How to get Sub graph with AVR's crossover?

[grn1969c10]

I have worked with REW for a few weeks now and found it to be a nifty tool for showing the full band response of each speaker - sub, mains, surrounds, whatever - which can help with placement. I have not figured out how people come up with nice graphs showing the rolloff of the AVR x-over.

REW's mono output split and connected to L & R analog inputs of the receiver which is set to stereo (or any other listening mode) yields the exact same full range frequency response as running the mono REW output directly to the speaker and matching levels. By full range, I mean whatever I set the measurement sweep to be is what I get. If I use 15 - 200hz I have a sub trying to crank out 200hz and mains thumping along at 40hz.

The receiver is an Onkyo TX-DS575. While an older unit, it is a DTS surround sound receiver and should have rudimentary bass management capabilities. (I think...) Any ideas on what I'm missing?

[Wayne A. Pflughaupt]

REW's mono output split and connected to L & R analog inputs of the receiver which is set to stereo (or any other listening mode) yields the exact same full range frequency response as running the mono REW output directly to the speaker...

"Directly to the speaker?" Your speakers have RCA inputs?

Regards,
Wayne

[brucek]

I have not figured out how people come up with nice graphs showing the rolloff of the AVR x-over.

Your mains speakers must be disconnected to show the sub only....

brucek

[grn1969c10]

Crummy description on my part. I shouldn't post before I have had breakfast. Let me rephrase. (I'm good now.)

Case 1:
A 15Hz to 200Hz REW measurement output directly to sub via RCA jack yields same result as REW to AVR and then AVR output to sub. I do have to readjust the levels, but with both test set to a similar target level of 75db they are essentially identical plots. The AVR does not appear to apply any crossover to the sub pre-out even with mains set to "large". My impression was subwoofers shouldn't be passed high frequencies to prevent unnecessary distortion.

Case 2:
A 15Hz to 200Hz REW measurement output to AVR and AVR output to mains show no low end rolloff from crossover. Setting mains to "large" in the receiver menu yields identical REW plot to setting mains to "small". The plots are identical and output is decent down to the 40 - 45hz range. My impression was "small" mains shouldn't even be passed such low frequencies to prevent distortion.

[grn1969c10]

Your mains speakers must be disconnected to show the sub only....

brucek

Thanks bruce, I did forget to disable my center in one test which led me to some absurd conclusions, but my mistake was discovered and I retested.

[brucek]

I did forget to disable my center in one test

You should only need to disconnect the mains since you are in stereo mode. The center shouldn't matter - it's not active in stereo..

Do a test of your receiver by itself. Connect the output of the receivers sub-out to the REW soundcard line-in (as opposed to having the mic plugged there). Then you can get a correct response of the sub out crossover of the receiver without having your sub or the room involved. The only change in measuring would be that you have to fuss a bit with the AVR volume control to get the levels correct when doing the Check Levels routine.

brucek

[grn1969c10]

I'll give it a whirl...

[Wayne A. Pflughaupt]

We saw a case recently where a Member's receiver's crossover only worked with digital sources, IIR, like DVDs. Maybe yours is the same way...

Regards,
Wayne

[grn1969c10]

Bruce you are the man.

I ran a measurement with the mic cal file cleared, C-weight unchecked and sub pre-out looped into soundcard. Playing with the target settings I get a best target to measurement match at 12db/octave and 90Hz cutoff. The measurement was the same when I set mains to "small" (previously "large") and you can't even tell there are two different measurements on the plot.



I enabled the RS meter cal file and ran three new measurements. I must have skipped breakfast all last week because the effect of the crossover can clearly be seen. Purple and Green are from the AVR to sub with mains large and mains small in the settings menu. Blue is REW direct to sub. Proving the crossover is funtioning is good, but any ideas on why the measurements are so horrible? I did disconnect the mains!



Consistency in testing is generally a good thing (but I haven't used much of it). The major difference that I am aware of between last week and todays test is today I placed the mic about three feet in front of the sub pointing at the cone and set 85db target levels to counter higher levels of ambient noise from gusty winds. Last week I had the mic at the seating distance pointed toward the ceiling and 75db levels.

[brucek]

Proving the crossover is funtioning is good, but any ideas on why the measurements are so horrible?

The response is a combination of the actual response of the sub, plus the room effect.

The room is creating a rather large peak at 70Hz.

You'll find corners offer some better low end extension (something you could certainly use).

What are the response specs of your sub?

brucek

[grn1969c10]

It's a Cadence CSX-15. They apparantly do not publish any "tested" frequency response graphs for the sub.

This is a paragraph from a response I received from Jake at Cadence in regard to expected freqency response and the benefit of facing the sub into the corner for a "corner loading" effect which I tried for a week or two. (Now that has to be a run-on sentence! Sorry.)

Regarding the spec, the amplifier response is definitely 10-120Hz +/-3dB
LP (10Hz-500Hz ByPass) while the subwoofer response is higher as you
mentioned. Because it is a tuned enclosure there is no way for us to
predetermine optimum location in each listening environment, especially
considering room acoustic anomalies. We therefore tested the subwoofer
in various different positions and corner loading certainly allows for
increased reflection and bass wave expansion and lower response.

I did notice their website claiming under the subwoofer specifications,

Frequency Response: 20-500Hz

however the owners manual claims

Frequency Response: 25-250Hz

At any rate, it is a 15" sub with a 225 watt amp and should certainly be fine worst-case-scenario down to at least 30Hz, if not 25 or 20. I am investigating the possibility of DIY base traps. Do you really think my room could yield that bad a response? It's 16 feet wide by 26 feet long and has two large openings into additional rooms.

[grn1969c10]

I picked the current location for the sub after dragging it all around the room for a day taking measurements. All the other locations that I tried had a very bad null at 32Hz which completely wiped out the low end.

[brucek]

Do you really think my room could yield that bad a response?

Sure, but you can do a near-field of the sub in the middle of the room (or better outside) to eliminate the room as much as possible to find the true response.

I picked the current location for the sub after dragging it all around the room for a day taking measurements

No need to do that. The accepted method is to put the sub at your listening position and move around the room to different locations with the mic. Best method is to use the RTA feature of REW.

brucek

[grn1969c10]

Here are the results from about five different spots in the room.
Starting nearest the center of the wall by the TV and working counterclockwise around the room.
Gold, Blue, Green and then Purple and Red. The green dot marks the mic location.



In this round of measurements I noticed very little bass was gained by moving the sub into the corner or using corner loading. I tried corner loading at three different distances from the corner with no obvious benefit and deleted the results.





Crawling for bass:
I'm a little unclear about the crawling for bass solution. If I use pink noise, this would probably tell me where I'm exciting some room node which will give me a ton of gain in the 70hz region which I don't even want.

Using RTA:
I will have to read your sticky on this feature I haven't tried it yet.

But have no fear, while I'm making no progress at all improving the sound, each time I delve into it I learn something new!

[grn1969c10]

While running measurements with the sub and mains together, I noticed a nasty null at 305hz. The slope of the null was very steep. For fun I used the generator to produce that single frequency and walked around. Craziest thing I ever heard. Really loud in one spot and then barely audible 12 inches closer. Yeah, yeah, I knew what a null was. But knowin it and actually isolating it with a test tone and experiencing it are two different things! I then noticed if I turned off the sub, the null was lessened considerably (though it didn't go away).

I suppose with a 12db/octave crossover centered at 90hz I'm only reducing 305hz by about 20db? Apparently thats not enough to prevent some funky interactions between the sub and the mains.

[brucek]

Your sub does not appear to support its claimed response extension to 25Hz, regardless of where it's situated in the room. It would be nice to see a response measurement without room influence to see if you're fighting a losing battle......

[grn1969c10]

I have to study for an exam and don't have much time to spend on this today.

I placed the sound meter 1/4" from the center of the slotted port on the CSX-15, set levels at 90db, selected 1M sample, two sweeps and measured. I then pulled the sub out 3 feet from the wall and repeated but saw no change in the measured response. For a comparison I placed the meter 1/4" from the round port of my JBL-PSW1200 and repeated.

The old JBL is in gold and the new CSX-15 is shown in blue. Hopefully the port opening was the correct location to close mic the sub. I borrowed an eq to play with but it may be a week or two before I can mess with it.



Added: I think the JBL is listed as -6db @23hz
(So my meter, soundcard, room, or methods probably give up some accuracy!)

[brucek]

Hopefully the port opening was the correct location to close mic the sub

No, ported designs have to be mic'd at a distance where you feel both port and driver are acting in concert as one.

The response of the port and the response of the driver in a ported design are quite different. I won't go into it (you can read about that), but you need to be further away from the sub than in a sealed design to take a near-field reading, since the driver drops off and the port takes over to extend the bottom end. Outside would be best, but the middle of the room would do.

It does look as though you have a nasty peak at 70Hz that you are going to have to equalize. Fairly simple to do.

brucek

[grn1969c10]

10-4 boss.
What is the appropriate distance to ensure the port and speaker are both being picked up equally? I am guessing three feet with the mic splitting the difference between the center of the cone and the port? I might redo it later tonight if I get the chance.

[brucek]

I am guessing three feet with the mic splitting the difference between the center of the cone and the port?

Yeah, I suspect that's about right for a front firing port...

brucek

[tenzip]

You can go look at the "Subwoofer test" forum, to read how Illka does it when he measures ported subs. I'd imagine his method is fairly sound. Not sorry about the pun.

[grn1969c10]

Well, no great excitement here. I pulled the sub out so it had over six feet clearance from any wall. Set meter 36" in front and measured (light Blue). For kicks I also moved the mic closer to 19" in front and remeasured (Green). (Dark Blue) is a previous measurement made in error with mic basically in the port opening.

[grn1969c10]

You can go look at the "Subwoofer test" forum, to read how Illka does it when he measures ported subs. I'd imagine his method is fairly sound. Not sorry about the pun.

I have read some of his reviews, but not with the intent of paying enough attention to the details to repeat his process. Then again, I don't have the quality calibrated measuring gear he has and apparently my sunroom makes a lousy anechoic chamber. I'm looking into DIY bass traps but I'm not sure yet how I want to build the frame and cover them. I'm thinking about an OC705 equivalent 4" thick, running floor to ceiling in each front corner.

I'm just cranky because I was hoping the CSX-15 would be a big improvement over my antique, $25, yardsale JBL and while its better, I'm having a difficult time quantifying "better". I'm bummed out that my receiver has a fixed 2nd order crossover at 90hz and I'm annoyed my DVD player doesn't support DTS. I'm tired of going to school and I wish I still had my old job so I could afford to buy stuff. I'm just plain no fun to be around tonight.

But thanks anyway, a well played pun does cheer me up a little bit.

[brucek]

I'm looking into DIY bass traps

That's not going to reduce the peak at 70Hz. You'll need to equalize that with an inexpensive BFD.

Basically you're listening to the peak at 70Hz, as it overwhelms the situation. If you reduce it, it allows you to then turn up the wholesale volume of the sub and the low end will shine through.

brucek

[grn1969c10]

and the low end will shine through.

What?? A ray of hope? I borrowed a DSP1100p to play with but I forgot to look for 1/4" to rca adapters.