But now you're back to "is flat the reference" which has been gone over and over already in the thread 😀
I wouldn't be surprised if there were a very high "correlation" between early (pretty correlated) reflections and later (pretty uncorrelated) sound. In other words, that the phenomena attributed to the various windows of time delay were really attributable to the brain being abetted or hindered* in creating spatial sound percepts based on the corpus of reflections. (It is also possible that there are flaws in many studies of delay, and there are many, because they did not properly attend to whether their test signals were correlated or uncorrelated).
About the most puzzling simple issue for me has been how the tricky lab tests resulting in Fletcher-Munson-like curves relate to music - and it is even more complicated than I thought, if Earl's earlier comment is to be trusted. In turn, there's no determinative way to establish loudness (that is, the subjective experience of loudness which can be objecitvely measured, of course) before that is nailed down. Which earlier led me to suggest that physical measurement be based on music-like test sounds.
That is naive. It is also wrong in that it only addresses room reverb and at best might quiet some fo the debate about speaker design concepts. What might be helpful is music-like test signals (say, paired-comparisons, as we say in the trade) to help subjective evaluations (of course, objectively measured, as always).
*Being "hindered" is the desirable condition because it means you take these uncorrelated echos and conclude the sound is coming from a large hall, not from the little speaker boxes.
About the most puzzling simple issue for me has been how the tricky lab tests resulting in Fletcher-Munson-like curves relate to music - and it is even more complicated than I thought, if Earl's earlier comment is to be trusted. In turn, there's no determinative way to establish loudness (that is, the subjective experience of loudness which can be objecitvely measured, of course) before that is nailed down. Which earlier led me to suggest that physical measurement be based on music-like test sounds.
That is naive. It is also wrong in that it only addresses room reverb and at best might quiet some fo the debate about speaker design concepts. What might be helpful is music-like test signals (say, paired-comparisons, as we say in the trade) to help subjective evaluations (of course, objectively measured, as always).
*Being "hindered" is the desirable condition because it means you take these uncorrelated echos and conclude the sound is coming from a large hall, not from the little speaker boxes.
Last edited:
As long as there is no reference there can't be an answer to "is flat correct". Flat can be the reference just like the X-curve can be reference. We just need to define a meaningful reference. Even after 80 years of stereophony there is no such thing in the audio industry.
In the light of all facts I can't see that Earl is doing anything fundamentally wrong.
There're de facto standards that justify a high frequency rolloff. Audyssey does the same.
I don't think there's anything fundamentally wrong with Earl (or anyone) applying a small roll off in the top end either, depending on the directivity characteristics of the drivers.
The debate is only what the justification for it is, and whether there is one, other than "it sounds right". 🙂
and we know - thanks to guidance from our distinguished meaningful reference - that "it sounds right" is no justification 😉
Last edited:
The justification is that it's closer to (a mostly unknown) reference. The reference is the mixing/mastering engineer in his control room. For movies there's some consistency among the acoustics of theaters and dubbing stages which allows the definition of a reference curve.
The real reason for this discussion is audio's circle of confusion. With better standards we could make more progress.
Last edited:
This is essentially why the long thread. However, most new monitors are flat, but power/polar response is still all over the map and industry pros have all sorts of opinion of which monitor is more accurate--AND IT CAN EVEN DEPEND ON WHAT TYPE OF MUSIC THEY ARE RECORDING.😕 They don't like it when you point out how silly that is. Then throw in the differences of control room acoustics, which are all over the map and home listening environments, again all over the map, and it almost seems a miracle any of this is listenable. Perhaps not.
Here's to adaptation and not knowing the reference!
Really though, thankfully we are getting closer to having a reference and the movie industry is leading the way.
Dan
One of the reasons for the commonly prefered falling slope in
home reproduction - especially when a CD radiation pattern
is used - may have its origin in the average deviation of a
control room's acoustics from a home environment.
Maybe have a look at "Studio C"
BLACKBIRD STUDIO
Surely this may not be a typical control room ...
But what happens, when a recording is mastered e.g. in
"Studio C" and played back in one of our "average"
living rooms ?
In many cases the recording will tend to sound overbright
at home, even when using the very same speaker for
mastering and home listening.
Maybe "Studio C" is quite harmless, because there might
be broadband diffusion to be expected ...
But other control rooms may have significantly higher
absorption from mid to high frequencies compared to an
average living room.
Is a falling slope in a home speaker an adequate solution ?
Is it even close to the underlying problem ?
Even if we compensate for excess HF energy in the living
room by "downsloping", what do we create in that room ?
"Proper imageing" as intended by a hopefully skilled
sound engineer (working e.g. in "Studio C") ?
That puts some of our hairsplitting discussions in here
under question, doesn't it ?
Last edited:
it does! and therefore I keep on saying: there is no hope for accuracy as an ideal in that regard because it is absolutely impossible to recreate at home exactly what the sound engineer has heard in His control room
therefore realism should be the goal, realism defined subjectively as a sound source recognition when we can't tell the reproduced from the real and objectively... ...as far as we can
The movie and television industries do, at least, have standards. They aren't always followed, of course.
For film projection it wasn't too hard, you just had to be sure there was enough light on the screen, the lamp didn't flicker and the aspect ratio was correct. Real tight standards meant keeping the image stable and ambient light off the screen. Video is more complex with color balance, brightness, contrast, chroma. I used to make a living doing video display calibration in TV stations, editing suites and production houses. They were mostly very far out of standard.
But audio? Flat or any curve on axis isn't enough. There is so much more to consider - mostly having to do with the room and the power response. No wonder debates like this go on forever. Funny tho, after all the debate and back and forth, what do we come down to? Most folks find a slight downward sloping response on axis works best. Some others find a flat response works best (seems to be those who sit close). But not much else!
It really is not all over the place, is it? Nobody here is advocating a big midrange dip, or boosting 100Hz by 6dB, or even a super tight +/- 0.2dB flat response. Not at all. There have been a lot of hairs split. Seems we disagree more in theory than in practice.
For film projection it wasn't too hard, you just had to be sure there was enough light on the screen, the lamp didn't flicker and the aspect ratio was correct. Real tight standards meant keeping the image stable and ambient light off the screen. Video is more complex with color balance, brightness, contrast, chroma. I used to make a living doing video display calibration in TV stations, editing suites and production houses. They were mostly very far out of standard.
But audio? Flat or any curve on axis isn't enough. There is so much more to consider - mostly having to do with the room and the power response. No wonder debates like this go on forever. Funny tho, after all the debate and back and forth, what do we come down to? Most folks find a slight downward sloping response on axis works best. Some others find a flat response works best (seems to be those who sit close). But not much else!
It really is not all over the place, is it? Nobody here is advocating a big midrange dip, or boosting 100Hz by 6dB, or even a super tight +/- 0.2dB flat response. Not at all. There have been a lot of hairs split. Seems we disagree more in theory than in practice.
With proper standards this is achievable.
It will be even harder to reach realism without proper standards. For example, binaural recordings don't work very well for me because they don't follow the necessary standards, being my personal HRTFs.
Hi Dave
1) the fact that nonlinearities are a crucial factor in echo cancellers in no way indicates that the same thing is true for our hearing. The two have nothing to do with one another. The nonlinearities in the echo become unmasked as the linear portion of the echo is cancelled making them obvious. In the ear this masking is always present and acts to make the distortion products inaudible.
2) a 90 degeree phase shift does not decorrelate a signal and Greisinger does not say that either. He does say that they phase shift the LF by 90 degress and "it sounds better" but nowhere does he imply that this is because the two channels are decorrelated. But both your comment and his require further information to go any further. When you say 90 degrees phase shift, is this at a particular frequency, or at all frequencies at the same time. The two things are quite different and neither of you is clear on that point. The former simply results in a constant delay independent of frequency and perfect correlation at that delay time (note from Markus post that correlation is a time function not a scalar, so rotating a pure tone by 90 degees makes the correlation go to zero at time zero, but it is unity at the delay time). A constant 90 degree phase shift at all frequencies is quite a bit harder to analyze, but I would still tend to believe that the correlation is unity at some time. At any rate the constant delay, while being quite easy to do does not decorrelate anything. The 90 degree phase shift at all frequencies would be much harder to impliment as this would require a frequency dependent phase delay. I believe that a Hilbert transform does this.
The design of a decorrelation filter is a complex task as a google search will indicate.
In room acoustics decorrelation comes about only after a series of progressive reflections. It is interesting to note that in a model this decorrelation only comes about as a result of a fairly random selection of reflections, while a periodic or regular set of reflections will not decorrelate the reverberation field. The random nature of the reflections is REQUIRED to get a decorrelated reverb field which speaks directly to the design of an auditorium for good spaciousness. This also goes directly to what is meant by decorrelation. Without some randomization in the structure of a filter it will not decorrelate the input form the output. Decorrelation means "no linear relationship of one signal to the other". No "regular" manipulation of a signal is going to do that, only a non-regular, i.e. random, manipulation can achieve this.
but is there hope for proper standards established and followed in the foreseeable future?
there is not, at least in a free country, but why not? because in a free country the guiding principle for sound enginneers in their professional activity is also sounds good to me / I like it that way, it suffices to take a look at pro forums like gearslutz to realize the fact
therefore it is unachievable
there is a misunderstanding in Your argument - standards are critical for reproduction point of view, hence from accuracy perspective
I am talking about production here - let's try to produce something more realistically sounding with recordings that we are given by the industry instead of hopelessly trying to get something that particular sound engineer had in His mind and in His control room
Last edited:
There's always hope. Even in a free country. And there're de facto standards. Audyssey is currently making it into control rooms.
Well, I'm open for new ideas but I'm sceptic that any of the concepts you brought up so far do reliably produce what you're talking about or if those speakers just create pleasant random spatial effects. I'm in contact with Schupbach for an audition of his stereolith at my place. This is my latest DIY project:
An externally hosted image should be here but it was not working when we last tested it.
yes Markus - and You have expressed this scepticism many times
and that is a serious insult, don't You agree? 😉
quite impressive!
but really I am impressed, taking into account the sceptical person who has made such a substantial investment 😉but You are aware that the sound You will hear will be the sound of two FRS8 in a cardboard? 😉
an that just back-to-back might be not enough?
great that You will have also an opportunity to check out the original product as we suspect that there could be some clever filters inside 🙂
Last edited:
Pardon?
quite impressive!
but You are aware that the sound You will hear will be the sound of two FRS8 in a cardboard? 😉
an that just back-to-back might be not enough?
I'll throw in a 3rd FRS8 and a DCX and two subs.
What makes you think that there is a "clever" filter inside?
- Status
- Not open for further replies.