Evidence-based Speaker Designs

[tuga]

The same EBU/ITU recommendations also recommend a fixed flat operational room response curve (see page 7 in your above document) which all us Toole readers know its not the right target, see here more http://www.aes.org/e-lib/browse.cfm?elib=17839 (free download)

It looks flat up to 2kHz only.

The "correct" target curve is another area where there's lack of consensus.

https://www.audiosciencereview.com/...r-favorite-house-curve.2382/page-2#post-67187

Here's a different one:

 

[andreasmaaan]

In regard to your last sentence/conclusion, by effects, do you mean negative effects?

I actually meant simply any effect on localisation, whether positive or negative (although I was also suggesting that, where localisation effects are significant (i.e. shifting/spreading), they are likely to be perceived as positive by many listeners.

With reference to the EBU/ITU recommendations you quote, I don't think they are necessarily contradictory, I just think they are overly simplistic.

The European Broadcasting Union (EBU) and Telecommunications Union (ITU) recommend that you treat reflections earlier than 15 ms relative to the direct sound. Your goal is to tame these early reflection levels so they are at least 10 dB below the direct sound for all frequencies in the range 1 kHz to 8 kHz. (>15 dB below is better)

I'd refer again to the data presented in the graph in my post #856.

Early sidewall reflections within 5 or 6ms, if present, are likely to be relatively high in level (-6dB or higher), and therefore may well cause disturbances to the primary image.

However, these would occur only in very small rooms, where both speaker and listener are within 2 to 2.5m of the sidewall, and where the directivity is sufficiently wide to create lateral energy at the reflection point almost equal to the on-axis energy.

But look at the graph in post #856 from around 6 or 7ms and later: the requisite SPL for a delayed sound to cause "shift or spreading of the primary image" is so high relative to the source that, in normal rooms, lateral reflections are simply not going to manage it.

In other words, the thresholds for image shift/spread are significantly higher than the EBU/ITU recommendations imply.

Of course, all of the above is signal dependent, so it might be generously assumed that the EBU/ITU wish to leave a large buffer to ensure absolutely no shift/spread, even with pathological signals (and of course what I said earlier about control room objectives differing from listening room objectives also applies).

 

[Liberal Democracy]

Stereo is a distraction when evaluating loudspeakers.

Sorry in advance.

 

[Liberal Democracy]

Early Reflections Are Not Beneficial
by @Ethan Winer

This article is meant mainly as a rebuttal to those who believe that early reflections enhance sound quality in a typical home-sized listening room. By "home-sized" I mean a room less than about 20 feet wide, where reflections from nearby surfaces arrive soon after the direct sound from the loudspeakers. I assume most readers already know about early reflections and reflection points, but if not, this RFZ Article on the RealTraps web site explains in detail. One proponent in favor of early reflections is loudspeaker expert Dr. Floyd Toole, whose 2008 book "Sound Reproduction" has been deservedly well received. I'll address Dr. Toole's position later on.

http://ethanwiner.com/early_reflections.htm

This is wrong thinking , there's a difference between creating a norm to apply in a varied acoustic environment and creating a acoustic space to enjoy recordings via loudspeakers.

No recording engineer i know creates recordings assuming we live in dead environments.

It's not natural. Now if you could guarantee all listeners were in a controlled environment that would be different.

 

[tuga]

I actually meant simply any effect on localisation, whether positive or negative (although I was also suggesting that, where localisation effects are significant (i.e. shifting/spreading), they are likely to be perceived as positive by many listeners.

Comb filtering might be perceived as positively "lively" or "dynamic" or "exciting" by many listeners.

Many listeners are not particularly demanding, most audiophile newbies only focus of their unsophisticated attention on loudness and on bass extension and intensity.

Why is it wrong to like a BBC-dip or low- even-order harmonic distortion and other types of "euphonic" seasoning but not side-wall reflections?

Early sidewall reflections within 5 or 6ms, if present, are likely to be relatively high in level (-6dB or higher), and therefore may well cause disturbances to the primary image.

However, these would occur only in very small rooms, where both speaker and listener are within 2 to 2.5m of the sidewall, and where the directivity is sufficiently wide to create lateral energy at the reflection point almost equal to the on-axis energy.

This would be the typical British sitting room then...

 

[tuga]

Stereo is a distraction when evaluating loudspeakers.

Sorry in advance.

I agree. Except when you wish to evaluate stereo.

 

[Liberal Democracy]

I agree. Except when you wish to evaluate stereo.

You can't so let's leave that.
No one pretends otherwise.

 

[tuga]

This is wrong thinking , there's a difference between creating a norm to apply in a varied acoustic environment and creating a acoustic space to enjoy recordings via loudspeakers.

No recording engineer i know creates recordings assuming we live in dead environments.

It's not natural. Now if you could guarantee all listeners were in a conteoll environment that would be different.

There's a difference between using an audio system which acurately reproduces the mix for monitoring purposes and putting toghether a system for listening enjoyment of recorded music.

See what I did there? It's Science.

 

[thewas]

The "correct" target curve is another area where there's lack of consensus.

https://www.audiosciencereview.com/...r-favorite-house-curve.2382/page-2#post-67187

If you read above paper of Toole (or even better his book) you will see that all those fixed room target curves don't really make sense as they depend on the directivity of the loudspeaker, reverberation characteristics of the room and the listening distance. The only target which make sense and perceived as neutral is flat direct sound and smooth directivity and reverberation characteristics and in the modal region you want also to EQ peaks due to room modes.

Here is more reading material about it https://www.audiosciencereview.com/...ut-room-curve-targets-room-eq-and-more.10950/

 

[Liberal Democracy]

There's a difference between using an audio system which acurately reproduces the mix for monitoring purposes and putting toghether a system for listening enjoyment of recorded music.

See what I did there? It's Science.

No one here conflates those two distinct things .

 

[andreasmaaan]

Early Reflections Are Not Beneficial
by @Ethan Winer

This article is meant mainly as a rebuttal to those who believe that early reflections enhance sound quality in a typical home-sized listening room. By "home-sized" I mean a room less than about 20 feet wide, where reflections from nearby surfaces arrive soon after the direct sound from the loudspeakers. I assume most readers already know about early reflections and reflection points, but if not, this RFZ Article on the RealTraps web site explains in detail. One proponent in favor of early reflections is loudspeaker expert Dr. Floyd Toole, whose 2008 book "Sound Reproduction" has been deservedly well received. I'll address Dr. Toole's position later on.

http://ethanwiner.com/early_reflections.htm

I like Ethan's work, but there are a number of things I disagree with in this article. Here are just a couple of them:

But when played back in a small untreated room, the strong early reflections from nearby surfaces drown out the larger sounding reverb in the recording. This makes the music sound smaller and narrower, not larger and wider. As proof, headphones are fully anechoic having no reflections, yet music generally sounds larger on headphones than through loudspeakers.

Firstly, the research is quite clear that reflections need to be very early and very loud indeed to "drown out" direct sound, or even to shift/spread the perceived image.

Secondly, I am not someone who finds that music (other than potentially binaural recordings) sounds anything close to larger on headphones, so I just find that comment utterly strange.

One common myth is that rooms used mainly for playing stereo music should be treated differently, and less, than home theater rooms. But a big part of most movie sound tracks is music! Whether listening to music alone or as part of a movie, everything should come through clearly, with a flat frequency response and low distortion, and without strong early reflections clouding the sound.

It's not about movies vs music, but about stereo reproduction vs multichannel reproduction. In the latter case, the cues for spaciousness and envelopment are present in the recordings themselves. In the former case, they may need to be helped by the room.

However, I do agree with his general view that the smaller the room, the more likely it will be a good idea to absorb first reflections.

Why is it wrong to like a BBC-dip or low- even-order harmonic distortion and other types of "euphonic" seasoning but not side-wall reflections?

"Wrong"?

Firstly, all I've argued on this point is that (1) the evidence is stronger that most listeners don't prefer it and (2) it's a decision often best left to the engineer on a recording by recording basis.

Secondly, I am not arguing that the presence of strong sidewall reflections is more accurate to the source than otherwise.

I'm simply arguing (1) that most listeners prefer them because they tend to enhance envelopment, perceived spaciousness, and in some cases also apparent source width, and (2) that only in extreme cases do they significantly/negatively affect localisation (≈ imaging).

Moreover, we all prefer there to be some reflections present in our listening rooms. If you call my preferred reflections "euphonic" then you must call any reflections at all "euphonic" too, right?

And if you're against euphony of any kind, you must then be against reflections (and so on and so on ).

 

[tuga]

No one here conflates those two distinct things .

People are defending that distortion which results from wall reflections is acceptable because one or a dozen polls seem to indicate that people enjoy them.

The floor-bounce cancellation can be perceived as "fast", "tight" bass, as would a room null in the same frequency band. Some people like that too.

 

[andreasmaaan]

People are defending that distortion which results from wall reflections is acceptable because one or a dozen polls seem to indicate that people enjoy them.

C'mon @tuga, if by "people" you mean me and others, you must acknowledge that my defence of sidewall reflections is based on a lot more than just loudspeaker preference studies?

The floor-bounce cancellation can be perceived as "fast", "tight" bass, as would a room null in the same frequency band. Some people like that too.

This is a new topic obviously, but do you have any research to share on this?

 

[Liberal Democracy]

People are defending that distortion which results from wall reflections is acceptable because one or a dozen polls seem to indicate that people enjoy them.

The floor-bounce cancellation can be perceived as "fast", "tight" bass, as would a room null in the same frequency band. Some people like that too.

Don't forget it's humans that are listening to this sound , humans that live in reflective environments . Didn't used to , well shall we talk about acoustic environments and how they have informed music ?

It's quite simple as i see it , you make the off axis sound match as best you can to on axis so these reflections whenever they arrive don't cause any confusion in the brain .

Confusion as in ' is this a new sound or the one I heard just a millisecond ago. '

Visual affirmation helps IME , I relate sounds in space better when there's visual clues to assign the sounds with.

Can't just be me.

 

[tuga]

I like Ethan's work, but there are a number of things I disagree with in this article. Here are just a couple of them:


Firstly, the research is quite clear that reflections need to be very early and very loud indeed to "drown out" direct sound, or even to shift/spread the perceived image.

Secondly, I am not someone who finds that music (other than potentially binaural recordings) sounds anything close to larger on headphones, so I just find that comment utterly strange.


It's not about movies vs music, but about stereo reproduction vs multichannel reproduction. In the latter case, the cues for spaciousness and envelopment are present in the recordings themselves. In the former case, they may need to be helped by the room.

However, I do agree with his general view that the smaller the room, the more likely it will be a good idea to absorb first reflections.



"Wrong"?

Firstly, all I've argued on this point is that (1) the evidence is stronger that most listeners don't prefer it and (2) it's a decision often best left to the engineer on a recording by recording basis.

Secondly, I am not arguing that the presence of strong sidewall reflections is more accurate to the source than otherwise.

I'm simply arguing (1) that most listeners prefer them because they tend to enhance envelopment, perceived spaciousness, and in some cases also apparent source width, and (2) that only in extreme cases do they significantly/negatively affect localisation (≈ imaging).

Could you elaborate on why "only in extreme cases do they significantly/negatively affect localisation (≈ imaging)"?

Moreover, we all prefer there to be some reflections present in our listening rooms. If you call my preferred reflections "euphonic" then you must call any reflections at all "euphonic" too, right?

And if you're against euphony of any kind, you must then be against reflections (and so on and so on ).[/QUOTE]

Euphonic means pleasing. In other words, any distortion that you find pleasing is euphonic to you. It's a matter of taste. An it doesn't really matter whether a particular type of distortion is preferred by most listeners or not.

One can even be against euphonic distortion from a phylosophical stand point yet still enjoy the effect of some types of euphonic distortion.


The argument for wide-dispersion is based on the fact that existing research seems to indicate that most listeners prefer it. In my view, even though the polling data was gathered in a scientific manner, the fact that it is a taste-driven opinion is not enough to justify the validity of wide-dispersion.

Thanks for the interesting debate.

 

[tuga]

C'mon @tuga, if by "people" you mean me and others, you must acknowledge that my defence of sidewall reflections is based on a lot more than just loudspeaker preference studies?

Please correct me if I misunderstood but I would say that have used research to refute the claim that side-wall reflections affect imaging (and the recreation of the recorded cues) which is not the same as using it to defend side-wall reflections.

 

[maverickronin]

Im not sure the point you are making with your last paragraph. Perhaps you could elaborate. Again though, why are you assuming it should be "sharply placed"?

Here are 3 recordings of myself talking and walking around a room with different microphone configurations. It shows the differences in spatial effect. Different configurations lead to different localisation.

I haven't had time to listen to those samples you uploaded yet, but I think I know why we are talking past each other.

You seem to be only assuming the micing and recording of a live event while I'm talking about studio mixes which can and do place phantom sources very sharply.

 

[andreasmaaan]

Could you elaborate on why "only in extreme cases do they significantly/negatively affect localisation (≈ imaging)"?

For the reasons I mentioned in post #856, where I wrote:

Firstly, delayed sounds actually need to be very high in level to interfere with localisation dominance. It's unlikely that lateral reflections in a typical small room will be loud enough (or that delays will be long enough) to have a significant effect, except in pathological cases (e.g. small, asymmetrical rooms in which one speaker is very close to a sidewall while the other is not).

I know I've trotted out this graph from one of Toole's papers before, but it really is an excellent summary of the main research on the subjective effects of delayed sounds...

As you can see, it's unlikely that lateral reflections in a typical small room would be either high enough in level or late enough in time to cross the "shift or spreading of primary image" threshold (and moreover, even where they are loud and early enough to cause image shift or spreading, this is likely to be preferred by many listeners).

Secondly, I'm arguing that strong(ish) lateral reflections, even when they don't result in shift or spreading of the primary image, do tend to contribute to perceptions of spaciousness and envelopment.

In other words, , in most small rooms, the effects of strong lateral reflections on image localisation will be either non-existent or minimal relative to the effects on spaciousness and envelopment.

I argued that a room would need to be very small, and speakers and listener suboptimally placed in it, for lateral reflections to fall within this zone in which they would be likely to affect localisation (by shifting/spreading the image):

I consider rooms in which early lateral reflections arrive that fast and that strong to be "extreme".

Euphonic means pleasing. In other words, any distortion that you find pleasing is euphonic to you. It's a matter of taste. An it doesn't really matter whether a particular type of distortion is preferred by most listeners or not.

One can even be against euphonic distortion from a phylosophical stand point yet still enjoy the effect of some types of euphonic distortion.

It's not the "euphonic" part I object to; it's the "distortion".

If lateral reflections contribute positively to perceived spaciousness and envelopment without negatively affecting localisation (or anything else), then they can't be characterised as "distortion" (unless, as I argued earlier, you wish to characterise all reflections as distortion).

Please correct me if I misunderstood but I would say that have used research to refute the claim that side-wall reflections affect imaging (and the recreation of the recorded cues) which is not the same as using it to defend side-wall reflections.

Fair enough. Of the two long papers I posted, sure, one was more concerned with localisation, but the other also dealt in detail with research into the connection between lateral reflections and perceived spaciousness and envelopment. The only reason I didn't discuss the positive effects of lateral reflections in much detail in my actual posts was that nobody seemed to take issue with it.

The argument for wide-dispersion is based on the fact that existing research seems to indicate that most listeners prefer it. In my view, even though the polling data was gathered in a scientific manner, the fact that it is a taste-driven opinion is not enough to justify the validity of wide-dispersion.

Thanks for the interesting debate.

For the reasons I've discusses above, my views are not particularly based on research into listener preference for loudspeakers. Rather, they are based more on the psychoacoustics of delayed sound perception.

Thanks too for the interesting discussion

 

[March Audio]

]

A mirror relIf we place a mirror in the early reflection area


I've skim-read parts and I have many of his papers too.

Really? Then I'm not sure why you are posting some of the things you are.

Perhaps a more serious read might be appropriate.

 

[Koeitje]

The ranked order of the average preferences remains the same. That's not the same thing. It's possible for other trends to be hidden within those averages which would require preference rating subscores to tease out.

That is true, the difference in the actual scores is actually bigger for mono.