Comments on my first measurements?

[weverb]

elee532 wrote:

As suggested earlier, the first graph shows readings of each speaker. Red is the sub and mains together, green is the left front speaker, and gold is the right front speaker. Does this offer any insight into the dip around 169Hz?

Ok, I am going to take a stab at answering this but hopefully Wayne or brucek will give us better information. It looks like both of your mains may be causing the dip. I think you can address this by trying to move them around. If not, I still think it can be handled by a treatment panel. By the way, that is a good looking graph even with the 167Hz dip!

elee532 wrote:

The next graph shows my sub and mains with my pre-pro crossed over at 60Hz. The red line shows the crossover setting on my sub set at about one-o'clock (scale ranges from 40Hz to 160Hz). The gold line shows the crossover setting on the sub at about eleven-o'clock. The red line looks a bit flatter, but I wonder whether dialing up the crossover setting on the sub is contributing to my localization issue. I'm really confused by how the crossover setting on my sub works with the crossover setting of my pre-pro... and how do I go about finding the best setting for each??

I would put the sub xo back to the red line setting. I would think you would want to bypass the sub xo and only use the pre-pro crossover.

elee532 wrote:

The next graph shows my mains without the sub. Green is pre-pro set to large with no sub, black is 60Hz pre-pro crossover, and blue is 80Hz crossover. Not posting this one for any reason in particular, though it does seem an odd looking graph to me? ...actually, I just realized something, this contradicts my earlier posting showing no difference between 60Hz and 80Hz crossover setting. Guess I must have screwed up something the first time. Why do my mains dip like that between 30Hz and 80Hz?

That's what we should see. I think this is correct.

elee532 wrote:

Finally, I posted my first attempt at an ETC reading. To be honest, I have no idea how to read it.

You want to zoom in and only look at the first 10-20 milliseconds, not out to 2 seconds. You want to find all the major peaks in that time span and figure out the distances. Combinations of that distance will help you narrow down what surfaces need more absorption put on them. Here is an example of mine:



That second peak right after the initial one is either my back wall (couch is against my back wall) or the couch its self. The peak that I measured out to be 6 feet tells me something within 6 feet of the mains or mic needs to be addressed. I believe it is the ceiling. Then you can see the much smaller peaks that will then need to be address. This graph also helps when you trial place a panel to be sure you have it located correctly.

How did I do brucek/Wayne?

[brucek]

elee532, you have to switch the ETC plot from DBFS to %FS as weverbs plot shows.

The peaks will show the extra distance the signal had to travel to reach the mic. So a back wall at 3 feet from the mic will show up at 6 feet in the ETC.

brucek

[weverb]

So my comments were correct?

[JohnM]

weverb wrote:

So my comments were correct?

Seemed reasonable to me

The sub's own crossover should be set as high as it goes (or defeated if that is an option) so that the pre-pro is handling the crossover.

One thing the ETC highlights is fairly high distortion - those smaller peaks that happen before the main peak are distortion harmonics. Suggests your sub may be working fairly hard even though the measurement levels don't look especially high.

[elee532]

JohnM wrote:

The sub's own crossover should be set as high as it goes (or defeated if that is an option) so that the pre-pro is handling the crossover.

This is still confusing me. My pre-pro is set for 60Hz crossover. As the two measurements below show (red line shows the crossover setting on my sub set at about one-o'clock (120Hz?) - the dial on my sub ranges from 40Hz to 160Hz - and the gold line shows the crossover setting on the sub at about eleven-o'clock (80Hz?).)

This higher crossover setting on the sub shows a flatter response line. Doesn't this reflect that the sub is now producing output all the way up to 120Hz? Wouldn't this be contributing to my localization issue? If my pre-pro is set to 60hz crossover, shouldn't the sub only be getting signal at 60hz and below?

JohnM wrote:

One thing the ETC highlights is fairly high distortion - those smaller peaks that happen before the main peak are distortion harmonics. Suggests your sub may be working fairly hard even though the measurement levels don't look especially high.

This is really vexing as well. The gain on my sub is at about 50%. Moreover, the bass extension frequency switch on the sub is set to 28Hz and Damping is set to Low. I would think these settings would reflect about the easiest load possible on the sub?

Thanks for any additional info!

[JohnM]

elee532 wrote:

This is still confusing me. My pre-pro is set for 60Hz crossover. As the two measurements below show (red line shows the crossover setting on my sub set at about one-o'clock (120Hz?) - the dial on my sub ranges from 40Hz to 160Hz - and the gold line shows the crossover setting on the sub at about eleven-o'clock (80Hz?).)

This higher crossover setting on the sub shows a flatter response line. Doesn't this reflect that the sub is now producing output all the way up to 120Hz? Wouldn't this be contributing to my localization issue? If my pre-pro is set to 60hz crossover, shouldn't the sub only be getting signal at 60hz and below?

Crossovers provide a smooth transition, not a sharp cutoff. If you play with the target settings in REW you can see (in the target curve) what the crossover response looks like for various crossover frequencies for a subwoofer or non-full range speaker. For a 60Hz crossover the sub's output is down 6dB at 60Hz and down 25dB at 120Hz. By bringing the sub's own crossover into play as well you speed up the rolloff of the sub's output but create a dip in the transition between sub and mains.

The localisation issue may be a result of the distortion. The ETC shows significant harmonics from 2nd right up to 6th (so for a 50Hz input that means distortion products at 100, 150, 200, 250 and 300Hz), and these are at the output of the sub, they are not reduced by changing the crossover setting. It does seem odd, as high distortion would typically be seen in a sub being driven past its capabilities but from the extension you have your sub seems very capable - unless that extension is being achieved by boosting the sub's low end and that boosting is contributing to the distortion.

[elee532]

OK, as if I'm not confused enough already...

It seems I am getting completely different results depending on the Start and End Frequency that I use for my sweep. The blue line is when I use 0 to 200Hz for my sweep. The purple line is when I use 0 to 15,000Hz for me sweep.

The results are obviously dramatically different up to about 20Hz. Any insight as to what's going on?

Thanks!

[JohnM]

That is very odd. You don't have a procesor that applies any sort of dynamic EQ or dynamic loudness do you?

[elee532]

No dynamic EQ or loudness on my pre-pro (Emotiva LMC-1). When I take my readings, pre-pro reads "bypass mode, PCM48k." Sounds Effects are checked as disabled on the Enhancements tab for the playback devise in Windows. Here's another graph using various sweep settings.

Red = 0 to 200Hz
Gold - 0 to 500Hz
Green = 0 to 1,000hz
Aqua = - to 5,000Hz
Black = 0 to 15,000Hz

I have no idea what to make of this??

[JohnM]

The difference in the various sweeps is how long the signal spends at the lowest frequencies. If the sub is very non-linear below 20Hz changing the time spent there would affect the level of distortion products. You may get a sense of that from watching the driver as the sweep progresses. You might get more consistent results by changing the sweep start frequency to 10Hz or higher.

[elee532]

JohnM, thank for the reply! I understand what you are saying re: the length of time at each frequency during the sweep might differ depending on the sweep range. However, I'm kind of lost after that (i.e. non-linear sub?). I'm pretty much an amateur at this stuff.

FYI, I did try starting my sweep at 10Hz instead of 0, but the results were the same as above.

What does all of this say about my sub? Which results, if any, results are accurate?

Thanks!!

[weverb]

It almost looks like your amp is switching through the different damping options. It looks similar to the last three graphs on this page:

http://www.rythmikaudio.com/eq.html



Is there a way to check the incoming signal response to the sub to eliminate it as the cause? This might help narrow down where the problem is coming from. You know, test something like:

1. Signal coming out of the computer before the pre-pro.
2. Signal coming out of the pre-pro before the sub.
etc.

brucek has mentioned similar tests to others. If I remember I did something like it to test if my pre-pro's xo was working properly.

[brucek]

Is there a way to check the incoming signal response to the sub to eliminate it as the cause?

Test the receivers response out the LFE port by using REW.

Just as you would test a loopback cable from line-out to line-in to get a flat response, simply include the receiver in that loop and take a measure. You have to fuss with the levels a bit when doing the Check Levels routine, but it's easy enough.

Connect soundcard line-out to receiver AUX-in, and receiver Sub-out to soundcard line-in, and measure.

Be sure the meter cal file is cleared and the soundcard cal file is loaded.

The response should be just like one of the REW targets, except it will roll off down low. Measure from 0Hz.

brucek

[elee532]

brucek,

I'm think I am misunderstanding your instructions. When I follow what I think are the connection instructions you provided above, the output meter reads about -15 but the left/right input meters read nothing. Is the SPL meter supposed to be connected in the mix somewhere?

Thanks and sorry for my ignorance here!

[brucek]

Is the SPL meter supposed to be connected in the mix somewhere?

No, you're measuring the electronic frequency response of your receiver by itself, just as if it was a loopback cable. No meters, no speakers - you're measuring the electronics.

Connect soundcard line-out right channel to receiver AUX-in right channel, and receiver Sub-out to soundcard line-in right channel. Run the Check Levels and Calibrate routine and Measure.

but the left/right input meters

You'll only be looking at the right channel input meter. The left channel is not used.

brucek

[elee532]

Thanks brucek. I'm a little closer now, not sure what I was doing wrong before. However, when running Check Levels, nothing I seem to can get the Output and Right Input meters to balance. The output meter is always louder than the input meter by something like 12db.

Also, can you confirm for me that "soundcard line-out right channel" is the same as the right Front speaker connector?

Thanks for any advice!

[Ricci]

So this is a Rythmik sub? Which one 12" or 15"?

I've seen artifacts like that below 15hz before but not that dramatic. Looks almost like a cal file got added or something.You will get different results depending on your sweep range and duration. From what I have seen the longer the sweep stays in that range the more accurate it will be. A 1M 0-200hz sweep is going to be better than a 128k 0-24000hz sweep for the bass range. The very short sweeps seem to exaggerate the bass below 15hz a little. THIS IS OPPOSITE OF WHAT YOU ARE SEEING THOUGH! You usually have a LOT of compensation added to the readings in that range, which can do some screwey things from multiple cal files. I've never seen it affect things all the way up above 23hz though and with that big of a difference. Usually it's just a few db.

How hard is your sub being driven on these sweeps? Looks almost like a limiter is kicking in or something. The Rythmiks are small sealed subs and they are servo controlled and eq'd flat down to the really deep bass. That would cause the driver to be pushed very hard below 20hz even with a moderate volume sweep. What length were those sweeps? Can you keep the sweep range the same like 0-200hz and run one at each length and see if there is a difference there between the shorter and longer ones?

[brucek]

can you confirm for me that "soundcard line-out right channel" is the same as the right Front speaker connector?

Yep, usually.

However, when running Check Levels, nothing I seem to can get the Output and Right Input meters to balance. The output meter is always louder than the input meter by something like 12db.

I'm a bit confused by that statement. It makes no sense to me. You use the receivers volume control to set whatever level you desire on the soundcards input meter.

First connect a single loopback from line-out (right) to line-in (right) and do a Check Levels and Calibrate @75 and a Measure to be sure you get a flat line.

Now all your levels are setup in the soundcard. Then insert the receiver into that loopback and fuss with the receivers volume control when you run the Check levels routine. Don't touch the soundcards input or output levels, because you already have them in a good spot as if you were measuring a cable, but instead you've placed a receiver in the loop.

Get it?

brucek

[elee532]

brucek wrote:

First connect a single loopback from line-out (right) to line-in (right) and do a Check Levels and Calibrate @75 and a Measure to be sure you get a flat line.

brucek, I'm sorry, but I'm confused. When I connect line-out to line-in, I don't actually get any sound from my sub or speakers. So, what/where am I looking to get a 75db reading?

Thanks!

[brucek]

When I connect line-out to line-in, I don't actually get any sound from my sub or speakers. So, what/where am I looking to get a 75db reading?

After you run the Check Levels routine, you run the Calibrate routine. You don't hear any sound (of course), but you set the level to 75dB with the thumbwheel in the Calibrate routine to fool REW's SPL meter into thinking the level is 75dB. How else would REW know where to place the trace on the graph. You have to tell it that the level of the loopback signal is 75dB, so that's where the trace will lie. The advantage of doing the loopback first before you insert the receiver in the loop is to first verify that a cable without the receiver is indeed flat, and to get a good setup of the soundcards line-in and line-out before the receiver is inserted in the loop.

Then you remove the loopback and connect the receiver and go through the Check Levels routine using the receivers volume control to get the correct input level.

brucek

[elee532]

Thanks brucek! This is hopefully a correct attempt at measuring my pre-pro's response...

• I set Windows Volume to 100.
• I connected soundcard line-out (front right speaker) to soundcard line-in right channel.
• I cleared meter cal file and kept soundcard cal file loaded.
• I ran check levels.
• I adjusted the sweep level to -20 which is where the out and right meters werer equal
• I ran "Calibrate" using REW subwoofer cal signal as signal source
• I set to 75 and clicked finished
• I took a measurement using 0 to 15,000Hz, -20 level, 256k length, 1 sweep.

.
The result is shown in the first graph below (green line). Does it look as it's supposed to?

• Next, I connected soundcard line-out right channel to pre-pro aux-in right channel, and pre-pro sub-out to soundcard line-in right channel and ran Check Levels.
• I kept all settings the same as described above and only adjusted my pre-pro volume until out and right were equal (at around -20). I had to crank the volume on my pre-pro up to about 95%. Is this common?
• I ran a measurement using the same settings as above. The results are shown in the second graph (red line).

Do these both look as they should?

Thanks!

[weverb]

Looks correct to me. brucek will have to verify. It looks to be doing just like John stated:

JohnM wrote:

For a 60Hz crossover the sub's output is down 6dB at 60Hz and down 25dB at 120Hz.

[brucek]

Do these both look as they should?

Yep, they look correct.

I would display the measurement graph with an upper limit of 200Hz to get a better look, and I might add the 60Hz REW target line to give yourself assurance that it's tracking close to spec, but it looks fine. There's nothing wrong with your receiver....

brucek

[weverb]

Would you say brucek that this pointing to the sub as the problem with his graph variances?

[elee532]

Here's a closeup. Any idea what that weird "blip" is about at 120Hz? Also, is it normal that I needed to crank my pre-pro's volume up to 95% to get the output and input levels even?

I'll work on some of the measurements suggested above by Ricci next.

Thanks again for all the help!!