Home Theater Forum and Systems banner

Help with ETC interpretion

7K views 21 replies 5 participants last post by  EarlK 
#1 · (Edited)
Hi,
I am using ETC to help me locate and treat some reflections in my room. Hoping someone can help me understand how to interpret the measurement so I can determine how many feet the reflecting surface is from the SPL meter, in order to determine treatment location.

Following is a pic of my front left speaker's ETC response. I used the "use loopback as timing method" for this measurement, with both "sub sample timing adjustment" and "decimate IR" unchecked as I was told this gives you the actual TOF of the reflection. Problem is, I just don't know how to read this plot.

My hardware consists of UCA202 and RS SPL meter. I have the meter going to Rin on the 202, with Rout feeding my AVR. I added an RCA cable for loopback from Lout to Lin on the 202. Is that correct? Couldn't think of any other way the loopback could/should be connected.

Is T1 the loopback signal? If so, that doesn't make sense as the initial peak (T2) indicates arrival of the direct signal at 25ms, which corresponds to around 28ft. That can't be correct as I'm measuring around 12'-13' from the speaker.



Thanks in advance for any clarification offered.
Floyd
 
See less See more
1
#5 ·
SAC, attached is the new .mdat file. I just made two new runs and verified that the loopback settings were correct in REW.
My speakers (both L and R) are no more than 12'6" from the SPL meter, both sitting just inside the right and left edges of my AT screen. Following are a couple of pics for reference. Interestingly enough, both sweeps exhibit a rather large spike in ETC approx 2.4ms after the original signal (if indeed that is original signal). I have been under the impression that spike was due to reflections off of my wood columns, which are close to 50% the distance from speaker to listening position.

Today, I took a bunch of measurements with what I can only interpret as the "blocking method" using a roll of linacoustic (about 50% leftover). Interestingly enough, I couldn't find any location while placing that roll (48" high) in any lateral positions to attenuate the largest spike. I then took a folded up piece of R-19 and measured while holding it to sides, front etc. right around the SPL meter. I was able to get a nice attenuation of that spike only when holding it in front/higher elevation of the meter. I am coming to the conclusion that this common spike between L and R speakers may very well be coming from the ceiling (which is untreated).

Anyway, I was hoping to be able to understand how I can interpret the ETC and use the string method to identify points that this particular reflection (and subsequently others) incident boundary surface is. Just not sure about how to interpret the TOF based on loopback and translate that appropriately.





Thanks,
Floyd
 

Attachments

#6 ·
T1 is indeed the loopback timing reference.

If the test signal is routed through a digital AVR however, or there are any other digital processors in the signal path to the speakers, there is additional processing delay introduced. Most AVRs have a "direct" mode (analog through) that allows the signal to remain in the analog mode. Use that mode to eliminate the extra delay.

The multichannel input/output path often stays in the analog realm also I think, so that may be a good option.

If you don't have an "analog through" setting on the AVR you could connect the loopback connection from the Left output of the AVR back to the UCA 202 left input and that will add the same extra delay to the loopback signal also. The numbers should then work out IF both the right an left speaker distances in the AVR are the same.

[I would expect that most of the extra delay is because of speaker distance settings in the AVR that adds delay to match the arrival times of the various speakers. These times are calculated in a relative manner, not on an absolute basis so if you set all speakers to the same distance there will probably be no distance delays added to any of the channels. The digital conversion delay will still exist however if you don't have the analog through mode .]
 
#7 · (Edited)
John, I have an Onkyo TX-NR708 AVR, with FR/FL pre-outs feeding a Rotel 2 channel amp then into my FR/FL speakers. I only have Audyssey enabled (no dynamic EQ or volume) and have the AVR in 2.1 stereo listening mode when taking measurements with the sub physically turned off. The UCA202 connects into the AVR via front panel AUX input, with connection to L or R input depending on speaker being measured. Indeed, I agree there is signal processing delay being introduced which is indicated in the info panel as "System Delay": 29ms.

I guess I thought this would be a bit more intuitive than what you're proposing, in connecting the loopback path out of the AVR? If the system delay is measurable as it appears to be, seems to me it's just a matter of interpreting the resulting ETC timing correctly.

So if T1 is the test signal pulse, then the measured delay through my signal chain is approx 29ms. Direct speaker signal is arriving at T2 to the mic, and reflections are arriving at T2 plus whatever the diff is between T2 and T3.

If that's correct, let's continue to analyze based on the following. Let's assume I have 12 feet from mic to speaker. If my T2-T3 delta is 2.54ms, or approx 2'6", then total delay from that reflection would be 14'6" (12' for direct + 2'6" for the additional distance traveled by the reflected signal) correct?
So, I should be looking for a reflecting surface around 7'6" from the microphone (half the travel time of that signal).

AND if all that's correct, I'm not sure what value the actual TOF is when measuring with loop back (for the analysis I am doing). The measured delay (T2-T3) on sweeps I took NOT using loop back are the same 2.54ms as the loop back method. The only diff I can see is that for non loop back, direct/peak energy spike is referenced at "0" time on X-axis, while in loop back "0" time corresponds to the test signal, while shifting everything else (initial direct energy and reflections) out on the axis.

If one knows their speaker distance to mic, measure the difference from initial signal to reflected signals and do the math, what does it matter if you use loopback or non loop back?

Maybe I'm just being dense, but I don't get it.
 
#16 · (Edited)
.......

AND if all that's correct, I'm not sure what value the actual TOF is when measuring with loop back (for the analysis I am doing). The measured delay (T2-T3) on sweeps I took NOT using loop back are the same 2.54ms as the loop back method. The only diff I can see is that for non loop back, direct/peak energy spike is referenced at "0" time on X-axis, while in loop back "0" time corresponds to the test signal, while shifting everything else (initial direct energy and reflections) out on the axis.

If one knows their speaker distance to mic, measure the difference from initial signal to reflected signals and do the math, what does it matter if you use loopback or non loop back?

Maybe I'm just being dense, but I don't get it.
Why are you worried about what the loopback value is? What color are the curtains in your kitchen? (There are ways to determine this, but its not germane to the issue at hand.)

If this is enabled and you have no additional processing enabled in the AVR, the time that the signal takes to travel from source to mic should be the Time of Flight. Period.
That's all you need. From this you can determine the time of flight converted to distance by multiplying the TOF by 1.13foot/ms. The time differentials between direct and indirect signals are of no practical value for this application. ...Maybe when you start setting delay lines...

What is so complex about this?

But let's take this two steps further. You imagine that you "know" the acoustic center of the speaker. Really? This is NOT always true! Is your speaker a signal aligned full range single driver speaker? Is it a signal aligned coax driver such as in a Bag End coaxial? Probably not. Taken top another practical extreme often encountered in pro-sound, what if the source is a multi-element array?

The acoustic origin of the speaker - the point at which the sound 'appears' to emanate from the speaker is NOT the tweeter (as so many want to assume) or necessarily the baffle. Confuse this a bit more by including a horn, where the acoustical origin is at some amorphous spot somewhere inside the throat, determinable only by measurement...And these are 'well-behaved' examples! We haven't even mentioned astigmatism in CD horns where the acoustic center moves with frequency, or the fact that many drivers literally offset at different frequency ranges and move forward to a different position only to move back to the normal excursion range in other frequency ranges (as Don Keele demonstrated at the Atlanta '91 Loudspeaker Design Seminar with the woofer removed from his reference B&Ws!).

If only we could simply 'define' the acoustic center of a speaker simply by looking at it!

The point is, if you already know the distance (by intuition), why bother? Because you are going to use the MEASUREMENTS to determine the actual vector paths of travel and points of incidence. And these times, and hence derived distances of travel, are referenced to the MEASURED values - the actual acoustic center - not to your intuited acoustic center. You wanted accurate...right?

And the use of the string is a basic conceptual device. It is not the recommended procedure to use as a rule. It provides a concrete demonstration and visual reinforcement of the concept!

Once this is accomplished, I would recommend that you use the blocking technique...unless you simply enjoy monopolizing other folks time and taking several days to resolve the individual reflections paths and points of incidence. In this way the process, once understood based on a good understanding of the concept, can be relatively quickly performed.

And I will go still further...if you are going to do this with any regularity, I find it difficult to imagine not using the Polar ETC and a laser pointer and letting a PolarETC program calculate the 3space coordinates and simply dialing these coordinates into a transit with the mic replaced by a laser pointer and letting the laser pointer point to the incident spot on the wall as fast as you can move the cursor to the reflection and adjust the transit coordinates!

This isn't intended to be a character building exercise. Its a necessary bit of busy work used to identify points of incidence in order to then facilitate treatment options in accordance to one's desired target acoustical response model. In other words, the first few times someone does this its exciting and fascinating. Thereafter its work...and some would rather focus on moving the goal forward and not simply standing in admiration of the hammer they use to complete various tasks - if you catch the analogy.

Thus, to regress a bit, the focus is NOT on the string. Its not on guessing the acoustical center of the source. (And what if the source you were measuring was flown 15 feet over your head? Are you into repeatedly climbing up there on a ladder or scissors jack with your string? And then there are the other points of incidence...ceilings come to mind...what fun.....) the focus is improperly configuring the test platform so that these issues become trivial and one less issue about which to worry - allowing you to focus on the larger issues at hand.

So, I have taken you a bit further than where your current focus lies, but i hope it puts this process into a bit better perspective. We use the loopback correction as we are relying on the measurement to provide accurate values. Not because we want to add a bunch of approximate estimations into what is otherwise an extremely accurate process. And such errors, while perhaps not critical in a very short time frame and corresponding distance, become increasingly critical in longer times and distances, where the identification of incident points could easily be off by a few feet - significant enough to effect the treatment and response. And also significant enough to require what could be a pretty extensive period of re-measurement, re-placement of treatment and re-verification.

For some this time is money. For all its an avoidable pain in the backside.

The point here is not to see how many goofy variables we can introduce and explain (at least not from my perspective...and if it becomes this, there are quite a few neater 'anomalies' to explore...). The purpose is to properly configure the platform, make reliable measurements that avoid the amazing serendipitous journey of exploration, and to cut to the chase and determine the necessary information, namely the effective point of boundary incidence and to move on toe determining what treatment is most effective in obtaining the overall response that is desired.

So try not to get lost in the weeds and try to remain focused on the larger goal. Once that is understood there will be plenty of time to explore and to better appreciate the various associated minutia and variables that one can 'construct' that can be explored.
 
#8 ·
Here's a pic I created using your mdat file .

As you can see , I think this ETC spike is created by a reflective surface very near your speakers .

> I think what I would do is cut a piece of string 14.6" long , firmly attach one end to the speaker & the other end firmly to the measuring position .
> Then pull out the slackness of the string towards a point on the wall or roof ( until all the slack gets taken up & the angles look good ) . This is the area I would acoustically treat .

> It'll be interesting to hear what SAC says is the best approach/interpretation . In the meantime I need to go back to my copy of "Sound System Engineering" to get some pointers from the grand-dad of ETC .

<> :sn:
 

Attachments

#9 ·
Floyd,
I was only addressing why the loopback to impulse arrival time (T1 to T2) did not correspond to the actual distance measured between the speaker and the LP. If you eliminate the AVR delay in one of the ways suggested, it should then agree with your measurement.

I am no expert on the overall general question, but would expect that EarlK's suggestion would work just fine. That is, I don't see why one needs to measure direct distance using the time from T1 to T2 when you can just measure it easily. That method probably has more utility in large venue where the distance is not as easily determined. I am just guessing about that though.

In summary, I also would expect the 14' 6" string that EarlK suggested to work well enough to locate the reflection point.
 
#12 ·
LOL...I'm such a dolt sometimes. I didn't even think to disable Audyssey as I was previously doing some measurements to determine response after it was run, and quite frankly it's been a while.

For subsequent measurements, would you include my Rotel amp in the chain, or drive my fronts directly from the AVR?

Interestingly enough, I did end up doing just what you suggest regarding the "blocking" method. I stapled a doubled up piece of R-19 about 12" square to a stick that I could position around what I thought could be the path of the reflection. I was suspect that possibly the reflection was coming off the wood bullnose trim on my stage, so the linacoutic roll was placed there just to rule it out (which it did as there was no effect).

I also ended up doing the string trick based on the previous measurements. After measuring my actual speaker distance to mic at 12' and adding is the delay from T2-T3, total length was around 14'8". Guess where that touched something....right on the ceiling between speaker and mic. I marked it and then held the insulation with the stick while I remeasured. Bingo! That single piece of insulation brought that reflection down to around -10dB! Pic below for measurement with time marked for ref:




Now, as to that nasty direct signal, I am using American Acoustic tower speakers circa 1988 or there abouts:eek: Probably safe to say their construction and technology used at the time is a major contributor there. I've spent so much on the theater build, I decided to use these until I could afford an upgrade.

Guess what's on Floyd's list to Santa this year???? Matter of fact, I am planning on getting new bookshelf's in prior to the time I have off in Dec. I have Axiom QS8's for surround and VP150 for center currently. Would take any suggestions on what to consider for front mains if anyone has an opinion on that.

So, will probably put further efforts on hold until I get my upgrade. I've been a little suspicious of the current speakers since day one.
 
#18 ·
Guess what's on Floyd's list to Santa this year???? Matter of fact, I am planning on getting new bookshelf's in prior to the time I have off in Dec. I have Axiom QS8's for surround and VP150 for center currently. Would take any suggestions on what to consider for front mains if anyone has an opinion on that.
I'm sure you could get plenty of suggestions if you ask in the proper area of the forum. I'd guess that many responses will be motivated by brand loyalty though. One of the thoughts I've had is that if I ever wind up looking for a new array again, I may well look at self-powered studio monitors, ideally using a processor with balanced outputs. I've honestly no idea where one might go to get good quality speaker reviews as I've never seen one that included blind listening against the reference in any fashion. If I don't go with the studio monitor idea, I'd probably DIY something, which I might actually do anyway. Even if I don't manage to educate myself on all of the ins and outs of speakers, there are plenty of published DIY designs, or close enough to DIY (e.g. the stuff from Fitzmaurice).
 
#21 ·
Little late updating this.....I did get rid of the long system delay by putting my AVR in "Direct" mode instead of "Stereo" and proceeded to measure and track down some reflections.

As an experiment, I decided to track down the largest reflection in my ETC measurement of my front right speaker, which was measured at 2.22ms or 2.5 feet after the initial direct speaker signal. The direct speaker signal is normalized at T= 0 on the x axis, with the spike measured at the 2.22 ms later as shown following (note, this was a previous effort NOT using loopback with T=0 at impulse peak):



Next I needed to do some simple calculations in order to use the "string method" to find out specific areas to check where the actual reflection surface is. First I measured the distance from my Radio Shack SPL meter to the actual acoustic center of the speaker being measured. That distance was 12' (as best as I could measure). The reflection being measured takes an additional 2.5' to arrive, which means that the total distance of the reflected signal is 12' + 2.5' or 14' 6".

The concept of using the string method is that you take a length of string which is equal to the distance traveled of the signal (in this case 14'6"), tie off one end to your measurement mic, and the other end to the acoustic center of your speaker. If you have helpers available, they can hold either end. In my case I didn't so I had to gin up tie offs. You then take up (I just slid a finger along the length) the string slack between the mic and speaker extending it in all axis's and note what boundary surfaces it touches.
In my case, it touched the ceiling about 1/2 way between the mic and speaker. Following are a few pic for clarification.

Mic at measurement position with one end of string tied off:



Other end of string tied off at speaker:



Location on ceiling where string touched:



Visual of overall string path with my "treatment stick" holding it up:



I next took another measurement while holding that insulation batt to the ceiling. Following is that measurement with time reference noted. Notice that the spike has been attenuated. Original measurement shown directly following again for ease of comparison.





I have the same spike occurring on the left speaker at similar time, so appears that a ceiling panel is in my future.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top