Help with understanding my graphs

For both Graphs the SPL was attached to the FastTrack Pro and the FastTrack Pro was connected to the receiver. The sub out on the receiver is going in to the 1124P and the 1124P is going to the sub.

Ok, every time I created a set of filters, I all the sudden got a very big dip in my graphs at about 30 to 40 hz. What I started to realize is that my base measurement was never that way. I couldn't create a good graph becaue the 1124 is not adding valleys. I may just be overlooking something.

So what I did was take a reading with the setup above with the IN/OUT button IN but there are no filter settings (all red LEDs are dark) stored in the 1124 and then I took another measurement with the IN/OUT button off (OUT I suppose) so that teh 1124 would not apply filters even if they were set. I would expect them to produce at least somewhat similar readings.

Here are the results with the IN/OUT button off (EQ bypassed)



And with IN/OUT On (EQ on but no filters set)




Any help is appreciated

[brucek]

Ok, every time I created a set of filters, I all the sudden got a very big dip in my graphs at about 30 to 40 hz

If by this you mean that the second graph shows the response dropping away at 30 to 40Hz, I see what you mean.

This is certainly not normal for a BFD and you either have some filter set or you have a fault on the unit. They simply don't do what your graph shows. The IN/OUT simply shuts off the filters section in the DSP. It isn't a hard bypass as you might expect.

If you want to do the response check of your BFD, you can use REW to accomplish that. It might be worthwhile in your case.

Here's how.

When you calibrate the soundcard, you have a cable from line-in to line-out on your soundcard. After you've completed that calibration, leave the cable in place and do a response measurement of the cable itself (this shows that the soundcard cal file is working). You should get a perfect flat response.

Now insert the BFD into that cable and do a response check. (obviously you'll need two cables). The measurement will show the response of the BFD in the loop. (Be sure to have the meter cal file clear so it isn't added into the mix).

Now you can play with the IN/OUT switch and do some response checks of the unit itself and really see what the story is...

By the way, when you use REW for measuring a sub, set the Graph axis of the vertical from 45dB to 105dB (using the Graph Axis icon in REW. Also best to set the horizontal axis from about 10Hz to 200Hz.

When you measure a device, such as the BFD, you can exaggerate the vertical axis to something like 70dB to 85dB, so you can really see the small response differences.

Here's a response graph of my 1100P with the IN/OUT switch in and then out. I had to separate the two graphs, because if I didn't, you wouldn't know there were two. That's how close the response is between IN and OUT...... See how low the BFD goes in its response - all the way to 2Hz and that's with a fairly exaggerated vertical scale. The 1100P actually has a better response than the 1124P.....





brucek

brucek

Ok, I will do these tests for sanity.

I think I may have figured it out. On equalizer 10 there is a left and right bank. I don't fully understand how those inteact with the outputs, but I assume output 1 is left and output 2 is right. I cleared the eq settngs on the left channel, but my sub was plugged in to right channel output and even though I thought when you slect the left EQ button, the right filter was still filtring the right channel.

Actually, I'm not making sense, but essentially when I removed both filter from left and right I got much more consisten graphs betwen IN and OUT.

Thanks Bruce

[brucek]

On equalizer 10 there is a left and right bank.

Most people use program 4 and 5.

Note the input/output Channel 1 on the rear is the Left Channel and the input/output Channel 2 on the rear is the Right Channel.

brucek

Thanks Bruce. Can you tell me why they use 4 and 5?

[brucek]

Can you tell me why they use 4 and 5?

No reason other than when a BFD is new it comes preloaded with programs 1 - 3 set for feedback destroy mode and 4 - 5 for parametric amplifier mode, and 7 - 10 for something else, etc, etc. So people seemed to simply use 4 or 5 since all the filters were set for PA.

My advice to anyone has always been to take a few minutes when the get the thing and set every filter in every program to OF mode.

Personally, I use program 2 for my stereo two channel music and program 5 for 5.1 HT. Makes it easy to remember.

brucek

[clubfoot]

Nice tip there brucek, I'm going to use it.

Ok, I have gone and calibrated everything again and it looks good.

I spent a lot of time moving the sub around last night and most of my graphs look like the first one I posted with the dips around 34 and 60. Now that I have 6 GIK bass traps in the room stacked 2 high in 3 of teh corners (there is no real 4th corner to speak of) I dont have as dramatic of a peak at 42hz.

I was able to move it to reduce the severity of the 34hz dip, but then I had a dip that went off the chart at 65hz. It was a very narrow dip. Anyway, that was with the sub almost mid point in the wall.

So, I think I want to try something. How do I test the sub curve and minimize the effects of teh room? Do I measure it close to the sub and point it at the sub? Can this be done? I just want a reference curve. This is an ep500 and I want to make sure the DSP isn't causing some of this.

Also, what causes large drops like that? The room, the sub. I know the answer is both, but I can move it around a lot and even when it isn't as steep, it is still there. I was thinking of trying it in the back of the room to see if it dramatically changed the graph.

Also, people mention setting front speakers to small. I don't have that setting on the onkyo nr1000. I can set the crossover point for all 7 speakers. I have teh sub and the fronts both set to 80hz. Is that right? For many of the tests I disconnected the mains.

Good think I've been working out :-)

[brucek]

How do I test the sub

The best of course is outside away from any boundaries. Barring that, put it in the middle of the room away from as many boundaries as possible and put the mic about a foot away from the main driver. It will give a decent near field approximation. You do have the floor as a boundary, but unless you're prepared to hang the sub from the ceiling, you have to live with that..

Also, what causes large drops like that? The room, the sub

The room. The sub will output exactly what its specs say, so it's really not required to test it.

I don't have that setting on the onkyo nr1000

Yeah, you do. If you can set a crossover, that means the mains are set to small. Their low frequency is high passed at the crossover frequency. The sub is low passed at the crossover frequency. The two combine for a flat response....

brucek

Thanks bruce. On the crossover question, what I meant was that I don't have that small / large setting like I do on my other Onkyo. I have it set at 80hz. Is that the right crossover point for the mains?

When I do drive all the speakers, I do see the graph flatten out across teh 75db line which I believe is the goal

[brucek]

I have it set at 80hz. Is that the right crossover point for the mains?

Yeah, that's fairly standard. It suits most situations. If your mains are challenged, you might go higher and if your mains are really capable, you might go to 60Hz. You lose equalization potential when you go lower though. When you go higher you risk localization and loss of soundstage.

brucek

I'm hoping the M80 v2 is strong enough...

I just ran tests from othe side of room and my valleys all move.

I'll post some pics soon.

[clubfoot]

Have you tried moving your seat location?

Well, that is a good point. The answer is no because I know my butt will be in a specific place. However, I see how that might help me see how my "position" is affecting the curve. I do have a bit of a square room. I do notice that it sound different back by the bar area than at my seated spot.

[brucek]

I do notice that it sound different back by the bar area than at my seated spot.

I wonder if those drinks you had at the bar had any effect?

Ha Ha - Well I get much fewer large dips with the sub on the right side of the room.

I played with the EQ settings all night.

I'm obviously new to all this. I didn't realize I needed to calibrate the SPL each time I loaded new settings to the EQ. My level changes quite a bit.

However, each time I take the recommended settings and load teh filters and retest, the curve is flatter, but is significantly above the target line. I recalibrate the SPL before I measure and then have the software set the target.

Am I missing something?

Also, can someone point me to a good yet not perfect response graph. I'm not looking for perfection, but I would like to understnad how many db on each side of the target is considered acceptable.

Thanks,

Donald

[brucek]

Am I missing something?

No, not at all. Every time a filter is added, the overall level as measured by band limited pink noise will change. You do indeed need to recalibrate. Only takes a second though.

how many db on each side of the target is considered acceptable

Depends. If you have a narrow spike, then you usually leave it alone because you won't hear it. If you have a wide peak, then you can usually let it go when it gets around 3dB. It all depends on the frequency and the bandwidth of the peak.

Post your graph and we'll let you know.... use a vertical from 45dB to 105dB and a horizontal from 15Hz to 200Hz.....

brucek