ETC Graphs

[Bob_99]

Ethan or Bryan,

When looking at ETC data, what frequency range gives the best picture of what's happening (500Hz to 5000Hz)? Also, is it better to break it down into smaller ranges (1000Hz-2000Hz,2000Hz-3000Hz,etc.)?

Thank you.

Bob

[Ethan Winer]

Bob_99 wrote:

When looking at ETC data, what frequency range gives the best picture of what's happening (500Hz to 5000Hz)? Also, is it better to break it down into smaller ranges (1000Hz-2000Hz,2000Hz-3000Hz,etc.)?

What's the goal? I assume you're looking for problems caused by early reflections, and that's mostly in the midrange. Not sure if there's an advantage to breaking it down into small ranges. But I admit I haven't spent a lot of time with that feature in ETF.

--Ethan

[Bob_99]

Ethan,

I am looking for problems with early reflections but also just trying to understand more about acoustics and was wondering if this could be useful to determine how to treat first reflections. Here is what started me thinking about it:





You can see that the reflection at about 2 ms is more pronounce when using a narrower bandwidth. I'm not sure if that's real or just a result of the method used in the software calculations. If it's real, would it be reasonable to say that I need something thicker at the first reflection point?

Thanks again.

Bob

[Bob_99]

Ethan,

Just curious if you have any input regarding my last post.

Thanks.

Bob

[Ethan Winer]

Bob_99 wrote:

Just curious if you have any input regarding my last post.

LOL, I was pretty busy all week.

The first reflection that hits in the top graph is only a few dB below the primary sound, so that's a problem for sure. That it's 2 ms later means a path difference of about 2 feet. So that could be from a side wall if the speakers are close to the walls, or the floor, or both. I generally look at the full range to see it all in one graph. Regardless, if you treat the side walls and ceiling, and optionally the floor, at the usual places, those reflections should drop to less than 15 dB.

--Ethan

[Bob_99]

Thanks Ethan. I know how busy you are and I really hesitated to nudge you for an answer. I appreciate the time you put into the forum.

Bob

[Doug Plumb]

The ETC graph can be tricky - you are always trading off frequency resolution for time resolution and this is the heisenberg principle and cannot be escaped.

If you want lots of detail in time then you must use a wide frequency range. I like to look at individual driver radiation and have about 1 ms of resolution so for a two way speaker that has a crossover of 2.5 KHz I will choose a band of 1000 Hz - 2000 Hz for the mid/woofer and about 3000 - 4000 Hz for the tweeter. Notice that the bandwidth is the same in both cases = 1000 Hz. This gives me 1/1000 sec or 1 ms of resolution.

If I want a detailed view of cabinet diffraction I know that most of it comes from the tweeter and I can use a wide band on the tweeter so a band of 3000 - 8000 Hz gives a resolution of 1 / (8000-3000) = 1/5000 = 0.2 ms.

To accurately quantify a single reflection you need to use a wide enough band to isolate it from the others. The 1 ms band used in the above top graph isn't enough to really isolate that single 2 ms reflection. Try using a 5 KHz band in the tweeters operating range then you will see it clearly. Try 5 KHz - 10 KHz.

RPlusD has a great manual (so people are telling me) and you can download and read it for free. The methods used in this manual are more evolved in terms of methods than most. Essentially RPlusD was created to apply better measurement techniques. It was a redesign of ETF from the bottom up. All ETF users get RPlusD free of charge (for ETF purchase after Sept 1 2000).

The manual can be used as a guide for most audio measurement software in use today.

[Doug Plumb]

The second graph posted above does isolate that reflection and give a true indication of its actual level. The resolution is 1/(20,000-15)= Very Small. So its a detailed look at time domain behaviour.

I get a spelling error for "behaviour" and wonder why you Americans keep spelling everything wrong 8-)

[Bob_99]

Doug,

It's great to see you posting on this forum and thank you for your response. Your answer is the information that I was looking for.

RPlusD has a great manual

Yes, it's excellent and I have read it several times always gleaning new information from it each time.

wonder why you Americans keep spelling everything wrong

If you ever watched American TV you would not wonder why Americans spell everything wrong. Especially when you get the idiot who thinks 'relief' is spelled "R-O-L-A-I-D-S".

Bob