Listening in an Anechoic Chamber - a report

[Thomas_A]

I found another interesting quote from Sigfried Linkwitz: "I have observed occasionally that highly directional loudspeakers in a highly absorptive environment with low reflections can produce a center phantom source, a female voice that floats in front of the line between the loudspeakers. In an anechoic room the center phantom source can even manifest inside the head, which mimics headphone listening" (https://www.linkwitzlab.com/TMT-Leipzig'10/TMT-Hearing spatial detail.pdf)

Cues for distance perception includes reflections. So that is why I think reflections are needed in room.

 

[dasdoing]

again, some audiophiles seam to have a big misscinception about distance and space in music playback. those things are allready "encoded" in the recordings.
"but, when I hear music on headphones it is all in between my ears".
It's because headphones are not realy stereo in the sense of the concept of recording>playback.
actualy there is space in headphone listening...it is just that hardpanned stuff is coming from the side which breaks 90% of the ilusion

 

[EEE272]

again, some audiophiles seam to have a big misscinception about distance and space in music playback. those things are allready "encoded" in the recordings.
"but, when I hear music on headphones it is all in between my ears".
It's because headphones are not realy stereo in the sense of the concept of recording>playback.
actualy there is space in headphone listening...it is just that hardpanned stuff is coming from the side which breaks 90% of the ilusion

I agree with you but I am not sure it explains why the sound would be in your head in an anechoic chamber. Here, you do have crosstalk.
Hence, there must be something about perceived distance that relates to room reflections.

 

[Newman]

Toole gives a lot of weight to Klippel's experiments on this topic, wherein he found that 50% of the perception of naturalness in recorded music playback came from the feeling of space, and 70% of the perception of pleasantness came from that same feeling of space. And Toole notes that laterally reflected sound is the main contributor to a feeling of space.

'Going anechoic' basically denies oneself access to that important 50-70%.

I also note the 2019 Kaplanis/Bech/Lokki et al paper, although it did conclude that low RT60 RT30 is preferred, it was not without limit. An optimal figure of 0.4 sec was noted, below which preference falls. An anechoic room has an RT60 RT30 of 0.0 sec. Too low.

So, Klippel leads us to think that some lateral reflections are important, and Kaplanis et al is suggesting that we don't want too much of it: RT/T60 RT/T30 of 0.4 sec would need some treatment in typical home hifi environments.

[Edit: I just noticed the paper says “RT” as short for the mean RT30, so I have edited my text above. Doing so only makes the case for low RT weaker.]

Goldilocks would say "I told you so".

cheers

 

[Thomas_A]

again, some audiophiles seam to have a big misscinception about distance and space in music playback. those things are allready "encoded" in the recordings.
"but, when I hear music on headphones it is all in between my ears".
It's because headphones are not realy stereo in the sense of the concept of recording>playback.
actualy there is space in headphone listening...it is just that hardpanned stuff is coming from the side which breaks 90% of the ilusion

Yes but the reflections in the recording give cues to the recording event. A large or small hall can be heard etc. Even binaural recordings with HPs are difficult to get distance and direction due to lack of head movement etc

 

[Frgirard]

Yes but the reflections in the recording give cues to the recording event. A large or small hall can be heard etc. Even binaural recordings with HPs are difficult to get distance and direction due to lack of head movement etc

STEREO SOUND RECORDING
Regarding stereo sound recording, there are several ways to do it with microphones whose characteristics may vary but they are all based on one of the following 2 main principles. Either intensity stereo or phase stereo.

THE INTENSITY STEREO
Its principle is to recreate the sensation of stereo by the difference in intensity (understand sound volume) that each microphone will capture. Depending on the position of the source, one of the 2 microphones will be more directly oriented towards it at the expense of the 2nd microphone and will therefore capture more sound level.

Intensity stereo will have the advantage of having good mono compatibility because the capsules of the 2 microphones used are superimposed, but the stereo image it reproduces will not be as natural as when using phase stereo.
The XY pair uses intensity stereo. It is said to be phase coincidence.

LIVE STEREO
In the case of phase stereo, the 2 microphones used are not superimposed and therefore the sound signal will not reach the 2 microphones at the same time. Sound has a speed of 340 m/s in air so these time differences will be minimal but sufficient to recreate the feeling of stereo width.
Phase stereo is used in the ORTF and AB couple sound pickups. It is not recommended when the recording to be made must have good mono compatibility.

La prise de son stéréo (Part.1) – MixCorner

il existe plusieurs façons de faire avec des micros dont les caractéristiques peuvent varier mais elles sont toutes basées sur l'un des 2 grands principes suivants. Soit la stéréo d'intensité, soit…

www.mixcorner.com

 

[youngho]

I also note the 2019 Kaplanis/Bech/Lokki et al paper, although it did conclude that low RT60 is preferred, it was not without limit. An optimal figure of 0.4 sec was noted, below which preference falls. An anechoic room has an RT60 of 0.0 sec. Too low.

Actually, it was the opposite. "It is also important to note that the current results suggested that a negative preference is apparent for acoustical conditions with RT higher than 0.4 s...It is evident that a critical value seems to exist, close to the recommended mean value of 0.4 s, above which assessors’ preference degrades" (bolding mine).

 

[Newman]

Yes but the key point is that negative preference only begins at RT30 higher than 0.4 sec.

They didn’t say the lower the better. They said you have to get above RT30 of 0.4 sec before any negative preference arises.

And RT30 of 0.4 sec is pretty close to RT60 of 0.8 sec. Which is pretty easy to achieve in home environments with some non-extreme treatment.

 

[youngho]

Yes but the key point is that negative preference only begins at RT30 higher than 0.4 sec.

They didn’t say the lower the better. They said you have to get above RT30 of 0.4 sec before any negative preference arises.

And RT30 of 0.4 sec is pretty close to RT60 of 0.8 sec. Which is pretty easy to achieve in home environments with some non-extreme treatment.

Right, I agree with what you're saying here, but what you previously wrote, and I will quote, just like I did for the paper, was "An optimal figure of 0.4 sec was noted, below which preference falls." What they wrote was "A critical value seems to exist, close to the recommended mean value of 0.4 s, above which assessors’ preference degrades." Kaplanis does not suggest "we don't want too much of it," but rather "a negative preference is apparent for acoustical conditions with RT higher than 0.4 s."

There's a crucial difference between "optimal" (as in ideal or best possible) and "critical" (as in "threshold" or "boundary"), but most importantly is what I pointed out above, which is that "below" is the opposite of "above." Also, I don't see a need to paraphrase when direct quotes are available.

Toole has suggested "reverberation time" (unspecified in these links, but usually T60 in most of his and Olive's papers, from what I can tell) of 0.5 sec or below (https://www.audioholics.com/room-acoustics/room-reflections-human-adaptation and https://gearspace.com/board/showpost.php?p=15187387&postcount=61), and https://users.aalto.fi/~ktlokki/Publs/mst_laukkanen.pdf suggests different strokes for different folks.

EDIT: I wrote "0.5 sec or below," but I should have written "below 0.5 sec"

 

[Newman]

Yes that’s why I wrote my second post, which you agree with.

 

[youngho]

Yes that’s why I wrote my second post, which you agree with.

Ah, you had not acknowledged your error, so the second post seemed to be some sort of justification. RT60 of 0.8 is easy to achieve, but that is too high for most audio purposes, hence the references I provided.

 

[Newman]

Well your 2019 reference concludes, without paraphrasing, that their findings relating to assessors’ preferences may not reveal the judgments and sentiments of an average listener.

ie don’t quote them — at least not for the sort of discussion we are having here.

 

[youngho]

Well your 2019 reference concludes, without paraphrasing, that their findings relating to assessors’ preferences may not reveal the judgments and sentiments of an average listener.

ie don’t quote them — at least not for the sort of discussion we are having here.

Yes, I am aware of this limitation, which I mentioned already in post #69 of this thread (https://www.audiosciencereview.com/...ic-chamber-a-report.30274/page-4#post-1065746): "The Bech/Lokki JASA 2019 paper was relevant to listening rooms, but given that they used expert listeners, it's not surprising that a drier acoustic environment was overall preferred and still in line with Toole's statements."

You were the one to bring it up again in post #104 with a factually incorrect paraphrasing of the study's results. The reason I even mentioned them in the first place was in the context of responding to someone else's factually incorrect paraphrasing of another study by Lokki! That one focussed on concert hall acoustics, rather than listening rooms.

How about this? Try not to refer to a study with factually incorrectly paraphrasing of its results for "the sort of discussion we are having here," or for that matter, any discussion at all.

 

[Thomas_A]

STEREO SOUND RECORDING
Regarding stereo sound recording, there are several ways to do it with microphones whose characteristics may vary but they are all based on one of the following 2 main principles. Either intensity stereo or phase stereo.

THE INTENSITY STEREO
Its principle is to recreate the sensation of stereo by the difference in intensity (understand sound volume) that each microphone will capture. Depending on the position of the source, one of the 2 microphones will be more directly oriented towards it at the expense of the 2nd microphone and will therefore capture more sound level.

Intensity stereo will have the advantage of having good mono compatibility because the capsules of the 2 microphones used are superimposed, but the stereo image it reproduces will not be as natural as when using phase stereo.
The XY pair uses intensity stereo. It is said to be phase coincidence.

LIVE STEREO
In the case of phase stereo, the 2 microphones used are not superimposed and therefore the sound signal will not reach the 2 microphones at the same time. Sound has a speed of 340 m/s in air so these time differences will be minimal but sufficient to recreate the feeling of stereo width.
Phase stereo is used in the ORTF and AB couple sound pickups. It is not recommended when the recording to be made must have good mono compatibility.

La prise de son stéréo (Part.1) – MixCorner

il existe plusieurs façons de faire avec des micros dont les caractéristiques peuvent varier mais elles sont toutes basées sur l'un des 2 grands principes suivants. Soit la stéréo d'intensité, soit…

www.mixcorner.com

Understood, but what has that to do with distance and direction perception to the sound source when e.g. comparing HP and speaker/room listening?

 

[Newman]

How about this? Try not to refer to a study with factually incorrectly paraphrasing of its results for "the sort of discussion we are having here," or for that matter, any discussion at all.

I made a mistake and I made a later post that agreed with your correction. Not enough for you apparently! Even when the report may not be relevant to the general public, by the authors’ own admission, yet you carry on as if I was a climate denier. A bit of perspective please.

Here is my best understanding of / contribution to the broad issues of the thread:-

• The Klippel research is considered important, and says that a feeling of space, mainly caused by laterally reflected sound, is the single biggest factor in perceptions of naturalness and pleasantness.
• That doesn’t mean the more the better, but we surely don’t want to completely obliterate it! So, how much is enough? How much is too much?
• That’s where the 2019 Kaplanis et al paper might help. It found that preference only starts to fall after RT30 exceeds 0.4 sec (RT60 over 0.8 sec). It is too early to say that this applies to the general public, but even if it does, there is other evidence that the expert panels used by Kaplanis prefer a lower RT than the general public, so if the findings do turn out to have application to the general public, it will very likely be a higher RT before preference starts to fall.
• If one wants to take guidance from this for practical action today as a home listener, then one might do well to treat one’s room if its mean RT60 is over 1 sec. OTOH keep some reverberance in there, since it has an important role. When Toole attempted to provide some guidance on room treatment, he seemed to be suggesting about 25-35% of wall coverage with good quality absorption. He also seems to be saying that more socially friendly rooms can get by with much less treatment, intelligently applied, and still sound very enjoyable to the audiophile. He added a lot more devil-in-the-detail, too, but….read the book(s)!

PS as an aside, it is IMO strange that serious researchers choose to use RT measures in the 2019 paper to characterise domestic listening spaces that cannot create a statistically diffuse sound field due to being too small. What were they thinking? But I know we can all put a microphone in our room and run software that will give ”an RT number”, even though I question its choice…

cheers

 

[youngho]

I made a mistake and I made a later post that agreed with your correction. Not enough for you apparently! Even when the report may not be relevant to the general public, by the authors’ own admission, yet you carry on as if I was a climate denier. A bit of perspective please.

It's a small matter to say, "I'm wrong." I say it a lot. I even edited one of my posts above to indicate where I made a mistake. "Yes, but" is not the same thing. I decline to respond to you further, unless specifically requested, even though I have some thoughts on what you wrote in post #115. You have done well to shut down discussion on my part with multiple posts, like #112 above.

 

[Cote Dazur]

During the development of Silver 5L I spent two days in a 5m x 5m x 2m room

How was the perception of depth? The reason I ask is that today I placed speakers outside, I am very surprised that the left right image is great, as expected, but it’s absolutly flat. No depth. In my listening room, recording dependent of course, the sound has depth. I always thought having taken care of avoiding reflection, was the major reason for my glorious stereo imaging, width and depth. But here I am outside, zero reflection, no depth!

 

[sarumbear]

How was the perception of depth? The reason I ask is that today I placed speakers outside, I am very surprised that the left right image is great, as expected, but it’s absolutly flat. No depth. In my listening room, recording dependent of course, the sound has depth. I always thought having taken care of avoiding reflection, was the major reason for my glorious stereo imaging, width and depth. But here I am outside, zero reflection, no depth!

As you said. It depends on the source.

 

[Axo1989]

And RT30 of 0.4 sec is pretty close to RT60 of 0.8 sec. Which is pretty easy to achieve in home environments with some non-extreme treatment.

Not how those metrics work. RT30 is measured over the smaller decay delta, then multiplied. For a linear decay, the value is the same as RT60. In other words, RT30 is the method/protocol, RT is the metric. The purpose of RT30 is to give an equivalent result in a noisier environment, where RT60 doesn't work accurately.

The various RT measures are stated as “RTXX”, where XX is a number of dB. For example, RT20 is the RT value extrapolated from the time required for sound to decay by 20 dB,from -5 dB to- 25 dB. The initial 5 dB of decay is ignored, because in many rooms, it consists mainly of early reflections, and not true reverberation, thus spoiling the uniformity of the slope. So if the time required for sound to decay by 20 dB (from -5 to -25 dB) is 400 ms, we can extrapolate to find RT by multiplying 400 ms by 3, giving 1.2 seconds, since 20 dB is 1/3 of 60 dB, and the decay curve is presumed to be linear.

Tangentially—but apropos of the thread—I prefer a dry room. Mine measures ~200 ms above 60 Hz (the room has no additional treatment, the sonic characteristics are due to traditional Japanese architecture).

 

[sarumbear]

Not how those metrics work. RT30 is measured over the smaller decay delta, then multiplied. For a linear decay, the value is the same as RT60. In other words, RT30 is the method/protocol, RT is the metric. The purpose of RT30 is to give an equivalent result in a noisier environment, where RT60 doesn't work accurately.

It’s nice to see people who seems to know acoustics joining the conversation.