ETC - ISD gap question - AVS Forum

AVS Forum > Audio > Audio Theory, Setup, and Chat > ETC - ISD gap question

Audio Theory, Setup, and Chat

jim19611961's Avatar jim19611961
07:33 AM Liked: 88
post #1 of 57
07-24-2012 | Posts: 1,421
Joined: Mar 2011


My isd gap is around 14ms. Is this too short for you folks?

If I were to somehow remove the 14ms peak, is the peak around 22ms of sufficient amplitude for a isd termination?
jim19611961's Avatar jim19611961
10:01 AM Liked: 88
post #2 of 57
07-24-2012 | Posts: 1,421
Joined: Mar 2011

dragonfyr's Avatar dragonfyr
10:40 AM Liked: 13
post #3 of 57
07-24-2012 | Posts: 809
Joined: May 2005
You are going to want to address the very early diffraction/reflection issues occurring between ~0-2ms, as the indirect arrival at ~2ms is your first significant reflection.

Also, as I doubt your room is 25 feet long, you might also re-window your display to roughly 2x your room length in feet (e.g., if it is 15 feet deep, window the time axis to about 30ms for increased visible resolution.

yes...an ISD of ~13ms is fine..

But you would also want to use treatment (preferably 2D diffusion) redirecting energy to address the later anomalous peak at ~22.5ms as you want the decay to be as dense and smooth (following an exponential curve) as you can after the ISD is terminated, as a late sparse reflections can cause localization (etc.) issues..
jim19611961's Avatar jim19611961
01:53 PM Liked: 88
post #4 of 57
07-24-2012 | Posts: 1,421
Joined: Mar 2011
Actually, my room is 15.5W + 25.75 feet long. So it is 25+ feet long actually.

The 2ms peak is from the floor. Already have 3 rugs on top of each other and -15db is as good as its gotten (obviously). I dont know what else to do with it.

Fixing the 22.5 ms peak is easy. Give me a bit and I will post another ETC.
dragonfyr's Avatar dragonfyr
02:12 PM Liked: 13
post #5 of 57
07-24-2012 | Posts: 809
Joined: May 2005
There is a bit more to "fixing" it. And I hope the solution includes more than simply adding more damping (absorption) where the presence of such energy is actually desirable.

In general after the first significant reflection (FSR), you want t 'diffuse' soundfield to be prominently laterally oriented and well behaved - meaning that it is both temporally and spatially dense - lacking any anomalous high gain sparse reflections that cause problems with imaging and localization - and following an exponential rate of decay (with a log scale, the rate of decay of a line tangent to the maximal gain points after the FSR will be a straight sloping line.)

Again, we need to know more about the room. If this is a fully treated room, there are no guarantees that what is up is valid unless all is effectively broadband.
What are the baseline measurements of both speaker and room? The proper process provides a systematic accounting/inventory of the myriad variables ranging from the speaker response itself, to the room and subsequently to the effects of each treatment and accoutremount in the room.

Exactly what steps have been taken, and what exactly are the characteristics of any applied treatment? Is there verification of this or is the performance simply assumed?

Also, the floor does not not account for all of the very early clutter in the early reflected energy. A floor reflection would be discrete - with the only issue being that of multiple driver signal alignment in the near field.

And the ETC is performed with ONE speaker driven at a time. And there appears to be issues with symmetry as well seeing as how the various peaks are not coincident from what appears to be at least two sources driven. (especially with regards to the FSR/ISD termination).

For much of that, reducing the time window will allow for the expansion of the detail in the display and allow greater focus on any region of arrival time.

There is much more to a systematic process. (Unfortunately this sounds all too like an wacked amirm challenge where the process is intentionally perverted.)

So at this point, we can give simplistic isolated advice based upon MUCH unreliable assumptions, but that is all without substantial gaps being filled in. And the conclusions will only be as good as the validity of the assumptions.

If you are interested in pursuing this in more depth, PM me w8th a YM or Skype contact, and we can talk as time permits.
jim19611961's Avatar jim19611961
04:07 PM Liked: 88
post #6 of 57
07-24-2012 | Posts: 1,421
Joined: Mar 2011

The Room

Imagine the brown panels as QRD, and Black as Absorption.

The Absorption Panels are:
beside each speaker are 12" thick, R19
Rear corners, 8.5" thick R19
Ceiling cloud (not shown) 2" OC703 with 6" R19 above that. Cover entire area between listening position and speakers.
Front left corner - 4" OC703 24" across corner with outward facing FRK
Front right corner 2" OC703 14" across corner (all the room their is with the door in the way) with 4-6" R19 behind also with outward facing FRK.

The QRD panels are:
single N7 throughout (12"Wx48"H) except flanking the listener where those are N17's (28"Wx24"H).


My "fixing" was changing the angles on some flat panels on side walls that redirect energy to the rear (not shown) to the left and right of listener.



I drive one channel at a time for ALL my tests.

As you can see, I did fix the 22.5ms peak fairly well.

Here is a better detailed ETC showing both channels only to 30ms



I have lots of info. Just tell me what is helpful and I will post it.
FOH's Avatar FOH
08:48 AM Liked: 213
post #7 of 57
07-25-2012 | Posts: 4,749
Joined: Apr 2010
Welcome Jim, I'll watch this one. It's a pleasure seeing ETC being utilized around these parts. Any PM material covered privately, that could benefit the community, it would be nice if that was shared here as well.

Best of luck.
jim19611961's Avatar jim19611961
10:04 AM Liked: 88
post #8 of 57
07-25-2012 | Posts: 1,421
Joined: Mar 2011
Hello there,

Chime in at any time smile.gif

I did have a lengthy discussion with Dragonfyr. Had it been notated and put in this thread, it would have been over 100 pages. But I can generalize the keys points (imo)

1) The first 2-3ms are very important. I have a -15db peak from the floor there. Still cant figure out how to deal with it other than putting an absorption panel on the floor that I would likely trip over. I know its a floor reflection cause when I do put a scrap piece of OC703 there, the peak goes to -21db.
2) Its better to terminate the ISD gap from a lateral direction. Currently my ISD-Gap termination of 13-14ms comes primarily from behind. I am looking at ways to change this. Is there a certain lateral degree that is preferred?
3) While a long ISD-Gap (18-25ms) is better in that it gives the perception of a larger space, as low as 10-12ms can work.
4) You ideally want a slow decaying dense sound field to follow the ISD termination.

There were many other thoughts covered, but these are the ones I recall at this moment. My questions here are for anyone. .
FOH's Avatar FOH
10:13 AM Liked: 213
post #9 of 57
07-25-2012 | Posts: 4,749
Joined: Apr 2010
Anyone, Dragon included, should be applauded for selflessly giving their time to others for such discussion. Dragon has solicited the same to me, and although I've not taken him up on it I do very much appreciate it. His ETC knowledge is superior. I just wish as much as possible can be played out for all, here forum style. I know the minutia could fill pages, as you say, ... but all the better IMO.

Best of luck
jim19611961's Avatar jim19611961
11:07 AM Liked: 88
post #10 of 57
07-25-2012 | Posts: 1,421
Joined: Mar 2011
Anyone, Dragon included, should be applauded for selflessly giving their time to others for such discussion.

+1
AJinFLA's Avatar AJinFLA
02:09 PM Liked: 24
post #11 of 57
07-25-2012 | Posts: 1,579
Joined: Feb 2006
Quote:
Originally Posted by jim19611961 View Post

I have lots of info. Just tell me what is helpful and I will post it.
Ok.
How does the loudspeaker system/room sound? To your ears/brain. Before and after?
What is you goal perceptually? What specular speculation IS-Dis the gap measurements will tell you when you get there? How will you know? Thanks.

cheers,

AJ
jim19611961's Avatar jim19611961
05:10 PM Liked: 88
post #12 of 57
07-25-2012 | Posts: 1,421
Joined: Mar 2011
How does the loudspeaker system/room sound? To your ears/brain.

I think things are sounding very nice. But the spacial field is not quite right.

Imaging, great.

Freq response, very good. have null in the 220hz-280hz to deal with, but is minor.

Before and after?

Before and after what exactly?

What is you goal perceptually?

To keep the great imaging I have and enlarge the perceived size of the room.

What specular speculation IS-Dis the gap measurements will tell you when you get there? How will you know?

I dont understand this question.
jim19611961's Avatar jim19611961
05:58 PM Liked: 88
post #13 of 57
07-25-2012 | Posts: 1,421
Joined: Mar 2011


Hey Mark, here is the preliminary ETC plot after changing things according to what we talked about.

This is with a OC703 panel on the floor for the prior 2ms peak. I doubt I can keep it this way, but wanted to see what it looked like.
AJinFLA's Avatar AJinFLA
06:10 PM Liked: 24
post #14 of 57
07-25-2012 | Posts: 1,579
Joined: Feb 2006
Quote:
Originally Posted by jim19611961 View Post

I think things are sounding very nice. But the spacial field is not quite right.
Imaging, great.
I tend to conflate "imaging" with spatial perception of the soundfield. So I'm not quite sure what is/was "wrong" with your spatial reproduction. "Imaging great" means what? Ultra-precision that enhances the artificiality of the studio pan pot type stereo construct, to satiate even a Studiophile.....or a more natural/realistic type not-so-ultra-precision that one would tend to hear with live music (assuming one has heard live music!)? What is not quite right?
Your rendering appears to show a 2 ch system with 3 chairs nearfield to speakers well away from walls. Quite an unusual setup. What speakers are those? Are they the only acoustic sources? No subs? How many listeners?
Quote:
Originally Posted by jim19611961 View Post

Freq response, very good. have null in the 220hz-280hz to deal with, but is minor.
Be careful that isn't a measurement artifact like "floor bounce" that is largely imperceptible. There is a huge difference between a microphone and your binaural adaptive hearing system. Remember, our hearing system has evolved over time (unlike Studiophile knuckle dragging types) to adapt to non-open spaces (and stuff like "the ground"), so certain ugly looking measures may not be so ugly audibly. Inability to discern this can lead to sonically tragic consequences.
Quote:
Originally Posted by jim19611961 View Post

Before and after what exactly?
Those speaker wall bandages in the room rendering. I assume they weren't there when you purchased the place, unless an escaped Studiophile lived there?
What did it sound like prior to those bandages being applied?
Quote:
Originally Posted by jim19611961 View Post

To keep the great imaging I have and enlarge the perceived size of the room.
I see. So you intend to increase your spaciousness by killing it?
That is certainly a novel approach! How is it working out so far with your sighted listening vs what you remembered weeks (months?) ago hearing?
Quote:
Originally Posted by jim19611961 View Post

I dont understand this question.
Hah. On very rare occasions I like to say stuff a bit tongue-in-cheek.
Basically, when will the measurements you see, tell you what you want to hear?

cheers,

AJ
jim19611961's Avatar jim19611961
07:00 PM Liked: 88
post #15 of 57
07-25-2012 | Posts: 1,421
Joined: Mar 2011
Quote:
Originally Posted by AJinFLA View Post

. So you intend to increase your spaciousness by killing it?
That is certainly a novel approach! How is it working out so far with your sighted listening vs what you remembered weeks (months?) ago hearing?

Basically, when will the measurements you see, tell you what you want to hear?
cheers,
AJ

Killing it? Hmmm. I dont know if your trying to be constructive here or not. Ill be glad to answer your query if your willing to be specific.

As far as how the room sounded before ANY treatment was ever applied and now, its really night and day. The images were blurred and smeared before. The soundstage, not very deep. Some harsh flutter echo and a uneven freq response +/- 8db compared to now +/-3.5 (from 300hz-12k)(1/24th octave smoothing)

The measurements merely tell me if the model I am using is being achieved.

I have only given things a quick listen since my latest ETC post, but what I noticed from only yesterday is the speaker locality seems to have disappeared much more. In other words, the speakers as a source have become much more invisible. The added diffusion has made things seems smoother and more fluid sounding. A bit more detail in some of the images.

As far as my definition of what I mean by "great imaging", we really dont have a common frame a reference, you and I, for me to be able to answer that. But I will say that generally I dont like the sound or images to sound too clinical. I like things to sound naturally. I prefer analog sound to most digital. I like warm tonalities, not bright. But mostly, in regards to imaging, I like to be able to tell where everything is. And hear "space" or "air" around each image. I like to be able to tell where the wall boundaries of the recording are.

Hopefully, this helps

Jim
AJinFLA's Avatar AJinFLA
08:22 PM Liked: 24
post #16 of 57
07-25-2012 | Posts: 1,579
Joined: Feb 2006
Quote:
Originally Posted by jim19611961 View Post

The Absorption Panels are:
beside each speaker are 12" thick, R19

Killing it? Hmmm.

Ya. Like, stone dead, kaput, bye-bye. Unless the opposite of reality happened and spaciousness increased by putting 1 foot thick absorbers to annihilate lateral reflections from your nearfield triangle setup?
Quote:
Originally Posted by jim19611961 View Post

As far as how the room sounded before ANY treatment was ever applied and now, its really night and day. The images were blurred and smeared before. The soundstage, not very deep.
The soundstage of what acoustic sources? What furnishings in the room other than the 3 chairs?
Quote:
Originally Posted by jim19611961 View Post

Some harsh flutter echo and a uneven freq response +/- 8db compared to now +/-3.5 (from 300hz-12k)(1/24th octave smoothing)
Do you have those FRs?
Quote:
Originally Posted by jim19611961 View Post

The measurements merely tell me if the model I am using is being achieved.
What model of home stereo reproduction are you using?
Quote:
Originally Posted by jim19611961 View Post

I have only given things a quick listen since my latest ETC post, but what I noticed from only yesterday is the speaker locality seems to have disappeared much more. In other words, the speakers as a source have become much more invisible. The added diffusion has made things seems smoother and more fluid sounding. A bit more detail in some of the images.
That's quite impressive, since source localization is usually something I associate either with the recording or source artifacts like diffraction and/or resonances, etc
Quote:
Originally Posted by jim19611961 View Post

As far as my definition of what I mean by "great imaging", we really dont have a common frame a reference, you and I, for me to be able to answer that.
Maybe, but we should have some commonality if you attend live, unamplified acoustic music events. How often do you do this, vs listen to analog say records on you stereo? Which forms the basis of what imaging "should sound like" to you? An LP or a live event?
Quote:
Originally Posted by jim19611961 View Post

But I will say that generally I dont like the sound or images to sound too clinical.
Not a medical guy so I have no idea what that means.
Quote:
Originally Posted by jim19611961 View Post

I like things to sound naturally. I prefer analog sound to most digital. I like warm tonalities, not bright. But mostly, in regards to imaging, I like to be able to tell where everything is. And hear "space" or "air" around each image. I like to be able to tell where the wall boundaries of the recording are.
Hopefully, this helps
Jim
If you're happy with the sound that's all that matters.

cheers,

AJ
Bigus's Avatar Bigus
10:11 PM Liked: 51
post #17 of 57
07-25-2012 | Posts: 4,258
Joined: Feb 2003
Jim, can't tell on my phone if you are new here or not. AJ is nearly uniformly against any form of intentional room treatments, mocking those who use them as "padded cell" types. Might help you know how to read his posts, especially when you can't understand what he is asking but have that feeling there is something lurking behind the question.

I will leave it to you to decide if he is constructive or not. At the least, questions from someone like AJ can be useful in getting you to reflect on your goals for the room, whether you are chasing good sound or good measurements, and getting you to think about how you know the sound will be at its best by reviewing all the data gathered.

As to your specific questions... you may not need to worry so much about the floor reflection. I can't see your pics at the moment so can't make many specific comments. You can apply diffusion on the back wall so that it isn't the source of the isd termination (1D). Getting it specifically redirected to the sides to provide lateral termination is a bit trickier. I'll see if I have any useful suggestions once I can view your attachments.
jim19611961's Avatar jim19611961
04:17 PM Liked: 88
post #18 of 57
07-26-2012 | Posts: 1,421
Joined: Mar 2011
Quote:
Originally Posted by Bigus View Post

As to your specific questions... you may not need to worry so much about the floor reflection.

Can you elaborate?

Ive spent several hours trying to deal with it. I am very frustrated at the moment. I have tried anything and everything I have in my house that will lay flat and cant get it under -15db without using OC703 on it. Untreated, it rises to about -11db

Does anyone reading this thread have any suggestions / experience with floor reflections?
Bigus's Avatar Bigus
07:50 AM Liked: 51
post #19 of 57
07-27-2012 | Posts: 4,258
Joined: Feb 2003
When you place something absorptive on the floor, does it sound better? If not, you may be treating the numbers and not the sound. The Fraunhofer Institute in Germany detailed the design of a reference room in which the contribution of floor reflection was investigated. They found that eliminating a high gain spike as measured by ETC using thick absorption had a negative perceptual efffect. ETC identifies the reflections but you must still choose what to do with them.
jim19611961's Avatar jim19611961
08:04 AM Liked: 88
post #20 of 57
07-27-2012 | Posts: 1,421
Joined: Mar 2011
Yes, how it sounds is the most meaningful endgame.

But the model I am trying to emulate was based on far more listening tests than I could render in 6 months probably. The model in this case calls for -20db or better up to your ISD. Perhaps -15db will end up sounding fine. But if I can achieve the model objective, id rather start there and have the option to remove the floor treatment than to not have an option.

On another thought, I was considering what the ISD gap actually means in terms of the perceived room size. If your ISD gap is 12ms for instance, since sound in a typical situation will travel 13.5 feet in 12ms, does that equate to "sensing" the wall boundary 13.5 feet away (a 27 foot wide room)? And for a 20ms ISD, 22.5 feet (45 foot wide room), and so on?
jim19611961's Avatar jim19611961
01:53 PM Liked: 88
post #21 of 57
07-29-2012 | Posts: 1,421
Joined: Mar 2011
A few more days, a few more results and questions.

When we are speaking of terminating the ISD gap, most seem to agree that the directionality of that termination should be lateral. For reference, lets say 0 degrees is directly in from of you and 180 degrees is directly behind. Is 140 degrees acceptable? Preferable?
I gave a try at 90 degrees, but I just couldnt get enough energy there to get near the -12db termination threshold. Best I could do at 90 degrees was -15db or so. And the added complication of the 90 termination is I get an equal peak(s) at 22ms.

So given the choice, would you rather:

1) Have a -16db to -14db termination at 90 degrees (12ms) with a following peak(s) at the same magnitude (-14-16db) at 22ms
2) Have a -10db to -12db termination at 140 degrees (12ms)

?


Here is where I am at the moment:


kiwi2's Avatar kiwi2
04:19 PM Liked: 144
post #22 of 57
07-29-2012 | Posts: 1,630
Joined: Nov 2010
Quote:
Originally Posted by jim19611961 View Post

Does anyone reading this thread have any suggestions / experience with floor reflections?

What speakers do you have? Would you consider trying different speakers in the room that have limited vertical dispersion. i.e.. an MTM or ribbon or line arrays?

Newform Research talks a bit about floor and ceiling bounce here... http://www.newformresearch.com/technology.htm
jim19611961's Avatar jim19611961
04:42 PM Liked: 88
post #23 of 57
07-29-2012 | Posts: 1,421
Joined: Mar 2011
Quote:
Originally Posted by kiwi2 View Post


Newform Research talks a bit about floor and ceiling bounce here... http://www.newformresearch.com/technology.htm

The only I saw there was a mention that their product has limited vertical dispersion. Was there something I missed?
kiwi2's Avatar kiwi2
05:24 PM Liked: 144
post #24 of 57
07-29-2012 | Posts: 1,630
Joined: Nov 2010
Quote:
Originally Posted by jim19611961 View Post

The only I saw there was a mention that their product has limited vertical dispersion. Was there something I missed?

All through their site they talk about the benefits of reduced vertical dispersion.
http://www.newformresearch.com/home-theater-loudspeakers.htm

One way to avoid floor and ceiling bounce is to have a speaker system that doesn't produce it as much in the first place.
jim19611961's Avatar jim19611961
08:54 AM Liked: 88
post #25 of 57
07-30-2012 | Posts: 1,421
Joined: Mar 2011
Here is my newest. Left = Black, Right = Red



Then I tried facing the mics to the opposit channel. That is, while the left speaker was running the ETC sweep I faced the mic at the right speaker, and visa versa. I got something odd to me.



Where the termination peak occurs here is a lull in the other graph.

Here is a new room model showing some of the changes ive made.



Gray = Deflection panel
Brown = QRD panel
Black = Absorption panel
jim19611961's Avatar jim19611961
09:00 PM Liked: 88
post #26 of 57
07-30-2012 | Posts: 1,421
Joined: Mar 2011


Thanks Dragon smile.gif
amirm's Avatar amirm
09:11 PM Liked: 547
post #27 of 57
07-30-2012 | Posts: 18,375
Joined: Jan 2002
Quote:
Originally Posted by jim19611961 View Post

So given the choice, would you rather:
1) Have a -16db to -14db termination at 90 degrees (12ms) with a following peak(s) at the same magnitude (-14-16db) at 22ms
2) Have a -10db to -12db termination at 140 degrees (12ms)
The amplitude numbers are inaccurate in a room where reflections can go through different material such as yours. The spectrum gets modified and therefore the computed value cannot be compared to other reflections modified differently. I explained this using your case in a summary post in the long thread: http://www.avsforum.com/t/1413173/does-sound-sounds-better-in-a-room-full-of-furniture-and-stuff-or-without/720#post_22260049

The important factor is the angle of the reflection. Reflections coming from the side tend to be beneficial, the back and front ones not so much. So if by 90 degrees you mean from the side, then that is a better attribute than having the same coming from near rear of the room.

That is of course if you want to make your ears happy. If you want to keep making the meter happy instead, then plug away as Dragon is directing you smile.gif.
Bigus's Avatar Bigus
09:21 PM Liked: 51
post #28 of 57
07-30-2012 | Posts: 4,258
Joined: Feb 2003
His absorption panels are fairly broadband being 8-12" thick, and thus the differences in broadband ETC measurements as a result if differing spectral content is minimized.
jim19611961's Avatar jim19611961
09:42 PM Liked: 88
post #29 of 57
07-30-2012 | Posts: 1,421
Joined: Mar 2011
Quote:
Originally Posted by amirm View Post

The amplitude numbers are inaccurate in a room where reflections can go through different material such as yours. The spectrum gets modified and therefore the computed value cannot be compared to other reflections modified differently. I explained this using your case in a summary post in the long thread: http://www.avsforum.com/t/1413173/does-sound-sounds-better-in-a-room-full-of-furniture-and-stuff-or-without/720#post_22260049
The important factor is the angle of the reflection. Reflections coming from the side tend to be beneficial, the back and front ones not so much. So if by 90 degrees you mean from the side, then that is a better attribute than having the same coming from near rear of the room.
That is of course if you want to make your ears happy. If you want to keep making the meter happy instead, then plug away as Dragon is directing you smile.gif.

My ISD termination comes from a 140 degree angle (almost midway between directly back and directly to the side), but I have significant later reflections (arriving at 18-25ms) from 90 degrees (directly left and right).

I dont know why one would think I or anyone would go to all this trouble merely to please a meter.
Bigus's Avatar Bigus
10:11 PM Liked: 51
post #30 of 57
07-30-2012 | Posts: 4,258
Joined: Feb 2003
Jim, Amir believes ETC "lies, and lies badly." Take his advice however you see fit.

Reply Audio Theory, Setup, and Chat

Subscribe to this Thread

Powered by vBadvanced CMPS v3.2.3