[rogerv]

Hi Everyone,

I'm working my way through the steps to use REW, and I've run into something that I don't understand. Here's a summary of where I'm at:

1) I have a Creative Labs "Blaster X-fi" external sound card. This card has separate right and left channels for line out. It has a single connector for line in.

2) I'm using a Radio Shack digital SPL meter.

I believe I have done the sound card calibration properly. I set up the cables as shown in the first attachment.

On this picture, I ran the right line output from the sound card to the right side of the splitter plugged into the sound card's line in. I believe that this Radio Shack Connector splits out left and right signals from the single pin. The picture shows a second splitter which is identical to the one plugged into the sound card for reference.

The second attachment is the sound card's uncorrected response. The corrected response is completely flat from 2 - 24Khz. I don't have a jpeg for this graph.

So am I on track as far as the setup and procedures for calibrating the sound card ?

I think I have done all of the steps correctly, but when I go to the acutal measurment phase of the process, I'm running into something I don't understand, so I thought it's possible that I've done something wrong with these steps, and would be best to check.

Measurement Phase:

My thought was that I would measure the subwoofer first on its own without involving the mains. My sub has a separate amplifier with line in connectors for signal, and a separate output level control. For REW, I took the Right Output from the sound card's line out and ran it into the sub's line in, and ran the SPL's output into the same right side connector that I used for the sound card calibration.

3) I ran the sound meter calibration procedure as described in the help files. For this, I played the subwoofer test tone, and set the numeric value on REW to match what was shown on my SPL meter. I adjusted the output at the sub's amp so that it read 75 db.

4) Next I did a measurment. This is where I don't understand something.

5) I clicked on the measurement icon, set the start Freq to 15 HZ, and the End Freq. to 200 Hz. When I pressed the "check levels" button, I get a message that says "Level is Low". I'm not sure what level this message is referring to. Is it the output from the Line Out right channel, or is is the output from the sub that is low, or is it something else ?

I was able to set the Level (dB FS) setting to make this message to go away. But when I did this, the measurements were all above the recommended 75 dB level. I don't seem able to find the right combination to make the "Level is Low" message go away while at the same time keeping the measured output at 75 dB. If anyone has any words of wisdom, I'd greatly appreciate it.

Thanks In Advance

Rog

 

[brucek]

So am I on track as far as the setup and procedures for calibrating the sound card ?

Yes, and the soundcard file is fine. BTW, there is no corrected and uncorrected soundcard file. They are identical - it's simply the soundcard calibration file.

For REW, I took the Right Output from the sound card's line out and ran it into the sub's line in, and ran the SPL's output into the same right side connector that I used for the sound card calibration.

Yes, you can feed the sub by itself, but it's highly recommended to feed your receiver/processors AUX input instead. If you want left and right channels tested, use a Y-splitter at that point.

Set the receiver into stereo mode and set the crossover you would normal use. For sub only, shut off the mains speakers.

Connect your SPL meter to the line-in.

Then run the Check Levels routine and adjust your receivers volume to receive 75dBSPL at the listening position on the actual SPL meter. Set the input volume so the REW VU is about -12dB.

Then run the Calibrate SPL routine.

Then Measure from 0-200Hz.

brucek

 

[Guiria]

I had this problem when I initially setup REW and it drove me crazy. I even ended up buying TrueRTA to use instead of REW but it was not the same. I finally figured out that changing from right to left (or left to right, I can't remember Either way just choose the one that is not selected now) input in the REW settings fixed my problem and I no longer received the "level is low" message.

Good Luck. I can post some pics later if my description doesn't make sense but I'd have to bring my laptop home from work...

 

[gperkins_1973]

I was having the same problem last night. I can calibrate the spl to 75db but even with the soundcard input/output volume on max I could only get -35db levels. I changed my spl meter setting from 80db to 70db and didn't have a level problem but what I did find was that the headroom was clipping or going into the red. Total nightmare.

cheers
graham

 

[Wayne A. Pflughaupt]

4) Next I did a measurment. This is where I don't understand something.

5) I clicked on the measurement icon, set the start Freq to 15 HZ, and the End Freq. to 200 Hz. When I pressed the "check levels" button, I get a message that says "Level is Low". I'm not sure what level this message is referring to. Is it the output from the Line Out right channel, or is is the output from the sub that is low, or is it something else ?

I've had the same problem of late. Try ignoring the "Level is Low" and go ahead and take a measurement. Mine will come out fine despite the low level warning.

Another alternative is to use the RTA feature instead.

Regards,
Wayne

 

[gperkins_1973]

Wayne,

Yeh! your right when I went ahead and measured it had a headroom of 16db in green so I guess that's fine. Its was doing the levels on the speakers that was a pain. I set the spl meter to 80db on the knob, set the spl meter to read 75db, yet on REW is it was really low but when i switched it to 70db on the spl meter and then set the level to 75db it seemed fine, I could hit the 18db level easy but when i went to measure it was clipping. Is the procedure the same for subwoofers as to speakers or are the set ups different. C weighting was on and it was set to slow. Is this correct?

cheers

Graham

 

[Wayne A. Pflughaupt]

Yeah, keep the meter at the 80 range, C weighting at all times, be it subs or w/ mains. The fast/slow has no bearing, it's just for the meter's own display.

Regards,
Wayne

 

[rogerv]

Thanks to everyone for the additional info and encouragement. :thankyou:

I have made some progress, and have gotten some measurements which I will upload in a second post. Here's some description of what I ran into while doing this (in case it's useful for anyone else). I'm still not sure what exactly was happening. I've got a suspicion about what was going on, but haven't gone back to reproduce the problem. Maybe someone knows off the top of their head.

Problem #1: Bad hum from the speakers.

I had a terrible hum whenever I selected the amp's connection from the soundcard. I couldn't make it go away, and initially thought it had something to do with cables and splitters or possibly the sound card itself. In troubleshooting this problem, I noticed that the power cord for the monitor was lying on top of the audio cable going to the amp. So, I unplugged the power cord to re-route it. The hum went away when I unplugged the monitor. After additional trouble shooting, I attributed the problem to either the LCD monitor that I was using (an IBM 19 inch LCD), or the digital cable coming out of the port replicator, or noise generated by the video card on the laptop's main board. Unplugging the monitor and using the laptop's display fixed it. The screen's a bit small, but I got rid of the hum.

Problem #2: Output Displays All looked the Same.

I really scratched my head on this one for quite a while. When I tried to make a measurement, I could hear the sweep tones generated in the speaker, but the generated results all looked like attachment #1. No matter what I did, it all looked the same.

So, I went back and double-checked settings etc. In fiddling with the settings, I noticed that default output and input devices were selected in the sound card settings. I clicked on those, and re-selected something different (I think "speakers" for output and "Line In" for input.) I then ran the "check levels" from the soundcard settings.

Another measurement then produced normal-looking results. I shut down the p.c., and re-started it to see if something was being changed. REW now worked properly. My hunch right now is that when I disconnected the laptop from the sound card yesterday to put it back downstairs in my office that the that Windows re-assigned the default sound card to something else (probably internal speaker). I'm thinking through a test case where I can try and re-create this. But, that's the computer nerd in me. The audio nerd is just happy that it worked and that I got some measurements.

Rog

 

[rogerv]

Another Newbies First Measurement

Hi Everyone,

I was able to get everything working right (I think), and have some results to post. The crossover for the amp is set to 60 Hz. Attachment #1 shows the response which used 8 iterations of the test tones.

I don't know what to make of the rather severe drop out around 110 Hz. The only thing that I could think of was that it might have something to do with the bass management system from the amp, or possibly something to do with the amp itself ?

So I re-ran the test with the sub disabled and only using the mains (second attachment). The drop out seems to have gone away. with the sub out of the picture. Would it be fair to say that since the drop out went away with only the mains involved that it's probably not room acoustics ?

Anyone have any thoughts about what might be going on, or other comments ?

All I can say is that I sure love playing around with this stuff. I haven't had this much fun for a long time

Rog

 

[brucek]

Would it be fair to say that since the drop out went away with only the mains involved that it's probably not room acoustics ?

No, it wouldn't be fair to say that. The sub and mains are in different physical positions in the room. They will most certainly suffer from different reflection cancellations from the room, since they are at different distances from the wall.

What the sub and mains are enjoying together is a room resonance at ~38Hz, evidenced by the shared peak at that frequency.

brucek

 

[glaufman]

One thing I find curious is that the mains seem to be putting out so much energy below the crossover... did you disable the xover for the mains only scan?

To test, run a scan with just the sub. I would guess it won't show the same dip, in which case I would guess whatever energy it's putting out at that frequency is cancelling with that of the mains, i.e. a phase issue. But it's just a guess.

 

[Wayne A. Pflughaupt]

One thing I find curious is that the mains seem to be putting out so much energy below the crossover...

Yeah, no kidding. That's one of the worst mains peak I've ever seen!

Regards,
Wayne

 

[rogerv]

Bruce,

Thanks for the reply.

Maybe you can help me understand this better. I thought that if the crossover was set at 60 hz. that the sub would handle <= 60 Hz., and the mains would take everything else. Is this dip being caused by interactions between what the sub is putting out and what the mains are putting out ? How does that work ?

Please bear with me. I'm not very versed in a lot of these types of technical details. Feel free to refer me to some outside documentation or threads that I can read if that works better.

Regards,

Rog

 

[rogerv]

Greg,

When I ran the mains only scan, I told the amp that there was no subwoofer. I set up the left and right speakers to full range. I think this eliminates all of the crossover processing, but maybe not.

The mains are B&W 603's. The Amp is a Sunfire Ultimate Receiver II. Sub is a HSU TN1220 HO.

I'll test the sub only tomorrow. (Gotta use the theater tonight)

Rog

 

[brucek]

I set up the left and right speakers to full range.

Why? Don't you want to see the mains as they are used in your system with the crossover engaged.

brucek

 

[glaufman]

When I ran the mains only scan, I told the amp that there was no subwoofer. I set up the left and right speakers to full range. I think this eliminates all of the crossover processing, but maybe not.

Yes, it would. But it's better to tell the amp (or processor) there IS a sub, and turn the sub off, so you get a better picture of what the mains are doing in the real world. Similar for the sub-only scan, don't tell the amp you've changed anything, but disconnect the mains to run the scan and see precisely the contribution from the sub.

I'll test the sub only tomorrow. (Gotta use the theater tonight)

Always good to stop tweaking for a while and enjoy what you've got!:T

 

[glaufman]

Maybe you can help me understand this better. I thought that if the crossover was set at 60 hz. that the sub would handle <= 60 Hz., and the mains would take everything else.

Basically true, but not quite. The crossover isn't a "brick wall" that stops precisely at 60Hz. Starting at a frequency just higher than the crossover point and moving lower, the crossover gently rolls off the mains as it gently rolls up the sub. If you started at a point just under the crossover and moved higher, you'd see the amp gently rolling off the sub and gently bringing up the mains. If you looked at the freq plot of the mains and the sub output of the amp (no acoustics involved) and overlayed them, you would see them crossing over each other at the point where you told the amp to set the crossover frequency (hence the name)... At this point, summing the individual contributions brings the combined response (assuming their in phase, and level matched, and...) back to flat. The point is that each one (sub and mains) put out useful and necessary sound energy for at least an octave above and below the crossover point respectively.

Is this dip being caused by interactions between what the sub is putting out and what the mains are putting out ? How does that work ?

Possibly. To test this we run the individual scans, and if they're both clean, then there's an interaction between them. Simply put, if they're in phase they sum nicely. If they're out of phase, the cancel. Anything in between, and you get some portion of the cancelling effect. Sort of.

Please bear with me. I'm not very versed in a lot of these types of technical details. Feel free to refer me to some outside documentation or threads that I can read if that works better.

No problem! We all start somewhere. I'm still learning from Wayne, Brucek, and everyone else here as well. :TThe pictures in this link might help:
http://en.wikipedia.org/wiki/File:Interference_of_two_waves.png

 

[brucek]

If you looked at the freq plot of the mains and the sub output of the amp (no acoustics involved) and overlayed them, you would see them crossing over each other at the point where you told the amp to set the crossover frequency

Something like this..........



brucek

 

[rogerv]

Thanks to everyone for the additional explanation. It helped a lot, and I think I'm starting to get a handle on some of this. I think I understand now what I was supposed to do and what we're looking for. Here's my understanding of what everyone has said so far:

I want to capture the sweeps from the mains with bass management and without the subs contribution to look at the characteristics of the curve for the mains. Similarly, I want to look at the characteristics of the curve for the sub with bass management and without the mains contribution to look at it's curve. The combination of the two curves should provide additional information about how the two components are interacting and possibly either explain the drop out around 110 Hz., or point to where to gather additional diagnostic information.

I understand now why everyone commented on how strange the curves for the mains looked. My thinking when I ran the mains without the bass management was that it would provide a base line of sorts that other measurements could be compared to. I've used a similar approach in some of the work that I've done in the past (software system integration testing). Probably the wrong approach for this situation, and I now recognize the differences between the two environments.

I have repeated the measurements, and have attached them to this post. I don't know if the attached are any better than the previous ones or not. There are four sweeps attached, which I have labeled on the graphs. The crossover for all four was set at 60 Hz. One is for the mains without the sub. I physically dis-connected the sub for this sweep. Two are for the sub without the mains. I have a switch on the sub's amp labeled "polarity." One sub sweep is with the polarity at zero, and the other is at 180. For both of these sweeps I disconnected both the left and right speakers. The final sweep is with the mains and sub, but with the "polarity" set to 180 on the sub's amp. The post that I put up yesterday was with this toggle set to zero, which is where I normally have it.

It still looks like there's a high peak around 40 Hz even with the bass management engaged. It still may be the "worst ... ever seen" even though the peak is lower than what I posted yesterday. So I guess one of the outstanding questions is why the mains are generating so much energy below the crossover frequency. This is in addition to why we have such a severe drop out slightly above 110 Hz.

A further piece of info is that I did do a number of sweeps on Thursday (which I didn't post) using different crossover frequencies. All of them show a peak at around this same 40 Hz. point. Some of them are a little narrower than others (I think), but I have no instances out of 12 measurements where this peak is not there. So, I don't know what might be causing this.

Any comments are appreciated as always.

Thanks and Regards,

Rog

 

[brucek]

one of the outstanding questions is why the mains are generating so much energy below the crossover frequency.

Not an outstanding question really - it's your room. You have a nasty resonance at 38Hz. It's as simple as that.

My office system has the same problem. I helped it somewhat by raising the crossover, and this offers more to the sub which has EQ on its path, but it still has that horrible peak. The sub enjoys the same peak, so when they combine with the mains it creates quite a horror show at that frequency.

If you took your speakers outside and measured them a few feet away you would be amazed how perfect they are. It's always the room.

brucek