Using Waterfall and ETC Graphs to analyze room response - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #1 of 162 Old 10-06-2012, 04:03 PM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646

This thread started as a PM converstation between myself and Jevansoh, who has been kind enough to assist me in understanding REQ waterfall and ETC graphs to analyze reflections and resonances in my listening room.  It has been moved to an open thread so we can share our progress with others, and perhaps receive other useful input.

 

Waterfall--doesn't look right

Conversation between markus767, jevansoh, and you

38x38px-ZC-392c8b23_avatar59748_1.jpeg
AustinJerry
Oct 1, 2012 at 3:10 pm

Here is a waterfall. It doesn't look right to me. Do I have settings incorrect? Please advise.

 

 

38x38px-ZC-0fad44ce_avatar-400-1.jpeg
jevansoh
Oct 1, 2012 at 3:33 pm
Jerry,

Please click "Limits" in REW and change the bottom limit from 0 to 30db. I can see 90db at the top (looking at the left hand side) and if that's what it is set to, then by setting to 30, we'll be mostly out of your noise floor and have a 60db spread. That will make the mountains in the foreground disappear (or closer to it) showing the floor and giving us more detail.

After setting a 60db spread using the limits function in REW (top right hand corner) by changing the top and bottom parameters to just show the top of the graph without cutting any off and the bottom 60db less than that, then also change the "right" parameter in the limits section from 200hz to 300hz.

Next, click "Controls" and set the time range to around 500ms. When you do that, if any parts of the waterfall are still outside the graph limits (the part at the bottom of the screen) then you'll have to raise it. Do this in 100ms increments. If, when you set the top/bottom numbers back at the Limits window to a 60db difference and that alone shows the entire waterfall, then you needn't perform this step.

I hope this helps. Feel free to ask any further questions and I'll be happy to guide you in any way I can.

Thanks so much for sharing this information with us. This is the most fun aspect of AVS for me! smile.gif

--J
 
38x38px-ZC-392c8b23_avatar59748_1.jpeg
AustinJerry
Oct 1, 2012 at 3:47 pm

Still doesn't look right.

 

 

 

38x38px-ZC-39edf2d1_av.png
markus767
Oct 1, 2012 at 4:04 pm
Hi Jerry,

In fact it does look "right", the problem is the underlying data. The problems starts with capturing LF data: the noise floor in our living rooms is way too high to capture the actual decay over 60dB. Then there's always a tradeoff between time and frequency resolution, i.e. if frequency resolution is good, time resolution suffers and vice versa. I didn't ask for waterfalls and now you know why wink.gif

Best, Markus
 
38x38px-ZC-0fad44ce_avatar-400-1.jpeg
jevansoh
Oct 1, 2012 at 4:07 pm
Jerry,

Actually, this is looking much better.

Unfortunately, it appears you have some major resonances here.

The time range is at 500ms. That was a start. This is now the only thing that needs changed to get a proper Waterfall.

I'd set it to 1000ms. If that makes the bottom disappear into the noise floor, then back it down to 900ms, etc, etc, until you can "just" see the floor like you can now at what appears to be around 175hz.

Also, it appears you may have some noisy conditions. Is this a dedicated room with any isolation or a family room/no isolation? I'd guess that latter. After seeing the full picture, depending on what time range you end up with, we may go tweak the bottom (back at the limits window) to 40db-45db instead as that is the noise floor in most residential common space homes without any proper isolation. It is very common, so don't feel bad. Very few people have noise floors less than 40db.

It's hard to get it right the first time or two. smile.gif Just keep playing with it and you'll get a nice looking graph that gives us plenty of detail and then the real fun begins. I can't wait to see the difference with Audyssey off vs. on vs. 4520. In my case, the resonances are worse with Audyssey on, even though the frequency response is better. That suggests some phase issues to me, but I'm not quite clear/sure what's happening and am curious if Markus has any input/ideas??

Send me another draft and I think we'll have it ready to post to the forum next time. wink.gif

Thanks again,

--J
 
38x38px-ZC-392c8b23_avatar59748_1.jpeg
AustinJerry
Oct 1, 2012 at 4:19 pm
I have tried 1,000 ms and it doesn't look much better. I think something else is wrong.

I don't have a dedicated listening room. The back of the room opens up to the rest of the ground floor (hallway, kitchen, dining room, etc.). The room has LF room treatments (four RealTraps mini-traps, two in each front corner), and otherwise the room is carpeted, with soft furniture. I am single, so taking measurements with the listening room in dead silence is not an issue. If I recall, REW shows my noise floor right around 35-40dB.
38x38px-ZC-0fad44ce_avatar-400-1.jpeg
jevansoh
Oct 1, 2012 at 4:51 pm
Jerry,

Go to the limits screen and set the bottom to 45db for now. Set the top to 100db.

Then go to controls and set the time range to 1000ms.

Send me what you get after doing that and let me take a look.

What type of mic setup are you using to do your measurements and have you done a soundcard and mic cal within REW?

What you sent before really did appear to be coming together nicely and was pretty much what I was expecting for the circumstances in which you described above, so I think we're on the right track.

Let me see this latest graph with these settings and we'll go from there.

Hang in there, we'll get it. Then you can always use these settings in the future and it won't be any hassle for further graphs. wink.gif (Unless you add more bass traps, which I HIGHLY recommend. Even cheap fluffy insulation does wonders, believe it or not. There's no reason to spend mega $$$ on brand name acoustical products as you've done, but they do look nice and if you have the money, it will save you some time and a bit of itching.)

--J

PS For extra credit, could you give me the full dimensions of your room(s)? L/W/H? Does your theater room just open up into the other rooms with a standard doorway opening of around 36" or less, in one area, or is it very open and basically one big room, acoustically speaking? If the latter, I need the full dimensions so I can calculate your estimated modes/resonances and see if they line up with the graph you're about to send. If they don't... We need to look into taking the measurements again and verify everything is set up correctly as your previous FR measurements may also be incorrect. If they do.... Then we're all good and you'll even have a plan for future treatment if you wish to make things better. Do you have any boominess now? Are you dissatisfied with anything in the lower frequency/bass range? Have you heard good tight bass in an accurate room before, for comparison, and how does your room compare?
 
38x38px-ZC-392c8b23_avatar59748_1.jpeg
AustinJerry
Oct 1, 2012 at 5:49 pm

 

 

OK, here is the latest iteration:

 

 

 

 

 

Rough room layout:

 

Ceiling height is 10'. With REW, I am using a SoundBlaster X-Fi soundcard with a EMM-6 individually-calibrated mic. The soundcard has been calibrated in REW properly, and I have loaded the appropriate calibration file for the mic.

 

I have worked continuously over the last several years to fine-tune my room and its equipment to achieve what I think is pretty good sound. To me, my bass quality is very good, both in smoothness and extension. The bass has been a primary focus, and I have experimented with room placement, the RealTraps treatments, upgrading to MultEQ XT32 (I also had an AS-EQ1 until I upgraded to the 4311), and adding a third sub (the one right behind the MLP.

 

The listening room:

 

 

 

 

This is the Pro calibration that I am using right now. Doesn't it look reasonably good?

 

 

38x38px-ZC-0fad44ce_avatar-400-1.jpeg
jevansoh
Oct 1, 2012 at 7:23 pm
Jerry,

Lookin great!

I will respond in detail this evening.

For now, by golly, I think you've got it re the Waterfall!

I'd probably set it to show 15-300hz and now set the bottom to 40db which seems to be your noise floor, as you previously stated. (That's what I do here to make certain that lowest mode shows)

Although I haven't looked in detail at everything, I did take a good but quick look at the waterfall, and once you set the above, it's ready to post to the forum and share.

I'd like to make my comments, at least the ones on the waterfall, in the forum so everyone can benefit if it's okay with you. I just figure if I'm going to do the analysis everyone might enjoy in seeing the results and we may get more people to participate.

Now I want to see this same waterfall with these same settings, both with Audyssey on and off and then again when you get the 4520.

Thanks, and again, I'll submit a detailed reply on the forum later this evening if that's okay with you.

--J
 
 
AustinJerry is offline  
Sponsored Links
Advertisement
 
post #2 of 162 Old 10-06-2012, 04:09 PM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646

Update

Conversation between jevansoh and you

38x38px-ZC-392c8b23_avatar59748_1.jpeg
AustinJerry
Yesterday at 2:28 pm

Hello,

 

Since we last exchanged PM's, several things have happened. First of all, I took delivery of the Denon 4520, went to the trouble of ripping out my 4311 and replacing it with the 4520, only to find that the 4520 had a defective right surround channel. I returned it the next day for credit. You may have seen my report in the 4520 thread. I am discouraged, to say the least.

 

Second, I have had a number of discussions with GIK regarding taming the 40Hz and 70Hz resonances that you pointed out on the waterfall graph. We discussed the Scopus and the Soffit traps. While these products may help with the resonance issues, my room does not allow placement of either of the traps in sufficient quantity to make a difference. I would have to make some significant alterations, either to the current location of speakers and subs, or removing/relocating existing room treatments (specifically my RealTraps minitraps) in order to fit in the GIK traps. I'm not ready to make drastic changes at this time, but may try and be creative if I can prove that changes will bring about improvements. It's a tough decision.

 

Finally, I believe you expressed a desire for posters to include more waterfall and impulse graphs. Previously, you were very helpful in assisting me in generating and interpreting waterfall graphs. Now, if you could indulge me, I need some guidance with impulse graphs, specifically the ETC graph. Here is my first effort, using a 20-20,000 Hz measurement at the MLP with Audysssey on and DEQ off:

 

 

 

 

Have I set the parameters correctly? If yes, what does this graph tell me? The distance from the main pulse to the first reflection is 8.1 ft. There is no wall surface 8.1 feet from the MLP, including the ceiling. Any guidance you might provide would be greatly appreciated.

 

One additional note. While I was waiting delivery of the 4520, I was preparing for moving the 4311 to the bedroom system, and decommissioning the 4308 that was previously in the bedroom (which I plan to sell). As an exercise, I ran two final Audyssey Pro calibrations, one with the 4311 (MultEQ XT32), and one with the 4308 (MultEQ XT). Both calibrations were run in the downstairs listening room, using the exact same eight measuring points and all other factors the same. The REW measurements of the results of this controlled test clearly show the technical differences between XT and XT32. While this topic has been discussed on and off in the Audyssey and the Pro threads, would you think it would be useful to post these results for others to view and comment? I know you are contemplating an upgrade to XT32, so perhaps you might find the results interesting as well.

 

38x38px-ZC-392c8b23_avatar59748_1.jpeg
AustinJerry
Yesterday at 2:36 pm

 

Edit: I just realized that the 8.1 feet needs to be divided by 2. So I am looking for a surface that is 4 ft from the MLP. The back wall is almost exactly 4 ft behind the MLP. Is this the culprit? If yes, how can I tame a reflection there? It already has a 24x48 ATS Acoustics panel. How can I tell what frequency is the issue?

 

I'm starting to enjoy this!

38x38px-ZC-0fad44ce_avatar-400-1.jpeg
jevansoh
Yesterday at 9:50 pm
Hi Jerry,

Lots and lots to cover here.

First, I need you to reset the parameters on the ETC because honestly, I can't tell a whole lot of detail from it.

It's not awful, and a very good first attempt, but we can do better.

We really only need to look at the first 40ms or so. Please go to the limits window and set the right to .040. Then set the bottom to around -50 or so, as we also want as much vertical info/detail as possible instead of it all being squished together in the middle of the screen/window.

Then resubmit to me and I will tell you everything you need to know.

For now though, I can say one thing. Whoever told you to divide by 2 is just plain dead wrong. Sorry, but I don't know how else to say it.

If your first reflection is at 8.1ft (not 8.1ms, right, but you used the right mouse button, pointed at 0, then dragged all the way over to the first spike while holding down the CTRL key, right?? That's how you convert the ms into feet. REW does it for you as long as you use this technique) that means the total path length, IN ADDITION TO the time it took for the direct sound to get to you (which has no reflections and is shown as the initial spike at 0 in REW) is 8.1ft.

So then what you have to do is (you can imagine this after you get some practice in and then verify you've taken care of it by running another ETC and seeing if the spike is gone) cut a piece of string.

You determine the length of the string by Adding the ACTUAL distance between the microphone and the mid-range driver of your speaker to this 8.1ft measurement.

Then attach one end at the mic, without moving it (it's okay to make the string longer to give you a place to tie it to at both ends, but be sure and mark on it the actual starting point at 0.0'/at mic and the actual 8.1ft+distance between mic and mid-range driver) and the other end to a tri-pod or other temporary device to hold the other end as close to the mid-range driver (including the height of it - literally right in front of it as close as you can get without touching it/damaging the driver) then pull the string taught.

With the string pulled taught simply move the string around, keeping it taught, in all directions until it touches a hard reflective surface. When it does (this isn't necessarily "exact" but pretty darn close, so if the string touches a surface but isn't totally taught or if it "almost" touches a surface but doesn't quite reach, that could still be your culprit) you've found the culprit.

It's possible it may reach two or even three (uncommon) surfaces which means it is actually several reflections at that point in time/distance and you'll need to take care of all those reflections.

You need to try to tame all reflections that are above 20db between 0ms and 20ms. Anything that is below 20db already, you don't need to worry about, as those reflections can be helpful.

Note I did NOT say to necessarily absorb them. That is one way, and the most common way, but you could also diffuse them, which has inherent absorption properties in it, or to save the energy and not over deaden the room, you could simply reflect it away to the back of the room or anywhere else outside the acoustic bubble surrounding the MLP.

If you do decide to go the easiest route and simply place some absorption there, make sure it is at least 4" thick and if at all possible leave a 2-4" air gap, 4" preferably.

These types of reflections that are shown on the ETC are specular in nature. They are not tied to one frequency or band of frequencies. They are ALL frequencies above Schroeder/transition range of about 300hz or so. So you definitely want to make certain your insulation (if using OC703 or mineral wood like Safe N Sound) is at least 4" thick to be able to absorb frequencies that low. If you use thinner material which will, for instance, only absorb down to 1khz, like let's say only for example, 1" of material, then you are simply absorbing the high frequencies and the mid/lower frequencies are still reflecting, the reflection/spike will still show, and all you've really done is EQ the top end/high frequencies out of the room without using electronics. This is called passive EQ'ing and you do NOT want to do this.

I hope I haven't missed anything and have answered all your questions. If I've forgotten something, please let me know, and I'll be sure to answer.

If something isn't clear, let me know and I'll try to clarify.

I'll go into more detail and do more analysis when you get me the new graph with the new parameters.

I DEFINITELY think you need to share those results. Those are the type of results I keep looking for and don't find!!!!

I also hate to see this conversation and our dialog back and forth not be able to help more than one person.

IMHO you need to start a room treatment / acoustics discussion thread, or find an active thread, and move our PM's (you have my permission) over there so everyone can benefit and our conversations can continue in public. wink.gif Just an idea though.

Thanks Jerry,

--Jason
 
38x38px-ZC-392c8b23_avatar59748_1.jpeg
AustinJerry
Yesterday at 10:58 pm

Hi Jason,

 

Thanks for the lengthy and helpful response! Here is the updated ETC graph:

 

 

How did I get the idea to divide the distance by two? I did a search for "how to interpret an ETC graph" and opened a posing on Hometheatershack.com. I'm sure I misinterpreted the post, thanks for correcting me. And yes, I used the CTRL-Right Click to determine the 8.1 feet distance.

 

The actual distance from the mid-range driver to the mic, which is at the MLP, should be the Audyssey distance for the speaker, no? All three front speakers are 9.4 Ft from the MLP. The ETC is generated by using a measurement with the mic at the MLP, using the mono REW signal fed to both left and right speakers, plus the subs. So I use the 9.4 ft distance, and not the sub distances, correct? This would make the string 17.5 feet long. I'm going to cut my string as soon as I finish this response.

 

Testing for understanding. Is this the correct way to use the string?

 

 

In the revised ETC, I see two reflections that are over 20dB (only slightly), the one at 8.1 ft, and a second one at 13.6 ft. I'll string them both. Once I determine the reflection point, we can discuss how it might be tamed.

 

And yes, it would be nice to get these exchanges into an open thread so others could read and participate. I'll think about that tomorrow as well.

 

Thanks again for your helpful input.  As soon as I have done the string test, I'll report back.

 

Jerry

 

 

 

 

 

 

 

38x38px-ZC-392c8b23_avatar59748_1.jpeg
AustinJerry
Yesterday at 11:23 pm

 

 

OK, that was easy, and the result was predictable. Here are two pictures of my listening room:

 

 

 

With the 17.5 ft string attached to the front left speaker, the reflection point is right in the middle of the brick fireplace on the left wall. With the string on the right speaker, the reflection point is right in the middle of the third ATS Acoustics panel on the right wall. I suspect the fireplace is the issue.

 

I could generate two new ETC graphs, this time measuring the left and right speakers independently, which I'll bet confirms the fireplace as the issue.

 

I only did the 17.5 ft string, not the 23 ft string, which I'll do tomorrow.

 

What do you think? How could the fireplace be treated?

 

Jerry

 

38x38px-ZC-0fad44ce_avatar-400-1.jpeg
jevansoh
Yesterday at 11:46 pm
Jerry,

Oops....Don't get mad at me (a little more work for you) but I forgot (I need to learn not to assume) to tell you that the ETC absolutely has to be performed for each individual speaker. It's okay if the subs are in the mix as again, we are only looking at, testing for, and treating the specular region (above Schroeder) but whether the subs are in the mix simply doesn't matter.

You have to leave it all hooked up the way you have it. You have it all hooked up right from your description in your last PM. You need to simply disconnect, from the speaker itself, the speaker wires either for the left/right speaker so you're only testing one at a time, otherwise who knows which results are accurate. Just because you found a "potentially" harmful reflection at both ends doesn't mean that both actually are harmful... That's why you need to do them one at a time. Sorry for not mentioning that before.

Also, I missed one other thing...My bad... You need to change (at the top left hand corner) from DB% to DBFS. Just click on it and the only other option is DBFS. This could make all the difference in the world... Don't go placing absorption or doing anything else until you change this. Also, when you change it, and if you keep the other parameters the same, it "should" then fill out the spikes to the full top/bottom of the screen, or close to it. If it doesn't, or doesn't "fully" then you can click the little buttons right by the DBFS option to make the spikes bigger/smaller. Make it so they just barely fit in your screen and you can also scroll the graphs so we can see the tops and bottoms of the spikes/peaks.

Once you do this, 0 will be at the top, and any spike that appears between 0 and -20 (up to 20ms) is what you need to worry about.

Can you do this one more time, since I didn't clarify, because I'm afraid I've confused you and have you looking at the wrong things...Again...Sorry about that. tongue.gif

I like that drawing. Did you find that example somewhere or did you make it? If it's yours, may I have your permission to use it to help other people understand as pictures are worth a thousand words and it's PERFECT! wink.gif

Thanks again, and I look forward to seeing the accurate graph. Hopefully it's not too late for you to get it to me still yet tonight as I don't know when I might be able to reply/help tomorrow. I will be around tomorrow evening, but want to try to help you tonight and point you in the right direction so you can have all of tomorrow to hunt all these down.

After you find the first few reflections with the string, I think you'll be able to simply visualize the string method and with a little trial and error not have to go to all the trouble of actually using the string then simply run another ETC when you think you've found the problem (and have the insulation placed of course) to verify you've got it.

Also, moving backwards, there was something I forgot to reply to.

You sent me pictures and dimensions of your room, but only kinda sorta. smile.gif By your room being open in the back, acoustically speaking, whatever is behind it is really part of the room too.

Can you tell me the full dimensions of the open space and send me pictures from your MLP facing backwards?

I have some ideas for some EASY and INEXPENSIVE treatment ideas for you but need some more info first.

C ya,

--Jason (Friends call me J)
 
38x38px-ZC-392c8b23_avatar59748_1.jpeg
AustinJerry
Today at 12:02 am
Hi J,

No problem , I will re-run the measurements and generate new ETC graphs. Sorry, too late to do it tonight, everything is shut down and I'm nodding off. I'll also provide the pictures and dimensions of my ground floor. No hurry on the response--I have been working on this for over two years now.

Jerry
 
AustinJerry is offline  
post #3 of 162 Old 10-06-2012, 04:22 PM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646

As promised, here is a quick drawing of my ground floor:

 

 

 

 

Here are several pictures showing how the listening room opens to the rest of the ground floor:

 

 

 

 

 

AustinJerry is offline  
post #4 of 162 Old 10-06-2012, 04:49 PM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646

Hello J,

 

I have taken a number of new measurements this morning using REW.  Hopefully I have all the measurements that we will need to complete the analysis.  All measurements were taken with the mix at the MLP.  Specifically, the measurements I collected were:

 

- Left front speaker only, Audyssey off and Audyssey on, 10-20,000 Hz

- Right front speaker only, Audyssey off and Audyssey on, 10-20,000 Hz

- Center speaker only, Audyssey off and Audyssey on.= 10-20,000 Hz

- Left and right front speakers, Audyssey off and Audyssey on, 10-20,000 Hz

 

- Left front sub only, Audyssey off and Audyssey on, 10-300 Hz

- Right front sub only, Audyssey off and Audyssey on, 10-300 Hz

- Rear sub only, Audyssey off and Audyssey on, 10-300 Hz

- Both front subs, Audyssey off and Audyssey on, 10-300 Hz

- All subs, Audyssey off and Audyssey on, 10-300 Hz

 

- Left and right front speakers and all three subs, Audyssey off and Audyssey on, 10-20,000 Hz

 

Here is the new ETC graph for the left front speaker:

 

 

I tried the string technique on the first three peaks (5ft, 7.1 ft, and 13 ft).  I'm not too sure about the results:

 

5ft - The front panel of my Left Wide speaker (suspect)

7.1ft - The front of the brick fireplace (reasonable)

13ft - The wall in the left rear corner (suspect)

 

Comments?

 

I also generated a new waterfall for the subs:

 

 

I am looking forward to your feedback.

 

- Jerry

 

Edit:  If you would like to look at the REW MDAT files, I have uploaded them to DropBox and can provide a public link to download.

AustinJerry is offline  
post #5 of 162 Old 10-06-2012, 06:23 PM
AVS Special Member
 
HopefulFred's Avatar
 
Join Date: Mar 2007
Location: Atlanta, GA
Posts: 2,695
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 97 Post(s)
Liked: 184
Thanks for posting Jerry. I don't have too much constructive feedback, but I'm interested in following along.

A couple things occurred to me that may be relevant (maybe not).

First, even though you've done the math to compute the total distance that corresponds to each specular reflection, I thought REW had a simpler way to display the graph (use loopback as reference -? something like that) where t=0 is when the signal is actually generated, so the initial spike doesn't show up at zero, but instead at the time that it actually was perceived by the mic. There may be plenty of reasons not to do that, including not wanting to remeasure, and maybe concerns over the latency of the rest of the signal chain.

Second, and this is mostly taken care of through the cautions that using the string only gets you to the approximate area of the specular reflection, but Snell's law should be applied; the angle of incidence must be equal to the angle of reflection. I'm not saying you need to whip out your handy-dandy protractor, but it's something to keep in mind. For instance, if the string were to be stretched straight out to the side of a speaker, to touch a side wall at a near-normal (90 degree) angle, the reflected sound shouldn't be expected to predominately head towards the MLP - that would be changing direction. You probably understood this already (at least intuitively at some level - or you wouldn't have been able to figure out where to put your panels). Also, there is an argument to be made that once the sound wave interacts with a rigid boundary, that boundary becomes an apparent source - propagating sound in "all" directions; but that's not entirely true for specular reflections - this would, however, explain some of the energy losses associated with the reflection. Most of the sound energy will maintain it's direction as though it were a billiard ball.

Third - I'm surprised that you didn't identify a floor-bounce. Did you check for them?

Thanks again for posting - and I hope I'm not muddying the waters by interjecting.

Fred
HopefulFred is offline  
post #6 of 162 Old 10-06-2012, 06:42 PM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646
Quote:
Originally Posted by HopefulFred View Post

Thanks for posting Jerry. I don't have too much constructive feedback, but I'm interested in following along.
A couple things occurred to me that may be relevant (maybe not).
First, even though you've done the math to compute the total distance that corresponds to each specular reflection, I thought REW had a simpler way to display the graph (use loopback as reference -? something like that) where t=0 is when the signal is actually generated, so the initial spike doesn't show up at zero, but instead at the time that it actually was perceived by the mic. There may be plenty of reasons not to do that, including not wanting to remeasure, and maybe concerns over the latency of the rest of the signal chain.
Second, and this is mostly taken care of through the cautions that using the string only gets you to the approximate area of the specular reflection, but Snell's law should be applied; the angle of incidence must be equal to the angle of reflection. I'm not saying you need to whip out your handy-dandy protractor, but it's something to keep in mind. For instance, if the string were to be stretched straight out to the side of a speaker, to touch a side wall at a near-normal (90 degree) angle, the reflected sound shouldn't be expected to predominately head towards the MLP - that would be changing direction. You probably understood this already (at least intuitively at some level - or you wouldn't have been able to figure out where to put your panels). Also, there is an argument to be made that once the sound wave interacts with a rigid boundary, that boundary becomes an apparent source - propagating sound in "all" directions; but that's not entirely true for specular reflections - this would, however, explain some of the energy losses associated with the reflection. Most of the sound energy will maintain it's direction as though it were a billiard ball.
Third - I'm surprised that you didn't identify a floor-bounce. Did you check for them?
Thanks again for posting - and I hope I'm not muddying the waters by interjecting.
Fred

Fred, thanks for your input, and welcome to the discussion. As for the distance calculation, in REW if you hold the CTRl key, right-click, and drag the cursor, the distance is displayed. Add the actual distance from the speaker to the mic (in my case 9.4 ft as calculated by the Audyssey calibration), that's how I arrived at the distances. So, the first reflection is at 5 ft, add 9.4 ft, and the string is 14.4 ft long, etc.

I understand that the angle of incidence must equal the angle of reflection, which is why only the reflection that is seemingly being caused by the fireplace along the left wall is the one that seems valid to me. The other two don't satisfy this law.

I have not considered either the floor bounce, or a ceiling reflection at this time. The analysis is still a work-in-progress. I'm not to sure how to check for either of these conditions, so any input would be appreciated.

- Jerry
AustinJerry is offline  
post #7 of 162 Old 10-06-2012, 07:18 PM
AVS Special Member
 
HopefulFred's Avatar
 
Join Date: Mar 2007
Location: Atlanta, GA
Posts: 2,695
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 97 Post(s)
Liked: 184
Quote:
Originally Posted by AustinJerry View Post

I have not considered either the floor bounce, or a ceiling reflection at this time.
- Jerry
I just mean that you shouldn't forget those surfaces when you are working with your string. There shouldn't be anything different about them, but sometimes we forget we are standing on and under reflective surfaces.
HopefulFred is offline  
post #8 of 162 Old 10-06-2012, 09:19 PM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 18,748
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 488 Post(s)
Liked: 607
Quote:
Originally Posted by HopefulFred View Post

Snell's law should be applied; the angle of incidence must be equal to the angle of reflection. ...For instance, if the string were to be stretched straight out to the side of a speaker, to touch a side wall at a near-normal (90 degree) angle, the reflected sound shouldn't be expected to predominately head towards the MLP - that would be changing direction.
If the string is the correct length and pulled taut, is it possible to be able to touch boundries where the angle of incidence is not the angle of reflection (i.e., not the shortest path from speaker to boundry to ear)?

Sanjay
sdurani is online now  
post #9 of 162 Old 10-06-2012, 11:42 PM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 4,716
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 302 Post(s)
Liked: 171
For calculating the arrival time of boundary reflections this tool can be used:
http://mehlau.net/audio/floorbounce/

For general conversions of frequency/distance/wavelength/time:
http://mehlau.net/audio/calculator/

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is online now  
post #10 of 162 Old 10-07-2012, 01:45 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 4,716
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 302 Post(s)
Liked: 171
While ETCs and waterfalls are very useful, it's important to understand that they represent an interpreted view at reality and not reality itself. It's crucial to understand how these graphs are calculated and how parameters like window time influence what the results show.

Tons of valuable information on these topics can be found at the REW forum. Look out for posts from John Mulcahy, which is the author of REW. For example: http://www.hometheatershack.com/forums/rew-forum/45927-waterfall-window-parameter-please-explain-me.html#post435994

Regarding ETCs and finding the matching reflection points: simply put a movable absorber at the point in question, then re-measure. The spike should be gone. Don't look just at first order reflections, reflections bouncing off two or three walls (for example near a corner) can be very high in level.

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is online now  
post #11 of 162 Old 10-07-2012, 08:50 AM
AVS Special Member
 
jim19611961's Avatar
 
Join Date: Mar 2011
Location: Texas
Posts: 1,246
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 79
I too had a reflection point at a fireplace. What I did was make a panel out of 1x2 that was 8" deep out of 2" oc703 withe 6" pink fluffy behind that that covered the fireplace, but was free standing so if I wanted to do a fire, I could just lift the panel and sit it in another room.

I saw that you were being guided to get the first 20ms of the ETC response down to -20db. This I agree with and will be worth the effort if you can get there.

]

Here is my ETC. I am not using REW, but the data is still relevant. Even though I am not quite there on the -20db threshold (have a floor bounce at about 2ms @-15db and another at 14-15ms @-18.5), I am finding that this works pretty well. The big hump at 24ms is my ISD gap termination. That is there on purpose.

Anyway, my ETC's looked something like yours before I delved completely into treatment, so I think you can get there.

BTW... I live about 2 1/4 hours from you smile.gif

My Room
http://www.gearslutz.com/board/studio-building-acoustics/817205-my-listening-room.html

My Music
http://rateyourmusic.com/~jim1961

My Equipment

Rega - Apollo
Rega - DAC
Goldpoint Passive
(2) Classe CA-100 bridged power amps (350w)
Jenzen Next ( http://www.troelsgravesen.dk/Jenzen-NEXT.htm )
...
jim19611961 is offline  
post #12 of 162 Old 10-07-2012, 09:35 AM
AVS Special Member
 
HopefulFred's Avatar
 
Join Date: Mar 2007
Location: Atlanta, GA
Posts: 2,695
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 97 Post(s)
Liked: 184
Quote:
Originally Posted by sdurani View Post

Quote:
Originally Posted by HopefulFred View Post

Snell's law should be applied; the angle of incidence must be equal to the angle of reflection. ...For instance, if the string were to be stretched straight out to the side of a speaker, to touch a side wall at a near-normal (90 degree) angle, the reflected sound shouldn't be expected to predominately head towards the MLP - that would be changing direction.
If the string is the correct length and pulled taut, is it possible to be able to touch boundries where the angle of incidence is not the angle of reflection (i.e., not the shortest path from speaker to boundry to ear)?
I don't want to derail this discussion, but I think you're missing something. For the sake of conversation and example, let's consider it only in two dimensions. If you were to draw the room as the OP has posted, you could use a string to scale to identify walls that might represent the total length needed to create the delay. You could measure the angles from there, but you'd find that the walls would need to coincide with the edge of an ellipse.
HopefulFred is offline  
post #13 of 162 Old 10-07-2012, 10:03 AM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 18,748
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 488 Post(s)
Liked: 607
Quote:
Originally Posted by HopefulFred View Post

I don't want to derail this discussion, but I think you're missing something. For the sake of conversation and example, let's consider it only in two dimensions. If you were to draw the room as the OP has posted, you could use a string to scale to identify walls that might represent the total length needed to create the delay. You could measure the angles from there, but you'd find that the walls would need to coincide with the edge of an ellipse.
Maybe I am missing something: where are the listener position and boundries in that diagram? Isn't the string supposed to go from the speaker to the boundry to the listener position? If so, then it will only touch each boundry at one location (unless your walls really are eliptical).

Sanjay
sdurani is online now  
post #14 of 162 Old 10-07-2012, 10:48 AM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646
Quote:
Originally Posted by sdurani View Post

If the string is the correct length and pulled taut, is it possible to be able to touch boundries where the angle of incidence is not the angle of reflection (i.e., not the shortest path from speaker to boundry to ear)?

 

Hello, Sanjay.  Thank you for your response.

 

Yes, using the distances calculated from the first three peaks in the ETC graph, 14.4 ft, 16.5 ft, and 19.6 ft, the first and third touch boundaries where the angles if incidence and reflection are not equal.  That is why I marked them as suspect.  At 14.4 ft, the string touches the side of my DSX Wide speaker, which is approximately 4 ft to the left and 2 ft in front of the left front speaker. At 19.6 ft, the string touches the drywall in the left rear corner of the room.  The second distance seems to coincide with the brick facing on my fireplace.  This early reflection point was initially discovered using the "mirror test", but I have not figured out the best way to address it.

 

Following Markus' suggestion, I used several ATS Acoustics 24x48 panels and tried placing them in front of the fireplace.  A repeat ETC measurement did not show any difference in the peak at 16.5 ft.

 

I need to test whether the reflections are the result of floor bounce or the ceiling, and would appreciate suggestions on how to do this.  It is not very easy to put a temporary treatment on a 10 ft ceiling.

 

Here is a picture of the three distances I was testing.  Only the green line looks to be valid.  The more I think about it, the 14.4 ft distance may sugest a ceiling reflection...

 

 

 

 

 

 

- Jerry

AustinJerry is offline  
post #15 of 162 Old 10-07-2012, 10:55 AM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646

One more question for anyone:  please put things in perspective for me.  Is the ETC showing a "terrible" problem, or a fairly minor problem?

 

It would be useful to train my ears to identify harmful reflections.  Does anyone know of a collection of audio files that demonstrates the effects of various levels of harmful room reflections, resonances, etc?  (OK, that actually was two questions, but I am excited about this learning experience!)

 

Thanks.

AustinJerry is offline  
post #16 of 162 Old 10-07-2012, 10:58 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 4,716
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 302 Post(s)
Liked: 171
Jerry, place the panel directly to the left or right side or top of the speaker, then re-measure. This will show you generally where to look for reflection points.

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is online now  
post #17 of 162 Old 10-07-2012, 11:03 AM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646
Quote:
Originally Posted by markus767 View Post

Jerry, place the panel directly to the left or right side or top of the speaker, then re-measure. This will show you generally where to look for reflection points.

 

Good idea, Markus.  How about immediately above or below to test for ceiling or floor reflections?

AustinJerry is offline  
post #18 of 162 Old 10-07-2012, 11:07 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 4,716
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 302 Post(s)
Liked: 171
Quote:
Originally Posted by AustinJerry View Post

One more question for anyone:  please put things in perspective for me.  Is the ETC showing a "terrible" problem, or a fairly minor problem?

That is a very good question and other's have written whole books with several hundreds of pages about it, e.g. Toole "Sound reproduction".

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is online now  
post #19 of 162 Old 10-07-2012, 11:07 AM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 18,748
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 488 Post(s)
Liked: 607
Quote:
Originally Posted by AustinJerry View Post

Only the green line looks to be valid.
Maybe I'm reading your diagram wrong, but the angle of incidence looks like the angle of reflection for the green line. Is it not?

Sanjay
sdurani is online now  
post #20 of 162 Old 10-07-2012, 11:10 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 4,716
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 302 Post(s)
Liked: 171
Quote:
Originally Posted by AustinJerry View Post

Good idea, Markus.  How about immediately above or below to test for ceiling or floor reflections?

That's how I eliminate the ceiling reflection:


Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is online now  
post #21 of 162 Old 10-07-2012, 11:13 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 17,768
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 493 Post(s)
Liked: 323
Quote:
Originally Posted by AustinJerry View Post

One more question for anyone:  please put things in perspective for me.  Is the ETC showing a "terrible" problem, or a fairly minor problem?
It potentially showing something beneficial let alone a "problem!" Reflections perceptually have different effects on us. Side reflections for example can be pretty pleasing. Front reflections are very hard to hear their effects as are the rear ones. Floor reflections change timbre but their effect is limited to reflections above 500 Hz (so a carper plus good padding can deal with them). Ceiling reflections are a bit tricky but probably better to not have them be there.

Even putting these points aside, reflections do not mean the amplitudes you see in ETC. That is because if they go through anything other than a bare wall, their spectrum changes. And once that is done, then you can't compare their amplitudes. Both listening tests and measurements show that a spike that is higher in amplitude may actually be lower in audibility than another!

People in the professional space have developed acuity with respect to reflections so if you go by their advice, all early reflections are bad and should be eliminated. Research into recreational listening like all of us do says that this advice does NOT generally apply to us . So please be careful in chasing the path of making measurements pretty. You may be sacrificing audio quality in the process.

What I just said is hotly debated in this forum. Fortunately every argument and then some was made recently and is available to read through. Likely will take you many hours if not days to read and you need high ability to get past bickering smile.gif. If you can do that, here it is: http://www.avsforum.com/t/1413173/does-sound-sounds-better-in-a-room-full-of-furniture-and-stuff-or-without
Quote:
It would be useful to train my ears to identify harmful reflections.  Does anyone know of a collection of audio files that demonstrates the effects of various levels of harmful room reflections, resonances, etc?  (OK, that actually was two questions, but I am excited about this learning experience!)

Thanks.
As I mention in the above thread, if you are convinced of the extensive amount of listening tests and research that testifies to some reflections being good and others having the characteristics I mentioned, then the strategy of how to deal with them does not require this type of measurements. You simply look at the layout of the room and decide on a course of action. Combine that with making sure you have sufficient amount of absorption in the room to make sure it is not "too live" and optimization of the bass and you are good to go.

Good luck smile.gif.

Amir
Founder, Madrona Digital
"Insist on Quality Engineering"

amirm is offline  
post #22 of 162 Old 10-07-2012, 11:22 AM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646

^ Thanks for your advice, Amir.  I'll try and wade through the thread you recommend.  And thanks for your interesting and educational articles in WSR.  They are among the more useful information that I have read recently.

AustinJerry is offline  
post #23 of 162 Old 10-07-2012, 11:25 AM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646
Quote:
Originally Posted by sdurani View Post

Maybe I'm reading your diagram wrong, but the angle of incidence looks like the angle of reflection for the green line. Is it not?

 

Yes, it looks as if that is the case, which is why I said that it looked valid.  I'm going to try some temporary treatment panels there and generate e freash ETC.

AustinJerry is offline  
post #24 of 162 Old 10-07-2012, 01:26 PM
AVS Special Member
 
HopefulFred's Avatar
 
Join Date: Mar 2007
Location: Atlanta, GA
Posts: 2,695
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 97 Post(s)
Liked: 184
Quote:
Originally Posted by sdurani View Post

Maybe I am missing something: where are the listener position and boundries in that diagram? Isn't the string supposed to go from the speaker to the boundry to the listener position? If so, then it will only touch each boundry at one location (unless your walls really are eliptical).
That was my point - that for there to be lot's of possibilities that satisfy both Snell's law and the total length, the room would need to be elliptical. Here's another example. I've built it from 3/4/5 right triangles so that I could do the math. I've drawn in the approximate normal lines and the surface that would be required to generate the reflection.



In this case, let's assume that the ETC indicates direct sound from 8 feet and strong reflection with total distance of 10 feet. There are technically infinite surfaces that might exist that could create that reflection. The only ones that make sense are the floor three feet below, the ceiling if 3 feet above, or a wall either 1 foot behind the listener or 1 foot behind the speaker (ignoring side walls, to simplify to two dimensions). I've drawn in the most likely floor bounce, which lands mid-way to the LP, given that the speaker and listener are at the same height. Clearly, the spot directly below the speaker couldn't be the source for the reflection, as it requires a total distance of 11.5 feet, and it would need to be angled, approximately as indicated to maximize reflected signal strength. Keeping the conversation only in two dimensions for the moment, there are other potential places, such as 1.8 feet below the speaker (the square root of the quanitity 1.8 squared plus 64 is 10, just like for the likely floor bounce). However, again, as shown, the required reflective surface would need to be angled.

This, as shown, would not be a likely area of actual confusion for anyone trying to do this. On the other hand if there were a side wall 2 feet from the speaker, a 10 feet long reflection path might be imagined on that wall - the string could certainly be made to touch the speaker, the wall, and the listener - however, the angle would not be correct. The wall could not be parallel to the line between the listener and the speaker and generate a strong reflection at the appropriate time. If they are all parallel - the wall reflection would be timed for just under 9 feet, and like the floor reflection , be midway between the LP and source. I don't know what the angle of the reflecting surface would need to be, but I know it would be tangent to an ellipse.

Sorry for the distraction.
HopefulFred is offline  
post #25 of 162 Old 10-07-2012, 03:40 PM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 18,748
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 488 Post(s)
Liked: 607
Quote:
Originally Posted by HopefulFred View Post

In this case, let's assume that the ETC indicates direct sound from 8 feet and strong reflection with total distance of 10 feet. There are technically infinite surfaces that might exist that could create that reflection.
I don't get the part about infinite surfaces. There should only be one point where the string can be pulled tight AND touch a boundry; unless your speaker is literally the exact same distance from more than one boundry, and how often does that happen? If you look at the previous diagram you posted of the elipse inside the rectangle, shrinking that rectangle (room) to touch the elipse would result in only one contact point per boundry, not infinite surfaces.

Earlier you said "if the string were to be stretched straight out to the side of a speaker, to touch a side wall at a near-normal (90 degree) angle, the reflected sound shouldn't be expected to predominately head towards the MLP". The reason I was questioning that comment is because if the string were to be stretched straight out to the side of a speaker (90 degrees), it would never touch a side wall. Take a look:

64190

The only place it looks like it would touch the side wall is where the angle of incidence IS the angle of reflection. So no need for the OP to be careful to observe Snell's law, since pulling the string taut seems to do that automatically.
Quote:
Originally Posted by HopefulFred View Post

Sorry for the distraction.
No need to apologize since it's not a distraction nor off topic to discuss the string method of finding reflections in a thread about using "ETC graphs to analyze room response".

Sanjay
sdurani is online now  
post #26 of 162 Old 10-07-2012, 04:55 PM
AVS Special Member
 
HopefulFred's Avatar
 
Join Date: Mar 2007
Location: Atlanta, GA
Posts: 2,695
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 97 Post(s)
Liked: 184
Quote:
Originally Posted by sdurani View Post

Quote:
Originally Posted by HopefulFred View Post

In this case, let's assume that the ETC indicates direct sound from 8 feet and strong reflection with total distance of 10 feet. There are technically infinite surfaces that might exist that could create that reflection.
I don't get the part about infinite surfaces. There should only be one point where the string can be pulled tight AND touch a boundry; unless your speaker is literally the exact same distance from more than one boundry, and how often does that happen? If you look at the previous diagram you posted of the elipse inside the rectangle, shrinking that rectangle (room) to touch the elipse would result in only one contact point per boundry, not infinite surfaces.
The image I posted first, with the ellipse, was not drawn with any rectangular room in the figure - I see a rectangle in the image you've reposted in your quote, but not in my original. The intention was to show that the elliptical room generates an infinite number of potential reflective surfaces all with the same total time for the reflection. The ellipse is the room boundary in that case. Also, even if there are multiple surfaces the same distance from the speaker, those won't result in the same delay unless they are also the same distance from the LP. See my second post and figure for some theoretical possibilities with different distances, such as a reflection behind the speaker looking like a reflection behind the listener on an ETC. If a room were a 3-dimensional elliptical sphere, the ETC would show direct sound, followed by a huge spike as all the first-order reflections impeded the LP at the same instant. Second order reflections would not be so organized.

I think somehow this confusion stems from the rectangle you see in the image I posted - it's probably just the edge of the image - it's not supposed to be a part of the room diagrammed.

So to summarize - finding multiple paths that satisfy the delay indicated on the ETC might be easy, but only one should satisfy Snell's law as well. I think my second post was more clear. In the first post I was trying to be brief, since I figured my point was largely academic.
HopefulFred is offline  
post #27 of 162 Old 10-07-2012, 11:16 PM
Senior Member
 
jevansoh's Avatar
 
Join Date: Oct 2010
Location: Southern, Ohio
Posts: 467
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 36
Hi Jerry et al,

First, I want to say how happy and excited I am that the public posting of our conversation via PM has sparked the exact discussion I had hoped it would.

Secondly, you've asked a lot of great questions, and every time I get ready to respond, someone (Markus is obviously a lot faster than me wink.gif ) has already answered.

I want you to know I am following along and if you have any specific questions for me or if I find any questions that haven't yet been answered or answers/comments with which I do not agree, I will be actively following this thread and will respond ASAP.

As for the string method and interpreting the results... While it is fun to discuss this complex topic of acoustics, theories, analysis, etc, including whether or not you should be attenuating the specular reflections to begin with you must first set a goal as to what you'd like to achieve.

My philosophy is that you should CHOOSE a model such as LEDE/RFZ, Ambechoic, based on specific criteria and what you wish to accomplish in your room. If you do not know which model you wish to attempt to achieve, then I'd recommend simply absorbing early reflections up to about 20ms and any high spikes after 20ms then look at FR and Waterfall and if everything is in line, listen this way for awhile. If you like it, great! You can always learn more and try other models. It's fun to experiment. I wouldn't worry about terminating the ISD gap in this case until you know specifically what your goals are.

Your ETC is absolutely not horrible. While I agree with Amir that reflections are not necessarily bad and can absolutely be good and are even wanted/needed in some models, it's where/when those reflections appear and at what amplitude compared against which model you wish to achieve that determine whether or not they should be there and not simply making a blanket statement that they are good or bad. wink.gif Obviously all of this is IMHO and take it for what it's worth, but I have been around for a bit and of course there is science to back some of this up. wink.gif

I do think you may be over thinking the string method a bit. If you're having trouble finding the exact spot of the reflection, then I'd simply place a panel (or a 2'x2' or even 15" x 24" or so piece of R30 fluffy stuff) each place you suspect might be the culprit then re-run the ETC. If the spike is gone, one by one remove each panel/piece of insulation until the spike reappears, running the ETC in between each time you move/remove a piece, of course. You won't necessarily learn the why or the how but you'll get the result you need and it's pretty painless/quick.

As for checking the ceiling reflection and finding it without a ton of work, time, and a bunch of holes in your ceiling, simply grab a 1x2x10' piece of lumber (about $2.00 at Lowes/HD) and that trusty piece of R30 and use this "stick" to hold it in place at the spot on the ceiling you think is causing the problem. Once you verify the right spot via a new ETC measurement, you can attach a permanent panel and only take the chance of breaking your back (or a sweat) once, climbing that ladder.

As to the gentleman who posted his ETC and mentioned the one remaining spike/termination of ISD... This is a great effort and you should be proud. I'm sure it took a lot of work/effort to get where you are. Please don't confuse one loan specular reflection that happens to be past 20ms on a graph a true termination of an ISD Gap. A true termination is no less than 12db down plus is followed by a diffuse return of decaying energy. It appears the energy on both sides of your loan spike is no more or less diffuse and is at roughly the same level. I would consider this more of a good start or if you've not chosen a LEDE/RFZ model I'd consider attenuating this loan spike, too. Again, great effort, but it's not quite finished.

If anyone has any questions that aren't answered (I am unfortunately only available about once a day around this time to participate on AVS as I'm in the process of starting a new acoustics company with some really exciting new QRD affordable diffuser options) before I can get to them, just know I am monitoring this thread and will respond ASAP.

Jerry, I can tell you're in good hands with Markus. He seems to know his stuff and quite frankly seems to have a better resource library than I do and I can't wait to read some of the links he's shared. He's someone's brain who I'd love to pick one day. smile.gif

Amir, thanks for your participation in this thread. Your knowledge is invaluable and I hope to have the honor of meeting you one day. I do not disagree with anything you've stated, however, I try not to advise people on specifically what to do or make blanket statements (not stating you did this either, just to be clear...just summing up) as to whether certain reflections are good or bad...Just that they can be good or bad... I think that's what you're saying too, right?

It's hard for people not familiar (and who haven't heard and/or spent a lot of time in) with rooms of different acoustical models to choose what they want for their own rooms because they have no point of reference. Therefore, I tend to recommend an easy to achieve (comparatively speaking, of course) mix/model somewhat approaching a LEDE room possibly with no termination of the ISG so folks can get a feel for this and see if they like the results. Then, as things are learned and experienced over time, the model can change along with growth of knowledge without having to tear down walls and start all over.

Thanks again to all who have participated and keep on!

--Jason

PS Jerry, if you've asked me any specific questions I've not yet responded to or if you'd like my opinion on something in addition to the information you've already received, feel free to ask. There's already a ton of good info here and it's possible I've inadvertently missed something.

Thanks to EVERYONE that Helps Make These Threads so Awesome!

jevansoh is offline  
post #28 of 162 Old 10-08-2012, 01:15 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 4,716
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 302 Post(s)
Liked: 171
Quote:
Originally Posted by amirm View Post

Fortunately every argument and then some was made recently and is available to read through. Likely will take you many hours if not days to read and you need high ability to get past bickering smile.gif. If you can do that, here it is: http://www.avsforum.com/t/1413173/does-sound-sounds-better-in-a-room-full-of-furniture-and-stuff-or-without

To save time just read everything by dragonfyr.

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is online now  
post #29 of 162 Old 10-08-2012, 07:44 AM
AVS Special Member
 
jim19611961's Avatar
 
Join Date: Mar 2011
Location: Texas
Posts: 1,246
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 79
Quote:
Originally Posted by jevansoh View Post


.. While it is fun to discuss this complex topic of acoustics, theories, analysis, etc, including whether or not you should be attenuating the specular reflections to begin with you must first set a goal as to what you'd like to achieve.
My philosophy is that you should CHOOSE a model such as LEDE/RFZ, Ambechoic, based on specific criteria and what you wish to accomplish in your room. If you do not know which model you wish to attempt to achieve, then I'd recommend simply absorbing early reflections up to about 20ms and any high spikes after 20ms then look at FR and Waterfall and if everything is in line, listen this way for awhile. If you like it, great! You can always learn more and try other models. It's fun to experiment. I wouldn't worry about terminating the ISD gap in this case until you know specifically what your goals are.

I agree that choosing a model is a good idea.
Quote:
Originally Posted by jevansoh View Post

As to the gentleman who posted his ETC and mentioned the one remaining spike/termination of ISD... This is a great effort and you should be proud. I'm sure it took a lot of work/effort to get where you are. Please don't confuse one loan specular reflection that happens to be past 20ms on a graph a true termination of an ISD Gap. A true termination is no less than 12db down plus is followed by a diffuse return of decaying energy. It appears the energy on both sides of your loan spike is no more or less diffuse and is at roughly the same level. I would consider this more of a good start or if you've not chosen a LEDE/RFZ model I'd consider attenuating this loan spike, too. Again, great effort, but it's not quite finished.

I dont want to detract from the thread to focus too much on my own plan, but furthering what I have done may help the OP, so I will make a few comments.

Its true that the LEDE / RFZ calls for several criteria, as I understand them, they are these:

1) A ISD gap of 12-25ms, preferably in the 20-25ms range.
2) -20db down or better within that gap
3) A termination peak of not less than -12db and arriving from a lateral direction (100-120 degrees)
4) a following diffuse sound field with a smooth exponential decay

Its true , I have not accomplished (and cant see how) 4), but have 1) 2) and 3).

So what becomes the point of discussion is what is the function and consequence of each parameter.

1) Basically allows you to hear the recording (direct sound) without the interference of also hearing your room within the gap length
2) This is the standard figure quoted to be sure to stay below the Haas threshold
3) The termination masks latter room reflections and establishes a psycho acoustic room boundary different than your actual room boundaries. This delay return also gives a sense of liveness to an otherwise fairly dead space established by the ISD gap.
4) Gives a added sense of spaciousness to the listening experience adds to the sense of being in a larger room

I realize I am over simplifying this a bit, but this is my understanding of the function for each of the parameters.

In my listening experience, having only 1) and 2) alone makes for a fairly dead listening experience. But you hear the recording very nicely and the soundstage is very tight and defined.

Adding 3) gives a sense of liveness and makes it less dead sounding. Without 4) I admit that the sense of spaciousness and/or large room boundary is not well established. But I do find having 3) without 4) sounds better to me than not having either.

As far as my room being unfinished, in reference to accomplishing the LEDE / RFZ in its entirety, yes, your right. But for me, I am finished. I cannot create a diffused sound field AND have a long termination gap in a room thats 15.5 feet wide. I just cannot make it work.

My Room
http://www.gearslutz.com/board/studio-building-acoustics/817205-my-listening-room.html

My Music
http://rateyourmusic.com/~jim1961

My Equipment

Rega - Apollo
Rega - DAC
Goldpoint Passive
(2) Classe CA-100 bridged power amps (350w)
Jenzen Next ( http://www.troelsgravesen.dk/Jenzen-NEXT.htm )
...
jim19611961 is offline  
post #30 of 162 Old 10-08-2012, 08:13 AM - Thread Starter
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 6,815
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 229 Post(s)
Liked: 646
Quote:
Originally Posted by jim19611961 View Post



I realize I am over simplifying this a bit, but this is my understanding of the function for each of the parameters.
 

 

Unfortunately, some of the discussions in this thread have evolved into theories that are not simple enough for me to understand, so no danger of oversimplifying.  While I appreciate the discussions and the advice, I am reminded of how much of a novice I am in this area.  However, I am an eager learner!

 

My next simple step will be to follow Markus' and Jevansoh's advice, i.e. to temporarily place treatments in several spots, re-run the ETC graphs, and use trial-and-error to improve things.  I'll report my progress...

AustinJerry is offline  
Reply Audio theory, Setup and Chat

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off