Full range measurements & averaging

**hifisponge** · **11-26-08, 01:32 PM**

Couple of quick questions from a new user (me).

I plan on using the DBX mic for full 20Hz to 20KHz measurements, and I'd like to know if the mic is to be used with the tip pointed to the ceiling or at the speaker.

I'd also like to do some averaging of the responses taken in a 3 foot radius around the prime seat. I'm wondering how this works. I know that REW will do the averaging for you, but once you move the mic to a new position wont the level change? For instance, if I take measurement A at 12 feet from the speaker and measurement B at 15 feet, I assume that the overall level will drop by a couple of dB at the 15 foot mark as compared to position A. Are the overall level differences accounted for by REW when it calcs the avg? Or do I have to adjust the level to 75dB each time I move the mic? Am I over-complicating things?

Also, does REW allow you to assign a weight to each of the measurement points when averaging, so I can give more weight to the measurement at the sweet spot than I do the others?

Thanks in advance!

- Tim

**brucek** · **11-26-08, 06:19 PM**

Quote:

I'd like to know if the mic is to be used with the tip pointed to the ceiling or at the speaker

Pointed at the ceiling. It's an omni-directional mic.

Quote:

wondering how this works. I know that REW will do the averaging for you, but once you move the mic to a new position wont the level change?

Yes, reset the levels to 75dBSPL and measure.. Touch them up before taking the average if you like.

Quote:

Are the overall level differences accounted for by REW when it calcs the avg? Or do I have to adjust the level to 75dB each time I move the mic?

REW simply takes any average of the responses you select. If one response is measured high, the average will be higher.

Quote:

does REW allow you to assign a weight to each of the measurement points when averaging, so I can give more weight to the measurement at the sweet spot than I do the others?

You can certainly use the Trace Offset adjustment to any of the responses, so that it changes the average. Remember to lock in the change with the Add offset to Data feature or the change won't affect the average.

brucek

**Wayne A. Pflughaupt** · **11-26-08, 06:37 PM**

Quote:

I'd also like to do some averaging of the responses taken in a 3 foot radius around the prime seat. I'm wondering how this works. I know that REW will do the averaging for you, but once you move the mic to a new position wont the level change? For instance, if I take measurement A at 12 feet from the speaker and measurement B at 15 feet, I assume that the overall level will drop by a couple of dB at the 15 foot mark as compared to position A.

I doubt 3 ft. is enough difference to register on the SPL meter. Certainly can't hurt to run the level-checking excercise again, but you can certainly get a valid measurement without doing it.

Regards,
Wayne

**lcaillo** · **11-26-08, 07:30 PM**

Quote:

Wayne A. Pflughaupt wrote:

I doubt 3 ft. is enough difference to register on the SPL meter. Certainly can't hurt to run the level-checking excercise again, but you can certainly get a valid measurement without doing it.

Regards,
Wayne

Whoa, Wayne! You might want to experiment. It depends a lot on the room, but a few inches would be enough in many rooms to get very different measurements. Maybe you have a large sweet spot, but that is atypical in my experience. Many of the rooms that I have measured have been very critical at the listening position. A few inches can make a big difference in measurements in some rooms.

**hifisponge** · **11-26-08, 07:52 PM**

Quote:

brucek wrote:

You can certainly use the Trace Offset adjustment to any of the responses, so that it changes the average. Remember to lock in the change with the Add offset to Data feature or the change won't affect the average.

brucek

Bruce -

Thanks for the response. Forgive my ignorance here, but are you saying that if I raise one trace higher in level than the others that the average will favor the higher level response?

Alternatively, I was thinking of taking 3 measurements in the sweet spot and one to each side, then averaging the five. This would give more weight to the measurements taken at the prime spot. Does your solution do the same thing?

- Tim

**hifisponge** · **11-26-08, 07:56 PM**

Quote:

lcaillo wrote:

Whoa, Wayne! You might want to experiment. It depends a lot on the room, but a few inches would be enough in many rooms to get very different measurements. Maybe you have a large sweet spot, but that is atypical in my experience. Many of the rooms that I have measured have been very critical at the listening position. A few inches can make a big difference in measurements in some rooms.

I believe that Wayne is saying that the overall SPL level wont change that much, not the FR.

My whole point of measuring in a radius around the prime seat and averaging is to minimize the affects that small mic movements can make on the FR. I want to see the common high and low spots within a listening window.

**lcaillo** · **11-26-08, 08:01 PM**

Quote:

hifisponge wrote:

I believe that Wayne is saying that the overall SPL level wont change that much, not the FR.

My whole point of measuring in a radius around the prime seat and averaging is to minimize the affects that small mic movements can make on the FR. I want to see the common high and low spots within a listening window.

...and depending on the environment and frequency they may vary considerably. That was my point based on my experience. When you determine FR you are measuring levels over various frequencies. It can be very dependent on position in some rooms.

**hifisponge** · **11-26-08, 08:17 PM**

Quote:

lcaillo wrote:

...and depending on the environment and frequency they may vary considerably. That was my point based on my experience. When you determine FR you are measuring levels over various frequencies. It can be very dependent on position in some rooms.

Yes, I could see that happening too. I probably recheck levels with each mic movement just to be safe.

**lcaillo** · **11-26-08, 08:20 PM**

Now you have the idea. Until you know that your location is not critcal, don't assume otherwise.

**Wayne A. Pflughaupt** · **11-26-08, 08:37 PM**

I can see your point, Leonard. I was (probably narrow-mildly) envisioning my own large family room set up. In small rooms, especially shoe-box or square ones, bass levels can vary noticeably at even a foot or so. And bass is what the OP is measuring.

Sorry...

Regards,
Wayne

**lcaillo** · **11-26-08, 08:47 PM**

We all tend to let our personal experience factor in to our comments. Unfortunately, I have a rather odd room that has some severe variations in response...the good news is that I am the only one who cares and I know where to sit!

**hifisponge** · **11-27-08, 01:46 AM**

Off the topic of my original post, but related to using REW, I want to make sure that I'm going about measuring the five main speakers in my HT the right way. I'm imagining that when I get to measuring the center and surrounds that I'm going to want to connect each of these speakers to one of the main left/right amp channels temporarily (and the REW output to the corresponding input on the processor). I would use the MCH inputs but these offer no bass management on my processor. Any problems with this approach, any alternatives?

**thewire** · **11-27-08, 01:50 AM**

Quote:

hifisponge wrote:

Off the topic of my original post, but related to using REW, I want to make sure that I'm going about measuring the five main speakers in my HT the right way. I'm imagining that when I get to measuring the center and surrounds that I'm going to want to connect each of these speakers to one of the main left/right amp channels temporarily (and the REW output to the corresponding input on the processor). I would use the MCH inputs but these offer no bass management on my processor. Any problems with this approach, any alternatives?

Play Pink PN on your CD player. Simply go to the Generator, select Pink PN, set the output level (example -33dB FS), save the file, burn it to disk. Then put the disk in the player and hookup the speaker or speakers so it is playing. Now go into REW and go to the RTA and use that. You could also use the Spectum Anylizer with White PN.

**hifisponge** · **11-27-08, 03:00 AM**

Steven -

What are the pros and cons of using the RTA or SA functions of REW as opposed to the chirp type signal? Are you suggesting the use of the RTA only for the mid and treble response? Because it doesn't seem like the RTA would have the resolution needed to correct the bass response of the main speakers.

- Tim

**thewire** · **11-27-08, 04:00 AM**

Quote:

hifisponge wrote:

Steven -

What are the pros and cons of using the RTA or SA functions of REW as opposed to the chirp type signal? Are you suggesting the use of the RTA only for the mid and treble response? Because it doesn't seem like the RTA would have the resolution needed to correct the bass response of the main speakers.

- Tim

You don't neccessarily need to use the RTA or Spectrum Anylizer. I think you could also use the measurements of the surrounds as you would the front speakers. To do that all you would need to do is swap the setting of the left and right speakers with the surrounds, then switch the wire connections. For the Spectrum Anylizer and RTA I use two averaging. I then select 1/48 octive (high resolution) for the bass or I use 1/6 smoothing applied to a saved measurement to look at the mains & subwoofer response together. It would be faster to use the RTA or Spectum Anylizer as you could make adjustments in Real-Time. It is up to you which way works best. 1/3 octive is good for the surrounds or mains, and 1/6 is slightly more higher in resolution. If you want to you can also use 1/24 octive for the mains or surrounds which is best.

**hifisponge** · **11-27-08, 04:28 AM**

^^^ Interesting. I hadn't even considered doing the measurements and filter adjustments in real time. I will have to play around with the RTA and SA. What do you mean by "I use two averaging"?

**thewire** · **11-27-08, 04:35 AM**

Quote:

hifisponge wrote:

^^^ Interesting. I hadn't even considered doing the measurements and filter adjustments in real time. I will have to play around with the RTA and SA. What do you mean by "I use two averaging"?

The averaging makes the signal on the screen slow down more so you can see it better instead of jumping all over very fast. I like the 2 averaging. The RTA measurements you will not be able to generate waterfalls for so if you need that use the chirps.

**hifisponge** · **11-27-08, 04:53 AM**

Quote:

thewire wrote:

The averaging makes the signal on the screen slow down more so you can see it better instead of jumping all over very fast. I like the 2 averaging. The RTA measurements you will not be able to generate waterfalls for so if you need that use the chirps.

Ahh, I get it. It's been a while since I've used an RTA and now that you mention it, the readout does jump around a lot. Good to know that REW allows you to slow things down.

I'll probably use the chirps just on the sub channel so I can set some filters to address resonances.

Thanks for the tips.

**brucek** · **11-27-08, 12:53 PM**

Quote:

Forgive my ignorance here, but are you saying that if I raise one trace higher in level than the others that the average will favor the higher level response?

REW averages the responses that are selected in the checkboxes. It's that simple. If you raise the wholesale level of a response using the Trace Offset feature and then select Add Offset to Data, then the average will be taken of the newly placed response against the other(s) selected.

Quote:

I'll probably use the chirps just on the sub channel so I can set some filters to address resonances.

No, for RTA you use the REW generators Pink Periodic Noise (Pink PN). For Spectrum Analyser use the REW generators White Periodic Noise (White PN). See here for all the settings to use for Spectrum Analyser and RTA feature.

brucek

**hifisponge** · **11-27-08, 04:50 PM**

Quote:

brucek wrote:

No, for RTA you use the REW generators Pink Periodic Noise (Pink PN). For Spectrum Analyser use the REW generators White Periodic Noise (White PN). See here for all the settings to use for Spectrum Analyser and RTA feature.

brucek

Bruce -

Sorry for the confusion. I wasn't saying that I would use the chirp tones with the RTA, I was saying that for the sub measurements I would use the static measure feature and then switch to RTA for the main channels. But thanks for that link. The RTA feature sure seems like it could be a useful tool, especially for sub integration.

Thanks again for your help!

- Tim

**bjs** · **11-27-08, 10:57 PM**

Quote:

hifisponge wrote:

Couple of quick questions from a new user (me).

I plan on using the DBX mic for full 20Hz to 20KHz measurements, and I'd like to know if the mic is to be used with the tip pointed to the ceiling or at the speaker.

You said up to 20kHz...so the orientation will make a noticeable difference at the high end.

Pointing at the ceiling reduces the response peaking but limits the high frequency response compared to pointing at the speaker. Whether it matters depends on what you are trying to do by measuring full range to 20kHz.

**hifisponge** · **11-28-08, 04:20 AM**

Quote:

bjs wrote:

You said up to 20kHz...so the orientation will make a noticeable difference at the high end.

Pointing at the ceiling reduces the response peaking but limits the high frequency response compared to pointing at the speaker. Whether it matters depends on what you are trying to do by measuring full range to 20kHz.

Exactly my reason for asking. Even though the DBX mic is omni directional, I know these mics don't have perfect polar response in the HF.

So are you saying that the reading will show more treble roll-off than there actually is if I point the mic up? I just want to know how to orient the mic so that the reading is accurate up to 20KHz.

Brucek - do you still recommend that I point the mic at the ceiling?

**brucek** · **11-28-08, 08:47 AM**

Quote:

do you still recommend that I point the mic at the ceiling?

For testing a home theater it's absolutely fine.

If you're doing speaker design, then you should get the ECM calibrated professionally and then consider all the factors involved in audio HF measurements.....baffle step etc..

brucek

**Wayne A. Pflughaupt** · **11-28-08, 09:18 AM**

Quote:

hifisponge wrote:

Exactly my reason for asking. Even though the DBX mic is omni directional, I know these mics don't have perfect polar response in the HF.

So are you saying that the reading will show more treble roll-off than there actually is if I point the mic up? I just want to know how to orient the mic so that the reading is accurate up to 20KHz.

Brucek - do you still recommend that I point the mic at the ceiling?

Why not take measurements both ways and see what happens? If you do get a high-freq roll-off with the mic pointed at the ceiling, it'll show up on the graph.

Regards,
Wayne

**hifisponge** · **11-28-08, 02:32 PM**

Quote:

Wayne A. Pflughaupt wrote:

Why not take measurements both ways and see what happens? If you do get a high-freq roll-off with the mic pointed at the ceiling, it'll show up on the graph.

Regards,
Wayne

Wayne -

At my 12 foot listening position I expect to see roll off in the treble reading above 10KHz whether the mic is horizontal or vertical, but I need to know how much of that is the mic's response contributing to the reading.

The mic is spec'd for flat(ish) response on one axis. I'm trying to find out which that is.

Ah well, most measurement mic's are meant to be used in the vertical position, so I guess I'll just assume that the DBX RTA-M is too.

- Tim

Advertisement