I probably said this before but I think power response is like RT. It's not a good descriptor for what we're interested in.
The early reflection pattern is the dominant factor when it comes to spatial properties. It's probably the dominant factor for loudness and timbre perception as well.
An example (?): The movie industry has a reference level which is based on a steady state measurement. When we apply the same measuring methodology to acoustically small spaces it turns out that the perceived loudness in a movie theater vs. a home theater is quite different. The perceived loudness difference is at least 10dB.
Don't go trying to put me in that camp
I fully agree that the direct sound has to be a long way down below the reflections for us to "let go" of it as the source of sound for localisation, however if we are already sitting near the critical distance where direct/reflected are comparable in level, with perhaps only a scant few dB advantage for the direct sound (something we found in our informal survey in "the big thread") then it only takes 10-15dB attenuation of the direct signal for this to be achieved.
Can a person wearing a woolly jumper standing a metre in front of the tweeter on a speaker attenuate the direct sound at high (>2Khz) frequencies by at least 10-15dB ? I haven't measured it but I believe so. (Obviously the measurement to prove this would need to be gated so its not including reflections...)
My point was this could be enough attenuation for the direct sound to lose its "grip" on our perception, so that we start to perceive more of the off axis sound than we usually would - thus it isn't a fair test to determine how important the off axis sound is in the usual case where the direct sound is not artificially attenuated.
Again I agree, however I have to wonder whether the threshold for reflections affecting perception of tonal balance is dfferent than it is for localisation.
Eg, the direct sound might have to be at least 10 dB below the reflections (depending on the time delay of the reflections..) before we start to localize the reflections instead of the real source - but perhaps the direct sound only has to be roughly equal or lower in level than the reflections before they start to contribute to perception of tonal balance.
My own unverified, unproven, anecdotal results (to appease those who say its not backed up by a peer reviewed research paper...) is that I start to notice a shift in tonal balance from the direct sounds's tonal balance to the overall room sound (room power response) at roughly the critical distance where the two are equal in level.
If the direct sound is several dB stronger than the reflected sound I perceive almost entirely the direct sounds tonal balance at high frequencies with little or no input from the room, provided that the RT of the room isn't excessively long such that I can hear sound coming back from the room in "gaps" in the music...
At the critical distance I notice a small but audible shift in tonal balance towards the rooms response, and when the direct signal is a few dB below the reflections I think the reflected signals do start to affect the perceived balance considerably, even before localisation shifts.
So power response might or might not matter depending on the specific listening environment and distance.
On the other hand, to get an "optimal" experience with sound reproduction, I think it really is necessary to be sitting at less than the critical distance so that the direct signal is at least a few dB stronger than any reflections or the reverberant field in general. In this case it would seem that power response is far less important, so power response may be most important in "non optimal" listening environments where the listener is beyond the critical distance.
Last edited:
Do you have a reference for this? I remember a Dolby discussion about dialog level in theaters large and small, but don't know if that is the same.
David
I wish I would but there's nothing besides anectdotal evidence. I guess questioning movie reference level publicly would be a major marketing disaster.
Virtually everybody I've talked to (including a re-recording engineer working at a high profile studio) doesn't listen at reference level at home because it's simply too loud.
I use to listen -20dB below reference level.
I'm not so sure it doesn't matter... I can't seem to escape it. If my listening axis tells me it wants a dB or two at some point but the room contribution doesn't for some reason agree, then it will just sound wrong. And I listen in the same half of the room as the speakers and I have over 100lbs of fibres about the place. When designing I find myself hedging as if the direct/room are, and should be equal, in the middle ranges at least.
Hi Markus
No photos of my ceiling device - sorry. Some measurements a long time ago to show that it did work - not as well as hoped, but perhaps better than expected considering its shallow extension. (I keep trying to find those old measurements, but alas I can't. I will need to do all new ones, but that won;t give a before and after).
Your point here is exactly the way that I see it as well. You have to control the reflection pattern (with the speakers and the room) or the sound just isn't right. If the direct field were dominate then this would not matter, but it does. In my room there is a dramatic change in the perception when one move forward into a direct field dominated situation or backward into a more nuetral direct/reverberant ratio. Again, if the direct field were dominate this should not happen, but it does. The direct field dominate has a headphone like perception where one losses the "space" of the room. It gives very precise localization oif the source images, but it is not natural, at least not to me. Moving back into the reverberant field and the sound is much more "spacious" - of course - but also much more natural. Its the VER pattern that does this and to me, its critical.
I should point out that imaging is not a strongpoint of Toole and Olives work. Floyd is not big on image - he is a classic freak and these recordings or their real life performances do not have strong images. There is no direct rating of "image" in Olives work either. He is big on coloration, spaciuosness, etc. So when they say that the direct field is dominate imaging is not a critical criteria as it is with me. What Folyd and Olive have done is very important and a very good first step, but it isn't the whole story. Ask them about the importance of "dynamics" - isn't even a factor. They completely discount any form of nonlinear distortion (which I agree with in part, but, as our previous discussion highlighted, I think that there are nonlinear perception effects which are not being considered. Toole and Olive don't consider them at all.)
Still no iPhone?
Here's a picture of my absorber hat - took me 5 minutes from taking the picture to sending this post:
An externally hosted image should be here but it was not working when we last tested it.
Although our speakers and damping is very similar, we get different spatial results. I don't get the larger sweet spot from 45° toe-in as you do. I also found the strong contralateral reflection to be rather detrimental.
So this is either due to cognitive differences or a distinct difference in the reflection pattern of our rooms. I'd like to dive deeper into that subject but you probably still don't have the time for doing measurements in your room?
My measurements say yes. Damping is >15dB. The absorber hangs forward by about 30cm.
iPhone?! No, I am not an Apple fan. I do have a Windows phone with a camera, and maybe I will do that, but I am just about to fly off to Boston and can't take the time to figure it out. (I have a very good Nikon system and I just can't see using my phone as a camera so I have never actually used it - well a couple of on-the-spot shots - they were terrible. So if I do take a picture it will be with my Nikon.)
Good idea on the ceiling bounce problem by the way. I could certainly do that since my speakers are behind curtains nobody would see it. Ceiling bounce is a big problem to correct.
Agreed, I would like to do this as well.
My time may free up as I catch up on orders (which I am doing slowly but surely). But I am also finding that as the economy gets better I get a lot more consulting inquiries. If I do get some time I will try and get the room measurements and we can pursue that topic. I have some subjective testing that I want to do with Lidia as well on thermal modulation. I have put that off for far too long as well. Too much work, too little time.
OT, but maybe this can help out others.
I consult all over the world. I need acces to almost everything that I do from everywhere I go. The "cloud" does this for me. Using "mesh" all my code, doc, and my entire Outlook database is synched between the cloud, my desktop and my laptop. No matter where I am on any computer that I use my data and docs are available (and backed up I might add). It's all flawless without any input required from me once I have it setup. The only glitch is if I have Visual STudio running on one computer and want to edit the code on another computer, Visual Studio will save the edits as two seperate filenames - not a problem, just an inconvenince renaming the files to be a single name.
I can actually log into my desktop remotely and run program from it. This is not very fast, but it does work if the need arises.
I consult all over the world. I need acces to almost everything that I do from everywhere I go. The "cloud" does this for me. Using "mesh" all my code, doc, and my entire Outlook database is synched between the cloud, my desktop and my laptop. No matter where I am on any computer that I use my data and docs are available (and backed up I might add). It's all flawless without any input required from me once I have it setup. The only glitch is if I have Visual STudio running on one computer and want to edit the code on another computer, Visual Studio will save the edits as two seperate filenames - not a problem, just an inconvenince renaming the files to be a single name.
I can actually log into my desktop remotely and run program from it. This is not very fast, but it does work if the need arises.
I need something a little more substantial in my case. That's the ceiling at 80 and 240. It is somewhat omnipresent though and partial absorption is liable to increase the depth of some of the nulls.
Attachments
The hat would also be difficult because the woofer is 70cm below the top panel. I also need to keep absorbers free of the room lighting and casting shadows, or move the lights.
These nulls change frequency in proportion with the measuring distance.
With several inches of absorber and some sub into the mix, it is doable. (Better than shown, I've overwritten the more recent shots in Holm)
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.