scrutinize the luma (Black and White) signal. This showed a lot of problems in the source video that were not visible in full color images. You could argue no one watches in black and white but if we are critical and want to find real problems, then we better do our best to find them. After all, what may be hidden now (with color or stereo in audio), may come right out in other scenarios we have not tested.

And remember, according to Dr. Toole, no speaker that has done well in mono, has lost in stereo. Goodness in mono directly translates to stereo.

I am saying the evaluation should be done in both, not just mono alone.

And also with the Neumann specifically, they are pair matched to .5 dB. And I believe that also translated to sonic goodness. Which I believe deserves evaluation.

 

[nerdemoji]

@amirm You have probably have been asked this question before, sorry. What would you say is the reason for your subjective speaker reviews/what does if add to the review?

 

[AudioJester]

Has anyone measured Roon EQ at dac outputs. Sometimes I dont feel its doing what the graph looks like, although probably speaker/room interacions??

 

[amirm]

@amirm You have probably have been asked this question before, sorry. What would you say is the reason for your subjective speaker reviews/what does if add to the review?

It serves a few functions:

1. Sub-bass handling. I have a track that clearly reveals if deep bass is distorted, not played, etc. Interpreting this from distortion graphs is hard as just about every speaker has a lot of bass distortion. It is easier to just listen to this track and try to do the perceptual analysis from THD.

2. Using EQ to determine audibility of frequency response errors. To the extent it does, then it provides a simple method for owners to fix their speakers using those filters.

3. Dynamic playback ability. I power speakers with a very high power amplifier. I then push them hard. Best speakers handle that with ease until my ears give up. Others, start to complain in a very audible way. I should probably get more strict here with SPL meters and such but as a general test, it works.

4. For me personally to keep learning the correlation between frequency response measurements and subjective experience.

5. Providing a "wet thumb in the air" assessment of whether I like what I am hearing.

 

[amirm]

Has anyone measured Roon EQ at dac outputs. Sometimes I dont feel its doing what the graph looks like, although probably speaker/room interacions??

I have once or twice with headphones and it is pretty close to what is intended.

 

[hmt]

You mean justify it after the fact as to look right? That is the very point I made.


The heck you are talking about? Subjective testing must be done in mono based on "science." Room effects need to be constant between speakers which is impossible to do with a pair of speakers. Room EQ must be deployed carefully to not penalize speakers due to room modes. There is nothing remotely scientific about putting two speakers in a room and playing Michael Jackson.

Casual, ad-hoc, listening tests need to be an appendix to a set of measurements. Reverse the equation and claim validity of your purely non-scientific listening and you have lost the plot with respect to what is science, and what is not.

But why all the drama. I personally also dont give much weight on the subjective part but the criticism is going a little overboard. Erin doesnt claim to be scientifically accurate with his subjective evaluation. He also puts several disclaimers.

 

[DanielT]

Don't think about it in terms of floor bounce "reduction" or "increase". That can only be done via absorption. What you're doing when you move the driver higher or lower is changing the frequency at which the first and strongest cancellation occurs. When you move the driver further from the floor the frequency goes down; when you move it closer the frequency rises. The strategy is then to carefully place the drivers (bass elements and midrange) such that the floor bounce frequency is pushed out of the driver's passband. I would guess that Genelec's recommendation is based on the understanding that frequency response aberrations are less audible the farther you push them out of the critical midrange where human hearing is most sensitive.

You can try playing with this calculator:

Floor/Ceiling Reflection Calculator

I would be skeptical of claims of hearing this phenomenon or absence of without a controlled experiment.

Hm, I wonder if dual bass drivers would reduce the problems that SBIR gives, first reflections,cancellation effects:

Edit:
I am stupid. Of course, I could just as easily use the FR10 as an example of my thoughts on double bass drivers:



Considering various carpets, coffee tables and more stuff in a normal furnished listening/living room, floor reflections should be less of a problem than ceiling reflections. At least that's what I'm guessing.
Best not to mention the first lateral reflex being or not. There was an intense debate about it a while ago (over dampened listening room = boring dry sound or precise and accurate).

In any case Peter Snell designed his Snell 1 with the aim of reducing, eliminating floor reflections. It would be interesting to hear a couple of these, or speakers with a similar construction principle:

 

[DanielT]

It serves a few functions:

1. Sub-bass handling. I have a track that clearly reveals if deep bass is distorted, not played, etc. Interpreting this from distortion graphs is hard as just about every speaker has a lot of bass distortion. It is easier to just listen to this track and try to do the perceptual analysis from THD.

2. Using EQ to determine audibility of frequency response errors. To the extent it does, then it provides a simple method for owners to fix their speakers using those filters.

3. Dynamic playback ability. I power speakers with a very high power amplifier. I then push them hard. Best speakers handle that with ease until my ears give up. Others, start to complain in a very audible way. I should probably get more strict here with SPL meters and such but as a general test, it works.

4. For me personally to keep learning the correlation between frequency response measurements and subjective experience.

5. Providing a "wet thumb in the air" assessment of whether I like what I am hearing.

That's how it is with testing what the stuff can handle by pushing the pedal to the metal. It reveals things.

Besides, it's something we men like to do regardless of the stuff being tested.

 

[sigbergaudio]

Pleas educate me.

Genelec advises to raise their monitors from the floor:
"Monitors should always be aimed towards the listening position. The higher the monitor is from the floor, the lower is the reflection induced frequency response irregularities. However, half room height placement should be avoided, as at low frequencies the ceiling is typically also a reflective surface."

That statement seems diametrically oppose to the idea of floor bounce reduction due to low placement of the woofer.
Which one is true?

The answer has practical implications to me: if I place my Genelec mains vertically, the distance to the floor is mere 16cm. Horizontal placement increases that distance to 53cm, but introduces other complications in my setup.

I'm not sure which Genelecs you have, but it would be poor advice to suggest to put most Genelecs on the floor for quite obvious reasons. Unless you are also sitting on the floor, the tweeter would be way too low.

 

[sergeauckland]

I am saying the evaluation should be done in both, not just mono alone.

And also with the Neumann specifically, they are pair matched to .5 dB. And I believe that also translated to sonic goodness. Which I believe deserves evaluation.

One test that I think is essential in evaluating loudspeakers is pair-matching. Accurate stereo relies on the two loudspeakers having very closely matching frequency responses. Any deviation results in the stereo image widening as the various frequencies in the signal reproduce at different levels between the two loudspeakers. That's why some manufacturers, like KEF, specify the closeness of matching. Measuring one loudspeaker only, tells one little about how well a pair will work in stereo, so measuring both, I think, is essential in a full review. It matters less if the frequency response of the single loudspeaker is very flat, as then it can be reasonably assumed that pair-matching will be good, but when the frequency response is poor, as so many loudspeakers are these days, then it's an unreasonable assumption that pair-matching will be good, even if the frequency response is actually quite poor.

I appreciate the logistical difficulties in having to transport and measure two loudspeakers rather than just one, but nevertheless, I think any review is incomplete without a measure of pair-matching.

S.