Originally Posted by Northward
Comes the music playback. Original sound ---- (x)ms ISD Gap ---- Heavy diffused feed back (-12dB!) from which the brain can in fact still extract info and endlessly tries to see if it indeed is a reminiscence of the original sound - so wants to find out constantly if this is a "new sound" or a reflection.
Think about it this way: how does the brain suppresses and fuses the reflections within the Haas effect but still allows new event to be heard within this gap? It's not like it shuts down sound altogether for 30ms. There is a "tool" in there that lets the brain know if this is a new sound or an "old" sound, a reflection. Without that tool, there would be no Haas effect possible as the brain would simply not know what to fuse. This is a simplistic explanation, but you get the point. I have always been fascinated by the brain's capacities btw. You think that what you hear is reality, but it isn't. It's a complete alternate reality, reconstructed by your brain.
The lack of ER from self-noises or musical playback only enhances that gymnastic as because of what we see, the brain is expecting ER it doesn't get neither from the "room playback" nor from self-noises. Note: this is an automated function, even if you are fully conscious of the deadness of the front wall you're facing, your hearing system process cannot be altered and thus still expects ER. Making you really tired after sometimes just a couple of hours. And making you take really strange Delay/Reverb decisions (a "fun" by-product of these issues)
I wrote the following text as a draft before Thomas's two posts above and then got sidetracked by work. When I later visited this thread to post the text I saw Thomas had very elegantly made some similar points, along with some great info about other factors. Thanks for the excellent posts, Thomas! My thoughts became redundant, but I thought it might still be useful to post what I wrote anyway:
Originally Posted by bwo
But work on the low frequencies first. Later you can move on to treating early specular reflections and then late ones.
This is the main point. In small rooms, diffusion is less important than the primary goal of reducing low frequency resonance. BIR nulls are also a bit less of a problem than resonance, given your head is not in a vice and we have two ears. Even though 125Hz is quite a long wavelength with an 8ms cycle, a fixed mono mic doesnt hear what your two ears hear as you move your head. Your head will rarely occupy the exact position of your mic. You'll certainly hear the dip at 125Hz, but it wont sound as bad as the chart seems
to indicate if it is BIR.
BIR nulls and modal nulls look bad on a static plot and they sound bad when listening to a pure sine, but they arent as bad as low frequency ringing when our ears are listening to music in a room. Our brains are pretty good at dealing with BIR and unmanaged specular reflections above Schroeder, but we arent good at dealing with masking caused by low frequency room resonances. This is also one of the reasons why we typically focus first on correcting resonant peaks rather than trying the much more difficult task of correcting deep nulls....coincidentally (or not) that's also why room DSP is pretty good at reducing resonances but less capable of improving nulls.
Humans love audio reflections. It's why we sing in the shower, add ping pong delays to reggae, enjoy choirs in cathedrals, love the drum sound in When The Levee Breaks etc etc.
This has a basis in evolution - like most animals with two ears, we evolved from organisms which possessed excellent location faculties (often based on deriving cues from reflections) to identify prey or foes. We rarely needed to decode those cues within a pile of low frequency resonances, although I guess that might apply when attacking a herd of stampeding bison, but we used our eyes for that
I'll say it again: a mono microphone does not hear what a mobile human head hears. We dont listen to music with our head in a vice. A mic doesnt have HRTF. A sine sweep is not music (despite all audio being made from sine waves). BIR is relative....and what we hear isnt necessarily the same as the way a chart displays the data. The displayed data is very useful and very informative, but BIR is not a reason for despair or aspirin unless you obsess over graphs and cant live with the graphs not looking
perfect. In your case, I'd be more focused on frequencies below 80Hz. Fix those and your room will sound better. More importantly, your mixes will translate well.
What you want is a room you can work in and be reasonably confident that what you are hearing is sufficiently accurate for you to make valid decisions. That doesnt mean you need a room response which looks perfect on a graph. Dont get caught up chasing BIR perfection unless you are building a mastering room (which you would probably not attempt to do with your room). Mastering rooms need to be as close to perfect as possible because that's the whole point of mastering - it's the final QA check before manufacturing. Home-based project studios typically dont need that level of precision.
You said this earlier:
Originally Posted by SirTralala
use the little intelligence for science i have to find some inexpensive, easy solutions to improve the sound of my room a bit (including the use of acoustic material I already built).
As you've discovered, the problem with that statement is two words: inexpensive and easy. Dealing properly with small rooms cheaply is not easy. But you can achieve enough by easy cheap methods to ensure you can work productively.
Here's one thing I'd do if I was determined to stick with a DIY solution: I'd return back to the beginning, take everything out of the room except for one speaker and the mic. Put one of the absorbers over the window. Put a square cube absorber in each of the 8 tri corners. Measure. Move mic up, down, sideways. (12 inch and/or 24 inch movements). Measure after each move. Do the same with the speaker. Add cloud. Measure again. The low frequency resonances should be reduced at this point. Add side wall specular absorbers. And so on. By the time you've done that, you should know what is causing the big null and how to reduce the problem. And the low frequency resonances will be controlled.
SirT, can I suggest your most efficient approach (if you really want to nail all of the problems and are unable to buy some Modex Plates) is to engage someone to assist you professionally? It's not a service I provide (I'm not an acoustician), but I can highly recommend Jeff Hedback
. He's not at all expensive and he is experienced with the kinds of problems you are encountering (and he knows a lot about space couplers).
PS dont worry about the mild bickering amongst the pros here - acoustics is a science, but the methods of fixing problems are as varied as the opinions on the relative merits. Whatever gets you to the point where you can work comfortably is all that matters....and there are multiple ways to achieve that. As Thomas says, you can count the people who really know this stuff on your fingers and toes. I'm certainly no expert, but Thomas most definitely is.