In my own personal testing with headphones and my DAW I really felt that diffusion under 400hz is unnecessary and creates too much mud which masks the critical dynamics of that frequency range.
i dont understand this test; how are you testing "diffusion under 400hz" with headphones?
Quote:
Originally Posted by OpusOfTrolls
noticed a drawing of an RFZ control room that had a LF diffusive rear wall. I said to myself how could this be any better than a dead wall? The diffuser doesn't absorb (AFAIK), but it does minimize the interference of the first reflection. Knowing physics there will be another reflection soon after, and I doubt it would be very predictable. But could it be beneficial?
are you insisting the total specular response of the RFZ model is casual?
Massive diffusers would break up LF resonance, but that is not an argument for their adoption.
I have read the Master Handbook of Acoustics by Alton Everest and on one particular section, noticed a drawing of an RFZ control room that had a LF diffusive rear wall. I said to myself how could this be any better than a dead wall? The diffuser doesn't absorb (AFAIK), but it does minimize the interference of the first reflection. Knowing physics there will be another reflection soon after, and I doubt it would be very predictable. But could it be beneficial?
Think about the wavelength of a LF soundwave, and how long it takes to actually realize the dynamic. Could adding temporal diffusion to a slow to begin with frequency band make the sound any better? It might, but I will say the effect is narrow in respect to the source audio. It will modify the dynamics and create masking in the frequency band. This is obviously not beneficial to critical listening, but I have heard many records with LF reverberation that without wouldn't have had the same conveyance.
So I do believe it is a matter of taste. In my own personal testing with headphones and my DAW I really felt that diffusion under 400hz is unnecessary and creates too much mud which masks the critical dynamics of that frequency range.
I agree with you, but note that I was not making an argument for massive diffusers as an alternative for absorption in a control room.
We were talking about his live room, and simply exploring the idea of large diffusers. While they are not a replacement for trapping, I think the concept is worth exploring in the live room because they would break up LF resonance while keeping energy in the room. Anyway, I was not trying to recommended this as a solution--I do agree that it's probably more effective, more predictable and safer (in terms of time and $) to focus on diffusion over 400 Hz.
I agree with you, but note that I was not making an argument for massive diffusers as an alternative for absorption in a control room.
We were talking about his live room, and simply exploring the idea of large diffusers. While they are not a replacement for trapping, I think the concept is worth exploring in the live room because they would break up LF resonance while keeping energy in the room. Anyway, I was not trying to recommended this as a solution--I do agree that it's probably more effective, more predictable and safer (in terms of time and $) to focus on diffusion over 400 Hz.
Sorry for my confusion, sometimes I don't backtrack to find context. LF Diffusion in a recording scenario would be a good thing IMO.
i dont understand this test; how are you testing "diffusion under 400hz" with headphones?
are you insisting the total specular response of the RFZ model is casual?
I HPF a reverberation plugin in parallel with the dry signal, and listen to it. Then I ask myself, could this reverb sound excellent on any source material?
---
Isn't all of it casual? What type of 'casual' are you referring to?
I HPF a reverberation plugin in parallel with the dry signal, and listen to it. Then I ask myself, could this reverb sound excellent on any source material?
FWIW a Schroeder reverb is nowhere near realistic, nor are many feedback designs. Maybe you are not considering the other varieties?
“Schroeder reverb”? Please enlighten me.*
Don´t get me wrong; I’m not disagreeing with your general conclusion that diffusion below about 400 Hz or so (room size dependent naturally) is not that useful. It´s the way you came to that conclusion that bothers me.
Do you realize that even talking about reverb below 400 Hz assumes an enormous space?
* EDIT: ok, you´re referring to the artificial reverberation suggested by Schroeder. Either way, to use a reverb plugin thinking that you simulate the effects of diffusion in a small acoustic space is just not right.
Don´t get me wrong; I’m not disagreeing with your general conclusion that diffusion below about 400 Hz or so (room size dependent naturally) is not that useful. It´s the way you came to that conclusion that bothers me.
Do you realize that even talking about reverb below 400 Hz assumes an enormous space?
..which can be simulated with DSP listening through headphones. But you need to use the right algorithm.
It would be very difficult for a DSP engineer to make one of these to simulate a real space. They are more likely to simulate the axial conditions with some phase randomization thrown in. It also does not account for any binaural effects unless told to do so, which may or may not be a realistic approximation.
There is only a handful of reverbs that come close to 'real'. I don't think I can name any that actually fulfill every detail. But you have a solid argument Jens, because even if the recording held every nuance of a space's diffusion and echo, there would still be the HRTF absent, which simulated cannot even come close to the real thing.
In my defense, I only use the reverb to approximate masking and dynamic. It was only the other day I realized the signifigance of any binaural effect.
I HPF a reverberation plugin in parallel with the dry signal, and listen to it. Then I ask myself, could this reverb sound excellent on any source material?
are you seriously making an attempt to consider headphones with a form of decay as an apples vs apples comparison with respect to diffusive surfaces within a
bounded acoustical space? oh, dear...
"I really felt that diffusion under 400hz is unnecessary and creates too much mud which masks the critical dynamics of that frequency range."
is this the type of logic breakdown where people think that by playing "reverb" (decay FX) from their speakers that magically the energy emitted from their source suddenly has statistical energy flows instead of specular wavelets?
Quote:
Originally Posted by OpusOfTrolls
Isn't all of it casual? What type of 'casual' are you referring to?
are you insisting the total specular response of the RFZ model is casual?
ca·su·al/?kaZHo?o?l/
Adjective:
Relaxed and unconcerned.
Noun:
A person who does something irregularly: "a number of casuals became regular customers". Synonyms: accidental - incidental - fortuitous - occasional
the diffusers were not just randomly placed with respect to the total specular response of RFZ. it is not "accidental" or just haphazard... there is actually some thought behind it...
playing a recording of a reverberant sound-field (or reverb FX) doesn't suddenly make the energy flows in your small acoustical space statistical.
Did you even read what I wrote? I am using HEADPHONES. That pretty much nulls out the ability of my small acoustical space to interfere with the signal.
Did you even read what I wrote? I am using HEADPHONES. That pretty much nulls out the ability of my small acoustical space to interfere with the signal.
yes, i read what you wrote.
you're equating headphones + an FX and attempting to compare apples to apples with "diffusion":
you said:
"I really felt that diffusion under 400hz is unnecessary and creates too much mud which masks the critical dynamics of that frequency range."
The simulation of the result of diffusion is the end effect of the acoustical function. While the model of the simulation is limited by the sound source, the result is within range of an approximation of a few key attributes that make the real system what it is.
My use of the word diffusion was incorrect, but I don't run circles in my head all day when I slap on some reverb and call the sound 'Simulated diffused acoustical interference wavelet manifold transfer function response'.
If you don't mind me asking, where in your room are you planning to mount the diffusers? Or did you just build them for experimentation?
Thanks again,
Tim
Three modules in the control room partial instead of the 'flutterfree'QRDs and some modules as 'mobilediffuser' for recording against the wall or on the floor.
I've build dozens of all kind diffusers esp. PRD(skylines) which I like very much and there's not much space left, guess I'm addicted to diffusion. and I like to do experiments.
One thing that was remarkable IMO with your stepped diffuser was the good transient or attack response(though quite normal with diffusion) despite the small dimensions of my DIY, example in the piano samples.
Three modules in the control room partial instead of the 'flutterfree'QRDs and some modules as 'mobilediffuser' for recording against the wall or on the floor.
I've build dozens of all kind diffusers esp. PRD(skylines) which I like very much and there's not much space left, guess I'm addicted to diffusion. and I like to do experiments.
One thing that was remarkable IMO with your stepped diffuser was the good transient or attack response(though quite normal with diffusion) despite the small dimensions of my DIY, example in the piano samples.
Fantastic, Schaap! Yes, while these diffusers are not deep, they are efficient at spatial scattering (as you can somewhat visualize from the video below):
If you get a chance, I'd love to see some photos of the diffusers in position in your room, or photos of any of the test configurations. E.g., photos of the "profiled modulation" (when the modules are oriented at different depths) would help people visualize what it all looks.
would the b2-frac array of five elements, at 2,1m width, perform like the n-36 or better if mounted in the suggested "0 50 60 50 0" order at different depths? if that mounting will lift the performance above the n-36 we don´t mind a little more work since it is still only straight cuts and mills.
today we had the carpenter at our studio to discuss the work and the materials, but since it is my duty to provide the detailed plans for the build and we have some other things to do first, it is no problem to switch to another design! so i probably will propose the build of a n-36 diffusor for the back wall instead.
one more question: how is the performance of the n-36 above 4000hz and is this of any significant relevance in a smallish room like ours? and what if we would put fractals on just some of the wells of the n-36 diffusor? could that be of any benefit to spread the performance to higher frequencies?
peace, rez
I don't know exactly how well the array will perform if mounted like that, because I've not tested it. I'm confident it will perform much better than it does now, but the n=36 stepped diffuser might still perform better.
Here's why I'm confident that profiled modulations (mounting modules at different heights) will boost the performance: Using the A1-LF design, I tested a variety of profiled modulations, including a 7-module low frequency fractal modulation, which I'll post an image of as soon as I upload it to my server.
I've not tested the n=36 stepped diffuser above 4000 Hz. Usually, the highest frequency of interest for predicting room acoustics acoustics is about 5500 Hz. This is just a hypothesis, but I expect the fractal diffuser (with a good profiled modulation applied to it) to outperform the n=36 stepped diffuser at higher frequencies (e.g., 10000 Hz).
Sure, you could put fractal-like cells on the n=36 stepped diffuser. I'm not sure if it would be worth the time and effort, but it would look neat and it would improve the high frequency diffusion (at the very least, it will create more HF scattering due to the increase in geometric complexity).
Hope this answers your questions! Does everything I've said make sense?
I've attached performance reports (pdf files) for the n = 36 optimized stepped diffuser.
The well width is 59 mm, and the sequence of well depths are given in the performance reports.
Unfortunately, I don't have the Reflex file because I was using a demo version of Reflex (it does not allow saving).
if i build a n36 stepped diffusor like in the provided pdf, should there be fins between the wells?
that might be a stupid question, but i´ve never built a diffusor and there are so many different designs out there.
i saw designs where between each well there was a fin of the depth of the whole diffusor and there were designs that had just the stepped wells.
Thanks for sharing all this.
Been reading this great thread and would like some opinions please.
I plan on making a rear wall diffusor for my little studio.
I was going to make a proper QRD from wood.
Then I saw it was going to cost me too much time and money.
Then, using QRdude, I designed a N17, with no fins, 25mm well widths, 20 cm deep. Seemed cool (except for being finless and for being made out of 16kg/m3 eps as opposed to a higher density eps).
Now I am looking at Tims simple blueprint and wondering:
Would that simple design be a better way forward for me? It would cost a similar amount to the EPS N17.
Or another way to look at it:
Would I be better off making Tim's simple 7 step design with 4 periods (I need to cover up to 1.8m of wall width)
OR
Would I be better of making a long 'high N', low density EPS stepped diffusor with narrow wells?
BTW, I was planning on coating the EPS with some tough water based polyurethane paint to create a very hard surface. A denser EPS increases the cost exponentially down here in South Africa.
Then, using QRdude, I designed a N17, with no fins,
QRDude can only predict (roughly) the performance of real QRDs (with fins i o w). Removing the fins will result in unpredictable drastically reduced performance. In order to make a finless design, you need to simulate the shape using BEM/FEM/FDTD (Reflex from AFMG being on example) that can accurately predict the actual performance.
thx for the help. i think i'm just gonna mount a wood wall for its aesthetic properties and then treat the room accordingly with trapping and possibly diffusers after the fact. simpler idea...better results.
QRDude can only predict (roughly) the performance of real QRDs (with fins i o w). Removing the fins will result in unpredictable drastically reduced performance. In order to make a finless design, you need to simulate the shape using BEM/FEM/FDTD (Reflex from AFMG being on example) that can accurately predict the actual performance.
New Reply 
Facebook
Twitter
LinkedIn
. and I like to do experiments
.
