LEDE / RFZ ideas applied to a critical listening room

[AwwDeOhh]

Originally Posted by jim1961

..
Auralex 2" studiofoam (not used anywhere except where I made a lampshade out of it)

LoL
At least you found some use for the stuff!

[jim1961]

Originally Posted by AwwDeOhh

LoL
At least you found some use for the stuff!

Yes, I was taken in initially by the stuff. In my case, using it was better than bare walls so I started putting it everywhere until one day I realized something was wrong. I probably dont have to point out what that was on this forum

[jim1961]

Originally Posted by bwo

I would start focusing on attenuation early reflections more now. You should be able to get them all below -25 dB. Redirecting on opposite sidewalls as already mentioned will attenuate more then absorption.

Ive gotten to this.



Marginally better than previously. Still, that -15db @ 2.5ms floor bounce bothers me but am at a loss as to what else to try.

Ive tried multiple carpets, no carpet, different arrangements of carpet. The only thing that makes it better is putting some scrap OC703 on the floor and that gets it down to -20db.

[aackthpt]

Box containing 703 with a highly perforated top (strong enough to stand on) with carpet on top, permanently in place on the floor. If it's in a traffic path rearrange a bit. Or maybe design a sculptural looking deflector, not that I'm sure how I'd hide a 2 foot wide panel. Perhaps as cat furniture, LOL.

[localhost127]

Originally Posted by aackthpt

Or maybe design a sculptural looking deflector,

+1: it only has to be temporary for measurements/subjective testing until a final decision is made.

[Syncamorea]

Originally Posted by jim1961

I probably dont have to point out what that was on this forum

Thanks for the giggle. So you've sworn off foamaholica?

[AwwDeOhh]

Originally Posted by jim1961

Yes, I was taken in initially by the stuff. In my case, using it was better than bare walls so I started putting it everywhere until one day I realized something was wrong. I probably dont have to point out what that was on this forum

been there.. done that, lol
It was the terrible results of that first foam purchase that got me interested in the subject to begin with... so i guess it wasn't a total loss!

[jim1961]

I have a new question regarding termination strength.



I put this ^^^^ here for reference to my question. Typically, termination strength is quoted to be -10 to -12db or so. So my thinking is 3-5db above the Haas detection threshold if your ISD is up 20ms. But what if your terminating later than 20ms? For instance at 25ms, the Haas threshold is about -19db instead of -15db for 10-20ms. Does this lessen the needed or desirable termination strength? For a 25ms ISD, would the termination strength now be -16 to -14db (keeping 3-5db above threshold) ?

Another thought is that at 25ms, a -10db ISD termination would be putting you very close to the "image broadening" threshold. We dont want that do we?

[jim1961]

This is where I am at

I realize its not that smooth, but the additional higher order reflections after the termination are better there in this fashion than not at all?



This compares where I am at to what I last posted here (Black = new, Orange = old)

(ignore the first 1 ms, the plot jumps around a bit)




My question is, is this ^^^^ roughly what ideally the ETC should look like for the LEDE / RFZ model?

That is, -18db in the 25-30ms range. -24db in the 30-35ms range, etcetera?

[bwo]

Here are my thoughts, but don't take them for granted. I'm no expert.

There's no absolutely answer to what the ideal length of ISD gap is for LEDE. 3-5 ms longer then the recorded signal is what's optimal, but this would of course vary from one record room to another. In a listening room, to make it quite long (25-30 ms) would probably be preferable to cover most of recordings.

Other then that, it's important how the late reflections from the rear arrive. They shoud be properly diffused which will make an exponential decaying diffuse tail as the RPG pic shows. In your case, you're missing some of this. You have a strong termination, but the field afterwards are not gradually falling all the way. Probably because you don't have enough diffusors close together. So you end up with some reflections as well and that's why you see some sticking out again in the area of 35-37 ms. A better diffuse field would also have everything closer together as oppose to spread out sparse reflections. The strong termination helps to mask this, but it's not optimal.

It's difficult to say for sure which is better. The orange seems to show less late arrival high gain reflections which is good. I'm assuming then that black one is showing specular reflections after 25 ms and not diffuse energy. Either way, diffuse energy should ideally be closer together.

There are indications by the way that an optimal termination should be in the area between -10/-15 dB to the direct signal. More importantly for you though, would be to improve the diffuse tail if that's possible. Quality diffusors placed together with the barker code. This will yield a more uniform backwave with less coloration.

And by the way; Get yourself a pair of CBT36 speakers if you can. Greatest speaker design breakthrough in years. Don Keele is a geniuos. Overlooked of course by typical audiophile community.

[aackthpt]

Originally Posted by jim1961

My question is, is this ^^^^ roughly what ideally the ETC should look like for the LEDE / RFZ model?

Yes.

I agree with most of what bwo said, with additional comments:

You appear to have a lot of very-early clutter (pre-2ms) that the LEDE/RFZ folks seems to think needs to be cleaned as much as possible. It does look like your previous one lacked that clutter.

It also looks like the old one possibly had a few less highish-gain reflections post-termination. But IMO to really assess this you need to show a longer time scale to see the decay to the noise floor.

As for the CBT speakers, they look fantastic and I'd love to hear a pair, but I think they come more from the Toole mindset that early reflections are good if they match the spectrum of the speaker so I don't think they match well with the LEDE/RFZ idea that OP is following. They might work well with it, but maybe not. "When Acoustical Philosophies Collide" might make for good television though.

bwo I presume you meant -10/-15 dB (not ms) to the direct.

[aackthpt]

Also wanted to note that LEDE/RFZ proponents claim that the strong termination does some pre-and post-masking to remove cues from other features of the ETC... you may very well find that removing clutter or smoothing the decay doesn't do a lot audibly if they are correct.

How does the sound compare between these two setups? Particularly on music with a lot of reverberation, or recorded in extremely large spaces?

[bwo]

Originally Posted by aackthpt

As for the CBT speakers, they look fantastic and I'd love to hear a pair, but I think they come more from the Toole mindset that early reflections are good if they match the spectrum of the speaker so I don't think they match well with the LEDE/RFZ idea that OP is following. They might work well with it, but maybe not. "When Acoustical Philosophies Collide" might make for good television though.

The important factor of CBT speakers is a much more uniform response. So you have a speaker with less coloration from the room as well as a more even freqeuncy response. How you treat the room is solely up to the owner. There's absolotely nothing that should make this speaker "match" worse with LEDE. On the contrary, a speaker with a more uniform polar response would only be benefical to a design that strives for accuracy. With CBT speakers it actually becomes easier to treat the room. With no vertical treatment, the early vertical reflections will be attenuated below -20 ms or more.

Originally Posted by aackthpt

bwo I presume you meant -10/-15 dB (not ms) to the direct.

Yes. Corrected it to dB.

[jim1961]

Originally Posted by aackthpt

But IMO to really assess this you need to show a longer time scale to see the decay to the noise floor.

[jim1961]

Originally Posted by bwo

The important factor of CBT speakers is a much more uniform response. So you have a speaker with less coloration from the room as well as a more even freqeuncy response. How you treat the room is solely up to the owner. There's absolotely nothing that should make this speaker "match" worse with LEDE. On the contrary, a speaker with a more uniform polar response would only be benefical to a design that strives for accuracy. With CBT speakers it actually becomes easier to treat the room. With no vertical treatment, the early vertical reflections will be attenuated below -20 ms or more.

Jenzen-NEXT

As for this subject, i do have some new speakers on the way ^^^^

Right now, I am using a 2-way design employing:

Vifa - P17WJ-00-08 - http://www.tymphany.com/files/resour...17WJ-00-08.pdf
Dynaudio D21af - http://www.gattiweb.com/images/dynaudio/d21af_1.pdf

Along with a subwoofer (2X 10" woofers in 5.9 cu ft ported box)

[akebrake]

Originally Posted by jim1961;8384692

[attach

314840[/attach]

My question is, is this ^^^^ roughly what ideally the ETC should look like for the LEDE / RFZ model?
That is, -18db in the 25-30ms range. -24db in the 30-35ms range, etcetera?

Regarding old ETC measurements.
LEDE / RFZ ideas applied to a critical listening room

TEF 12 pic from PD'A/Cox. LEDE/ RFZ
Note: This graph says nothing about the spectrum (level) of reflections. This is of coarse common with broadband ETC graphs.

I remember Chips Davis wanted to ”standardize” the 5 kHz ETC-sweep on the TEF analyzer so designers could compare their measurements. (80 ms span) A nice idea…
But in this this pic the time span i 40 ms which means 10 kHz swept. (Small room) This ETC "favors" 5 kHz frequency band because of the filtering.
So if you want to compare your room with ”historical rooms” by means of ETC you have to measure in the same way.
Or simulate the TEF-ETC (often published) by filtering the modern broadband ETC with a fairly broad (a couple of octaves) BP filter peaking at, in this case, 5 kHz (depending on number of frequcies swept).

Note also that after the termination you have a fairly strong, dense, reflected, diffuse-field with a decay tail. (which BWO points out) That room is not "dead" I guess. The reflections are ”moved in time” and probably re- introduced to enhence the diffuse decay by carefully placed reflectors/diffusors.

Best

[jim1961]

Originally Posted by akebrake

Regarding old ETC measurements.
LEDE / RFZ ideas applied to a critical listening room

TEF 12 pic from PD'A/Cox. LEDE/ RFZ
Note: This graph says nothing about the spectrum (level) of reflections. This is of coarse common with broadband ETC graphs.

I remember Chips Davis wanted to ”standardize” the 5 kHz ETC-sweep on the TEF analyzer so designers could compare their measurements. (80 ms span) A nice idea…
But in this this pic the time span i 40 ms which means 10 kHz swept. (Small room) This ETC "favors" 5 kHz frequency band because of the filtering.
So if you want to compare your room with ”historical rooms” by means of ETC you have to measure in the same way.
Or simulate the TEF-ETC (often published) by filtering the modern broadband ETC with a fairly broad (a couple of octaves) BP filter peaking at, in this case, 5 kHz (depending on number of frequcies swept).

Note also that after the termination you have a fairly strong, dense, reflected, diffuse-field with a decay tail. (which BWO points out) That room is not "dead" I guess. The reflections are ”moved in time” and probably re- introduced to enhence the diffuse decay by carefully placed reflectors/diffusors.

Best

I am not sure I am understanding what you mean here, but the bandwidth I am using on these ETC graphs is 450hz - 5.6K

[jim1961]

Originally Posted by aackthpt

Also wanted to note that LEDE/RFZ proponents claim that the strong termination does some pre-and post-masking to remove cues from other features of the ETC... you may very well find that removing clutter or smoothing the decay doesn't do a lot audibly if they are correct.

How does the sound compare between these two setups? Particularly on music with a lot of reverberation, or recorded in extremely large spaces?

I suppose what I am wrestling with at this point is how much (amplitude) and how long (out to what time in ms) I want to add room response after the termination.

Going back to my earlier graph (the orange one), there was a sense of liveliness, but not spaciousness if that makes sense. The "tail" as its being called seems to be the attributing factor to the spaciousness detail or aspect of this so ive been adding more of it.

One of the reasons I went with a strong termination was exactly to mask inconsistencies. In listening, I do not hear them.

As far as comparing the black and orange rooms, the black one does indeed seem more spacious. But in truth, I am no expert listener. Particularly in the area of making assessments on the particulars of the reverberant field following the termination. Simply speaking, if you go back to before I had a termination at all, things had gotten kinda dead. The terminator alone helped some, and this added "tailing" has helped a bit more. I now would not describe my room as dead sounding. But I am still not sure on many particulars including how I have gone about what I am doing here. Unfortunately, the ETC leaves out much, including the spectral data.



A one octave band study like this ^^^^ helps some.

Black = 500hz
Orange = 1k
Green = 2.1k
Purple = 4.2k

(ignore the blue line at the top)

------------------------

So at the end of the day, I dont know what a LEDE / RFZ room done perfectly to its fullest is supposed to sound like. Ditto for other acoustic room models.

[akebrake]

Originally Posted by jim1961

I am not sure I am understanding what you mean here, but the bandwidth I am using on these ETC graphs is 450hz - 5.6K

All analyzers / software programs are different and the early ETC's published (by Everest, RPG et al) are often without information about frequency span, filtering etc.
In the TEF analyzer you were usually choosing the ETC resolution by changing the number of Hz swept (span) in the ETC sweep. Not filtering after wards.

My point was that it's difficult to compare measurements of the "ideal ETC" from text books with an "ideal ETC" from another measurement. We don't know what it should look like.

Don't know if that was of any help..l

BTW:Same goes for inserting ETC curves into diagrams by Toole/ Haas etc as "proof" of psyco acoustic phenomena. Different things. (This is opinion)

[jim1961]

Originally Posted by akebrake

All analyzers / software programs are different and the early ETC's published (by Everest, RPG et al) are often without information about frequency span, filtering etc.
In the TEF analyzer you were usually choosing the ETC resolution by changing the number of Hz swept (span) in the ETC sweep. Not filtering after wards.

My point was that it's difficult to compare measurements of the "ideal ETC" from text books with an "ideal ETC" from another measurement. We don't know what it should look like.

Don't know if that was of any help..l

BTW:Same goes for inserting ETC curves into diagrams by Toole/ Haas etc as "proof" of psyco acoustic phenomena. Different things. (This is opinion)

Thanks for that added clarity.

What adds to the difficulty for someone like me or anyone for that matter trying to emulate the LEDE / RFZ model to their home listening environment or studio is the lack of comparable data. There are very few (any at all?) ETC's out there with this emulation in mind. Much less the test criteria and software settings as you point out. So for LEDE / RFZ, the goals for ETC are vague at best. While the ISD gap length and termination strength are fairly concisely defined, the following diffuse "tail" attributes are practically non existent. And what does exist, your now telling me cant be compared to my data.

[aackthpt]

If you want to see a real ETC of LEDE/RFZ that looks nearly ideal, and _bonus_ how it was done, how about nordenstam's?

LEDE room with Haas trigger

[akebrake]

Originally Posted by aackthpt

If you want to see a real ETC of LEDE/RFZ that looks nearly ideal, and _bonus_ how it was done, how about nordenstam's?

LEDE room with Haas trigger - Gearslutz.com

Thanks for that link! Very informative! (Missed that thread before.)
The human search engine is invaluable!

Looks like Andreas N. was using REW for his ETC’s.. We have to compare/ evaluate ETC’s from different software. We have no set up parameters for the measurement. So in IMHO the earlier discussion is still valid

Andreas measurement looks very nice. His ”Haas kicker” is sticking up alone above the main decay. That differs from the RPG pic and the reflected field is 6-8 dB stronger on the later one.

The ”Haas kicker” was an additional rear reflector to increase the termination of the ISD-gap in the early LEDE designs. Some times burried in the diffused reflections and not nescessarily stronger (on the ETC) or placed in the leading edge. Different ideas.

We believe in measurements, don’t we

Cheers

[Syncamorea]

Originally Posted by jim1961

But in truth, I am no expert listener. Particularly in the area of making assessments on the particulars of the reverberant field following the termination.

What reference recordings are you using to evaluate the changes made to the room acoustics?

[jim1961]

Originally Posted by aackthpt

If you want to see a real ETC of LEDE/RFZ that looks nearly ideal, and _bonus_ how it was done, how about nordenstam's?

LEDE room with Haas trigger - Gearslutz.com

Here is what Nordenstam's and My ETC look like when put into the same scale. They look a lot more similar put into this framework yes? Still, certain features stick out.

1) His tail is much smoother than mine and he has dealt with the early clutter better than I. Most of mine is a floor bounce which I have already discussed earlier in this thread and still dont know how to deal with properly.
2) His termination is at 19ms, mine is at 24ms.
3) While his "tail" is smoother, his immediate reflections following his terminator are very low. Looking at the models ive seen, you want the "tail" to gradually decline off the termination peak, not drop like a rock, yes?
4) His termination strength is a bit weak I think. -15.5db it looks like. Mine is -8.5db.


I have seen that thread before, but thank you in bringing up here because it does fit nicely with what is being talked about.

I too am using reflectors on the rear wall to get my Haas Kicker (the termination peak). Without a room diagram, I cant really cant tell how he is doing it though.

[aackthpt]

Originally Posted by akebrake

Looks like Andreas N. was using REW for his ETC’s.. We have to compare/ evaluate ETC’s from different software. We have no set up parameters for the measurement. So in IMHO the earlier discussion is still valid

You are using ARTA, right? I may ultimately do such a comparison myself, but you could just download REW and run a sweep with it right after an ARTA sweep. It's not difficult to learn -- gotta be simpler than ARTA. I'd be amazed if you get a different result though.

Didn't local post the ETC plots from the old LEDE workshops somewhere? For the life of me I can't find it (in a reasonable time) but it would be interesting to see if the even decay began at the termination (like the RFZ image) or if the termination sticks up out of the decay field, like nordenstam's. Maybe also this article would be interesting. I think it's highly probably the answer is "it doesn't really matter", though. I'd guess that except for someone who wants to be extremely pedantic, LEDE goals are less like rules and more like guidelines.

[Suomela]

Originally Posted by jim1961

//
1) His tail is much smoother than mine and he has dealt with the early clutter better than I. Most of mine is a floor bounce which I have already discussed earlier in this thread and still dont know how to deal with properly.
//

Have you tried placing a larger coffee table of suitable height in front of the listening position to try to suppress the effects of floor reflections?


/Suomela

[akebrake]

Originally Posted by aackthpt

You are using ARTA, right?

No, not yet. I'm a Mac guy and got REW and Fuzz M. (and an old TEF 12)

..... it would be interesting to see if the even decay began at the termination (like the RFZ image) or if the termination sticks up out of the decay field, like nordenstam's. Maybe also this article would be interesting.

Thanks for Don Davis article.
Bin a Syn Aud Con member for a long time but have not seen that one before.

The level of the "room sound" will probably make a difference.

Sorry for this short reply. Have to run away for a couple of hours.
See you

Best

[jim1961]

Originally Posted by Syncamorea

What reference recordings are you using to evaluate the changes made to the room acoustics?

Basically, whatever ive been listening to a lot lately so my memory imprint is strong as to how it sounded previously.

[jim1961]

Originally Posted by aackthpt

Maybe also this article would be interesting.

I took the ETC of the Master Sound Astoria sound room and added some bars to better see whats going on. This provides hope for me because clearly this graph has some waviness to it (not perfectly sloped). I am not sure how much should be read into this, but the article seems to say that the termination should be in the -5 to -10db range. That is a higher gain than the commonly used -12db figure that I see most.

I also notice these factors regarding the graph:

In the 30-50ms range, there are peaks in the -15db range. Much higher than most graphs ive seen and much higher than mine in that time range. Even after 60ms, there are peaks in the -22db range. Again, higher than I would expect. Its not till you get out past 110ms that they fall below -24db.

So either one of two things seems to be going on here.

1) These relatively high gain late reflections are masked by the termination

2) These late reflections are heard and not considered bad and are there by design

[aackthpt]

Originally Posted by akebrake

No, not yet. I'm a Mac guy and got REW and Fuzz M. (and an old TEF 12)

REW runs on Mac, you could easily do the comparison. Either that or bribe someone with beer. Unfortunately that someone can't be me as I've no Apple hardware.