Genelec on audio science

[Jakob1863]

<snip> I, for one, didn't know that the professional standards for room curves are flat (or at least it didn't stick when I read dr Toole's book). That's an important thing to know, and now I know it.

I would have bet to have linked/mentioned some of these publications before, but maybe i missed it (here) - (note, you shall not post on more than one forum ):

For an informal (more or less) article, see this:

https://tech.ebu.ch/docs/techreview/trev_274-hoeg.pdf
(Hoeg et al., Subjective assessment of audio quality – the means and methods within the EBU)
Figure 4 in this article reflects what the (linked before) Mäkivärti study reports too.

The more formal publication on that is:

https://tech.ebu.ch/docs/tech/tech3276.pdf
(Listening conditions for the assessment of sound programme material: monophonic and two–channel stereophonic)

and more related to the content assesment:
https://tech.ebu.ch/docs/tech/tech3286.pdf
(Assessment methods for the subjective evaluation of the quality of sound programme material – Music)

Edit:
And of course the ITU-R BS.1116-3, which covers the ABC/HR test method and some technical requirement as well:

https://www.itu.int/dms_pubrec/itu-r/rec/bs/R-REC-BS.1116-3-201502-I!!PDF-E.pdf
(Methods for the subjective assessment of small impairments in audio systems)

But refocusing a discussion on to the persons involved ("assertive in your unsubstantiated opinions") is unnecessary and destructive for the discussion. It may be true or it may not be true in a particular case - but that really doesn't matter. Let's just stick to discussing the topics. There is absolutely no need for venturing beyond that.

I second that.
But, as posted the last time, it might be appropriate to open a new thread for a discussion about discussion styles/habits, as in a multinational forum with a broad range of individuals there always will be potential misunderstanding/misinterpretation.
As a general rule, trying to interpret other posts in the best possible way helps a lot.

 

[svart-hvitt]

THAT IS ONE HELL OF A RASH GENERALISATION.
I am not alone in this observation. This thread is composed of more erroneous thinking than beneficial information. Take my reply as feedback and not a personal attack. Geez.

I do focus on the content. Some can't seem separate their content from their 'id'.

For consistency re your post you seem to behaving in the same way that you accuse me of.

You said, I said. Back to the thread.

NB! This post got a bit long, again. The twitters among us are advised to ignore this post.

@Wombat , if I'm wrong on a matter, then please point it out. @Floyd Toole has been very forthcoming, even benevolent in his behaviour trying to replace confusion with facts and insights. And I'm truly very grateful for that!



Having said that, I think the "room discussion" has been fruitful for many of us. At least for me... We have had a good discussion on steady-state vs direct sound. And who knew, beforehand, that Genelec's room compensation software GLM can be described thus:

"The red and green GLM curves are estimates of the perceived direct sound frequency response, after measurements performed at one or more locations, before and after compensation".

The important words here, and that may be where Genelec deviate from other providers of room software, are "perceived direct sound". And if we look at Genelec's in-room product performance guide (https://www.genelec.com/sites/default/files/media/Studio monitors/Catalogues/genelec_monitors_in-room_performance.pdf), Genelec seem to be more focussed on letting the direct sound dominate in the listening position than you'd expect from a seller in a hifi-shop. So there seems to be a SS vs DS theme that deserves more factual discussion to elevate insight.

Further, @Thomas Lund stated that "The program is furthermore designed and updated based on research and data from professional applications, which may be different from home requirements".

Which means that GLM as of 2010 may not be exactly the same as GLM as of 2020. That makes it a bit hard to understand GLM perfectly well from an outsider's position.

If I were to try and sum up a broad categorization of views on speakers and room, I'd do it like this:

1. Only the speaker matters, because we listen through the room. See to it that you use a speaker that is flat, smooth and of low distortion in anechoic chamber.
2. The speaker's integration in the room can be improved in the lower frequencies up to ca. 300 Hz. Use SS measurements to make compensation filters in the lower frequencies.
3. The speaker's integration in the room can be improved in all frequencies. Use SS measurements to make compensation filters in all frequencies.
4. The speaker's integration in the room can be improved in the lower frequencies up to ca. 300 Hz. Use measurements of perceived direct sound to make compensation filters in the lower frequencies.
   
5. The speaker's integration in the room can be improved in all frequencies. Use measurements of perceived direct sound to make compensation filters in all frequencies.

I guess @Cosmik is a proponent of (1).

Further, I have the impression that view (2) is a sort of consensus on ASR. And (2) is the view which is most supported by Toole's book, if I'm not mistaken.

However, I believe view (3) is pervasive outside of ASR. Audiolense and Acourate offer this kind of all-frequency-corrections, right?

View (4) and (5) are "novel" and based on input from Genelec. I believe Genelec in practice is somewhere between (4) and (5), maybe closer to (4), because compensation at higher frequencies in GLM are small.

From a research program point of view, this separation of ways to approach audio has some interesting implications. View (1) is a way where measurements and theory are most easily combined because you can control for the room 100 percent. A research program that can control for all factors is an attractive one because of its simplicity, where the beautiful math is the final judge. My experience tells me to be wary of such research programs because they often fail when confronting a reality where the human factor plays a role. However, in the case of audio, we do know - thanks to Toole et al. - that view (1)gives us sound of very high quality.

So deviating from (1), to go further down the list, means that we're trying to improve something that's pretty good already. It's always risky to leave a robust model - where we understand "everything" - to replace it with something that is more complicated and potentially confusing.

Once we leave (1), two new factors arise: (a) the room and (b) the human factor (perception). Both (a) and (b) represent sources of standard errors in the empirical data material that researchers collect and study. It's naturally easier to control for (a) the room than (b) the human factor; (a) is quantized by microphones while (b) is quantized by measurements (mostly surveys, but in some research programs they also measure what happens in the brain) the complexity of human perception.

For the sake of simplicity it's tempting to settle with research program (1), characterized by beautiful mathematics. However, needless to say, this research program was more or less solved before the millennium shift; any improvements since then have been evolutionary, small steps, and researchers have been in the know of most of the compromises made in speaker production.

If you don't feel comfortable choosing the uncomplicated way of (1) - maybe because you have data and experiences that tell you the world is a bit more complicated - you need to venture below (1). And as soon as you leave the simple world of aesthetically attractive mathematics, disagreement will flower among scientists. Sometimes, this disagreement is based on one party being confused; on other occasions there's real room for disagreement and "confusion" is not the correct way to characterize a rival scientist.

Looking at speakers from a factor approach view, we may think of the following factors:

i) Speaker
ii) Room compensation
iii) Human

If we have full control over all factors, we could make sound that will optimize any individual's utility at any point in time. In reality, however, there is an emerging consensus on (i) , i.e. what constitutes a good speaker. Isn't this consensus seen more easily in the world of professional monitors where there is less deviation in design than in the "hifi" and "high-end" world?

But there is no overall consensus on what is (ii) good room compensation, as far as I understand. Olive's survey from 2009 found that room software was the source of differences in perception, but he couldn't drill down to the exact factors that gave rise to differences in perception related to room DSP, did he? And I am not sure we will ever be able to understand the individual human at all times; because your preferences change according to time of day, season, emotional state and many other things. So factor (iii) may always be approached by averaging (median) many individuals preferences over time. In other words, an aesthetically unattractive compromise may be awaiting us at the end of the road, after all. But is that a reason to settle for the mathematical beauty of option (1)?


Confusingly yours,

 

[Cosmik]

NB! This post got a bit long, again. The twitters among us are advised to ignore this post.

@Wombat , if I'm wrong on a matter, then please point it out. @Floyd Toole has been very forthcoming, even benevolent in his behaviour trying to replace confusion with facts and insights. And I'm truly very grateful for that!



Having said that, I think the "room discussion" has been fruitful for many of us. At least for me... We have had a good discussion on steady-state vs direct sound. And who knew, beforehand, that Genelec's room compensation software GLM can be described thus:

"The red and green GLM curves are estimates of the perceived direct sound frequency response, after measurements performed at one or more locations, before and after compensation".

The important words here, and that may be where Genelec deviate from other providers of room software, are "perceived direct sound". And if we look at Genelec's in-room product performance guide (https://www.genelec.com/sites/default/files/media/Studio monitors/Catalogues/genelec_monitors_in-room_performance.pdf), Genelec seem to be more focussed on letting the direct sound dominate in the listening position than you'd expect from a seller in a hifi-shop. So there seems to be a SS vs DS theme that deserves more factual discussion to elevate insight.

Further, @Thomas Lund stated that "The program is furthermore designed and updated based on research and data from professional applications, which may be different from home requirements".

Which means that GLM as of 2010 may not be exactly the same as GLM as of 2020. That makes it a bit hard to understand GLM perfectly well from an outsider's position.

If I were to try and sum up a broad categorization of views on speakers and room, I'd do it like this:

1. Only the speaker matters, because we listen through the room. See to it that you use a speaker that is smooth and of low distortion in anechoic chamber.
2. The speaker's integration in the room can be improved in the lower frequencies up to ca. 300 Hz. Use SS measurements to make compensation filters in the lower frequencies.
3. The speaker's integration in the room can be improved in all frequencies. Use SS measurements to make compensation filters in all frequencies.
4. The speaker's integration in the room can be improved in the lower frequencies up to ca. 300 Hz. Use measurements of perceived direct sound to make compensation filters in the lower frequencies.
   
5. The speaker's integration in the room can be improved in all frequencies. Use measurements of perceived direct sound to make compensation filters in all frequencies.

I guess @Cosmik is a proponent of (1).

Further, I have the impression that view (2) is a sort of consensus on ASR. And (2) is the view which is most supported by Toole's book, if I'm not mistaken.

However, I believe view (3) is pervasive outside of ASR. Audiolense and Acourate offer this kind of all-frequency-corrections, right?

View (4) and (5) are "novel" and based on input from Genelec. I believe Genelec in practice is somewhere between (4) and (5), maybe closer to (4), because compensation at higher frequencies in GLM are small.

From a research program point of view, this separation of ways to approach audio has some interesting implications. View (1) is a way where measurements and theory are most easily combined because you can control for the room 100 percent. A research program that can control for all factors is an attractive one because of its simplicity, where the beautiful math is the final judge. My experience tells me to be wary of such research programs because they often fail when confronting a reality where the human factor plays a role. However, in the case of audio, we do know - thanks to Toole et al. - that view (1)gives us sound of very high quality.

So deviating from (1), to go further down the list, means that we're trying to improve something that's pretty good already. It's always risky to leave a robust model - where we understand "everything" - to replace it with something that is more complicated and potentially confusing.

Once we leave (1), two new factors arise: (a) the room and (b) the human factor (perception). Both (a) and (b) represent sources of standard errors in the empirical data material that researchers collect and study. It's naturally easier to control for (a) the room than (b) the human factor; (a) is quantized by microphones while (b) is quantized by measurements (mostly surveys, but in some research programs they also measure what happens in the brain) the complexity of human perception.

For the sake of simplicity it's tempting to settle with research program (1), characterized by beautiful mathematics. However, needless to say, this research program was more or less solved before the millennium shift; any improvements since then have been evolutionary, small steps, and researchers have been in the know of most of the compromises made in speaker production.

If you don't feel comfortable choosing the uncomplicated way of (1) - maybe because you have data and experiences that tell you the world is a bit more complicated - you need to venture below (1). And as soon as you leave the simple world of aesthetically attractive mathematics, disagreement will flower among scientists. Sometimes, this disagreement is based on one party being confused; on other occasions there's real room for disagreement and "confusion" is not the correct way to characterize a rival scientist.

Looking at speakers from a factor approach view, we may think of the following factors:

i) Speaker
ii) Room compensation
iii) Human

If we have full control over all factors, we could make sound that will optimize any individual's utility at any point in time. In reality, however, there is an emerging consensus on (i) , i.e. what constitutes a good speaker. Isn't this consensus seen more easily in the world of professional monitors where there is less deviation in design than in the "hifi" and "high-end" world?

But there is no overall consensus on what is (ii) good room compensation, as far as I understand. Olive's survey from 2009 found that room software was the source of differences in perception, but he couldn't drill down to the exact factors that gave rise to differences in perception related to room DSP, did he? And I am not sure we will ever be able to understand the individual human at all times; because your preferences change according to time of day, season, emotional state and many other things. So factor (iii) may always be approached by averaging (median) many individuals preferences over time. In other words, an aesthetically unattractive compromise may be awaiting us at the end of the road, after all. But is that a reason to settle for the mathematical beauty of option (1)?


Confusingly yours,

Yes, I'm a (1). I reserve judgement on the bass compensation because everyone seems so adamant it is necessary but I have a suspicion that this might be partly to do with ported speakers that are near universal, plus my sealed, phase-corrected and time-aligned speakers have wider baffles than modern speakers and maybe all that makes them sound drier towards the low end.

I control the roll-off but nothing else, and they sound good to me - and of course they are therefore 'straight' and and the recording is not being messed about with arbitrarily. What the room does corresponds to what the room would naturally do with that signal - for better or worse.

 

[Thomas savage]

I don't know man. I disagree with several of the things Svart-hvitt has said in this thread. But that's fine! Actually, because of Svart-hvitt's postings (some of which were erroneous in my view), this thread turned quite enlightening, much thanks to the input from people like dr. Toole and Thomas Lund from Genelec. I, for one, didn't know that the professional standards for room curves are flat (or at least it didn't stick when I read dr Toole's book). That's an important thing to know, and now I know it.

But refocusing a discussion on to the persons involved ("assertive in your unsubstantiated opinions") is unnecessary and destructive for the discussion. It may be true or it may not be true in a particular case - but that really doesn't matter. Let's just stick to discussing the topics. There is absolutely no need for venturing beyond that.

Mod, Responding to a reported post

Tbh on this occasion and considering others posts here by @svart-hvitt and their assertive tone relative to his actual understanding of the topics covered here by Toole etc I’m inclined to agree with @Wombat .

I’d like @svart-hvitt to show a little more scope in his posting style for the possibility he’s misunderstanding things as has been paitently highlighted by our esteemed learned freind @Floyd Toole on a few occasions now.

As a general comment to everyone please be carful quoting the likes of Toole and olive in defence of your argument, I see both of their names being used on audio forums like sticks to bash people when there often is gross miss understanding of their positions at play.

Thanks .

 

[oivavoi]

Yes, I'm a (1). I reserve judgement on the bass compensation because everyone seems so adamant it is necessary but I have a suspicion that this might be partly to do with ported speakers that are near universal, plus my sealed, phase-corrected and time-aligned speakers have wider baffles than modern speakers and maybe all that makes them sound drier towards the low end.

Anecdotal information: To my great astonishment, listening to the 8Cs in a very small room as I do at the moment is actually highly satisfying even without any EQ or acoustic treatment in the bass. This is the first time I listen to a speaker with deep bass extension in a small room where the bass doesn't feel muddy or overpowering. So my subjective ears tend to agree with this. It could be that the more even roll-off and better transient response of the sealed bass, together with the fact that more of the bass is perceived as direct sound because of the cardioid dispersion, somehow fools the brain into perceiving the bass as more "tidy". Because logically, it doesn't make too much sense. When the bass waves become longer than the room, it seems logical to me that one actually perceives the room more than the speakers. And when I get around to EQ'ing or putting up one more sub, it might improve even further of course.

EDIT: As to @svart-hvitt 's comment on correcting based on "perceived direct sound": I'm not sure that this expression really makes sense? Not to me at least? As I see it, all room correction products are based on the idea of "perceived direct sound". The assumption is that reflections, or the ambience of the room, gets fused with our perception of the direct sound, and that the direct sound therefore needs to be adjusted so that the total perception of the sound becomes more "correct".

I struggle to see how Genelec's approach can differ from this. When the speakers measure flat anechoically, then any in-room change of that will be based on reflections and cancellations that are caused by the room. Now there are different approaches to room correct, with a multitude of measurement and correction methods, but it all has to with making adjustments to the direct sound in order to correct for what the room does.

 

[oivavoi]

Mod, Responding to a reported post

Tbh on this occasion and considering others posts here by @svart-hvitt and their assertive tone relative to his actual understanding of the topics covered here by Toole etc I’m inclined to agree with @Wombat .

I’d like @svart-hvitt to show a little more scope in his posting style for the possibility he’s misunderstanding things as has been paitently highlighted by our esteemed learned freind @Floyd Toole on a few occasions now.

As a general comment to everyone please be carful quoting the likes of Toole and olive in defence of your argument, I see both of their names being used on audio forums like sticks to bash people when there often is gross miss understanding of their positions at play.

Thanks .

I see your point Thomas. Ok, I hereby make a vow to stop commenting on @Wombat 's commenting style, and let you do all the moderating, @Thomas savage . Be the change you want to see in the world, etc. Peace out

 

[Thomas savage]

I see your point Thomas. Ok, I hereby make a vow to stop commenting on @Wombat 's commenting style, and let you do all the moderating, @Thomas savage . Be the change you want to see i the world, etc. Peace out

Your ok, keep doing what your doing. I value the diffrent perspectives just as long as they are given in private and don’t clog up threads

 

[Floyd Toole]

A question to Dr. Floyd : What is your take on omnidirectional speakers, like these :
http://german-physiks.com/
Would SS and DS curves be similar?

If they are free of resonances, they should sound just fine. I had a pair of bipolar Mirage M1s back in Canada. They were almost horizontally omnidirectional and had similar direct, sound power and steady-state room responses (see Figure 7.20 in my book). The Ohm Walsh 2, a precursor to the German Physics concept, is shown in Figure 18.4(f), and it shows horizontally omnidirectional performance over much of the frequency range (it had a conventional tweeter as I recall). The original Ohm loudspeaker tried to do what German Physics say they have done, but they were very difficult to manufacture and were unstable with time. The idea is good, success is in the execution.

All such multidirectional loudspeakers generate an expanded stereo soundstage, and those with properly tailored directivity can expand the listening area - see Section 15.3 in my book. The tradeoff is usually a slightly less clear imaging because of abundant room reflections. Some people like it, some don't. No recordings are monitored with such loudspeakers so they must be considered to be "sound effect" generators for listeners suffering from the directional and spatial deprivation of conventional stereo.

I prefer the multichannel option, using multiple surround loudspeaker in good stereo upmix algorithms. Then the spatial enhancement can be turned on or off, up or down as the mood and music demand.

 

[oivavoi]

they must be considered to be "sound effect" generators for listeners suffering from the directional and spatial deprivation of conventional stereo.

That would be me.

I prefer the multichannel option, using multiple surround loudspeaker in good stereo upmix algorithms. Then the spatial enhancement can be turned on or off, up or down as the mood and music demand.

Another option is what I plan on doing: have two setups in two different parts of the house: an omni setup for the living room, for classical music which benefits a lot from spatial "enhancing", optimized for a wide listening area, and a more directional setup for the ego-cave for maximal imaging clarity on artificial multitracked stereo recordings.

 

[Floyd Toole]

That would be me.



Another option is what I plan on doing: have two setups in two different parts of the house: an omni setup for the living room, for classical music which benefits a lot from spatial "enhancing", optimized for a wide listening area, and a more directional setup for the ego-cave for maximal imaging clarity on artificial multitracked stereo recordings.

If you look at Section 7.4.6, p.188 in my book, you will see that I chose a similar solution in my Canadian custom home. Now I have only one room, so the multichannel solution is the only practical and, thankfully, rewarding, option.

 

[oivavoi]

If you look at Section 7.4.6, p.188 in my book, you will see that I chose a similar solution in my Canadian custom home. Now I have only one room, so the multichannel solution is the only practical and, thankfully, rewarding, option.

Indeed. Reread the section on my Kindle now. Those are more or less exactly the thoughts I have had on how to recreate a classical soundstage, vs the more confined soundstage from directional loudspeakers in stereo. I suspect that I might have picked up those thoughts from your writings, forgot where I read them, and then assumed that I came to think of it all by myself...

 

[Guermantes]

Yes, I'm a (1). I reserve judgement on the bass compensation because everyone seems so adamant it is necessary but I have a suspicion that this might be partly to do with ported speakers that are near universal, plus my sealed, phase-corrected and time-aligned speakers have wider baffles than modern speakers and maybe all that makes them sound drier towards the low end.

I control the roll-off but nothing else, and they sound good to me - and of course they are therefore 'straight' and and the recording is not being messed about with arbitrarily. What the room does corresponds to what the room would naturally do with that signal - for better or worse.

And sometimes the room is overwhelmingly bad. I wonder how much trickle-down there is from live sound engineering practice where correcting for the "room" is considered mandatory. Or the long history of equalising sound replay in motion picture theatres. Judging by the table of contents, I suspect Dr. Toole covers this in his book.

 

[Guermantes]

If you look at Section 7.4.6, p.188 in my book, you will see that I chose a similar solution in my Canadian custom home. Now I have only one room, so the multichannel solution is the only practical and, thankfully, rewarding, option.

Is your current setup this one? It's refreshingly domestic
https://audiosciencereview.com/forum/index.php?threads/dr-floyd-toole-system-with-pictures.1564/

 

[Guermantes]

Another option is what I plan on doing: have two setups in two different parts of the house: an omni setup for the living room, for classical music which benefits a lot from spatial "enhancing", optimized for a wide listening area, and a more directional setup for the ego-cave for maximal imaging clarity on artificial multitracked stereo recordings.

This is interesting regarding the spatial experience of recorded classical music. I enjoy listening to multichannel classical recordings because of the immersive aspect of "being in the hall". There is that desire to recreate the acoustic space of the performance environment within the listening room. But there are two acoustic convolutions in this: the performance space and the listening space, the overlaying of which I would have thought would appear as one convolution to the human ear. In this case are we hearing through the room? Which room are we hearing through? Perhaps both so that we hear the original instruments?

Sorry, probably off-topic.

 

[Floyd Toole]

Is your current setup this one? It's refreshingly domestic
https://audiosciencereview.com/forum/index.php?threads/dr-floyd-toole-system-with-pictures.1564/

Yes, and it is defiantly "domestic" . I attach a more elaborate pic.

The stereo listening configuration is a 12 ft triangle. For multichannel movies I move back a bit. The elevation loudspeakers are now in the garage, to be installed soon, and the JBL Synthesis 24 channel processor is on its way. A big. boy toy!

 

[Floyd Toole]

This is interesting regarding the spatial experience of recorded classical music. I enjoy listening to multichannel classical recordings because of the immersive aspect of "being in the hall". There is that desire to recreate the acoustic space of the performance environment within the listening room. But there are two acoustic convolutions in this: the performance space and the listening space, the overlaying of which I would have thought would appear as one convolution to the human ear. In this case are we hearing through the room? Which room are we hearing through? Perhaps both so that we hear the original instruments?

Sorry, probably off-topic.

Evidence to date is that in such a situation of superimposed spaces, the larger one wins. In other words, we can make a small room sound much larger, but we cannot make a large room sound small. The fact that "room sound" is largely incoherent, non-minimum phase, reflections makes it easier.

 

[Floyd Toole]

And sometimes the room is overwhelmingly bad. I wonder how much trickle-down there is from live sound engineering practice where correcting for the "room" is considered mandatory. Or the long history of equalising sound replay in motion picture theatres. Judging by the table of contents, I suspect Dr. Toole covers this in his book.

There is a lot of folklore about "equalizing the room" in both the pro and consumer worlds. Some of it makes no sense. I discuss both at length in the book, and spend an entire chapter on cinema sound, explaining that it got off to a bad start many decades ago, and it is time to fix it - if anyone is willing.

 

[Wombat]

I don't know man. I disagree with several of the things Svart-hvitt has said in this thread. But that's fine! Actually, because of Svart-hvitt's postings (some of which were erroneous in my view), this thread turned quite enlightening, much thanks to the input from people like dr. Toole and Thomas Lund from Genelec. I, for one, didn't know that the professional standards for room curves are flat (or at least it didn't stick when I read dr Toole's book). That's an important thing to know, and now I know it.

But refocusing a discussion on to the persons involved ("assertive in your unsubstantiated opinions") is unnecessary and destructive for the discussion. It may be true or it may not be true in a particular case - but that really doesn't matter. Let's just stick to discussing the topics. There is absolutely no need for venturing beyond that.

If Dr. Toole had not corrected the many erroneous conjectures or interpretations of his publications which were posted in this thread, then those errors would still be festering here. Posters should take care in the veracity of what they say. This takes a level of knowledge and understanding of the topic to enable using referenced parts of sources in context regarding the whole of the source piece. Being able to assess the credibility of published material is another research skill.

If this is not pointed out when necessary then how can a poster be aware of shortcomings in this regard? There is a difference between constructive advice and personal criticism.

 

[Pio2001]

Evidence to date is that in such a situation of superimposed spaces, the larger one wins. In other words, we can make a small room sound much larger, but we cannot make a large room sound small. The fact that "room sound" is largely incoherent, non-minimum phase, reflections makes it easier.

Interesting.
When I listen to recordings where the recording acoustic is present, I've got the feeling that I am hearing both rooms. I can hear the room where the recording was made, but I don't feel like I am inside it. I can hear the musicians playing in that room, but I'm listening to this from my room, if you see what I mean.
Like if there was no wall behind my speakers, but rather a large door, with the musicians' room on the other side of the door.

Exemples on youtube where I've got this impression :
Small room (Omnia - Poëtree Gröne Lunden, two harps and two voices) :

Church (ODO ensemble - Nami Nami, qanoun, ney, percussion and voice. The voice is closer to the mics and is less affected by the acoustic) :

FWIW, my room is untreated (not even with a curtain or a carpet), and the reverberation time is 0.44 s at 500 Hz.

Sorry for the invasive links... I don't find a way to format them as simple links instead of huge Youtube windows

 

[Cosmik]

Interesting.
When I listen to recordings where the recording acoustic is present, I've got the feeling that I am hearing both rooms. I can hear the room where the recording was made, but I don't feel like I am inside it. I can hear the musicians playing in that room, but I'm listening to this from my room, if you see what I mean.
Like if there was no wall behind my speakers, but rather a large door, with the musicians' room on the other side of the door.

Exemples on youtube where I've got this impression :

I think you are illustrating something highly relevant with a well-known phenomenon.

If you make a recording of your child playing recorder at a school concert while you are sitting in the audience, the results are usually highly disappointing. When you were in the school hall, it was as if you heard your child directly, and the hall acoustic registered separately. In effect you 'heard through the hall' to coin a phrase. But in contrast, the recording sounds 'hollow', coloured, indistinct. It does not sound the same.

The recording blends the hall with the direct sound, and listening to it you have no way of distinguishing between the two. The reason is likely to be that when you were in the hall, you used the wealth of information coming from two ears, micro head movements, directional information generated by ear shape and head related transfer function (HRTF), etc. to fully characterise the space and the acoustic sources within it, and your brain separated the sources from the hall. The hall still registered, but just as separate 'ambience' - hopefully pleasing.

Now, if you were a measurements-minded person, you might have done some tests in that hall, measuring the frequency response errors caused by the acoustics for a person sitting in the audience. It would look pretty bad: undulations all over the place. In order to improve the audience experience there would be no alternative: the performers would have to be 'modified' through applications of cotton wool, rubber bands, acoustic apertures, resonators, and so on, in order to pre-correct their frequency responses to counteract the hall's terrible effects. Alternatively, the concert could be turned into an electro-acoustic performance using pickups on the recorders, feeding the outputs through DSP and amplified speakers. The DSP might be called 'Room correction'.

The problem is, the audience members wouldn't hear a 'correct' version of the recorder player; they would hear a recorder player strapped with bits of cotton wool behind acoustic screens, etc. or some DSP-processed recorders.

The same is true of the 'room correction' advocates in audio. They believe unquestioningly what their microphones and laptops tell them about the speakers and room combined, and in doing so they simply ignore the well-known phenomenon you illustrate above. The result is, for example, a speaker with its treble boosted to the point where the instructions tell you not to listen to it for too long! The listener hears the 'correction' directly, not the blend of 'correction' and room acoustics. The 'uncorrected' neutral speaker would have sounded normal.

(This is also the reason that recordings are generally not made with a pair of mics placed where an audience member might sit. It was realised long ago that a recording sounds better, generally, if microphones are much closer to the performers, and the ambience level is controlled. Hence, many recordings use a multiplicity of microphones and are carefully mixed.)