Doing the REW and BFD thing - general discussion and questions along the way

[Mitch G]

So, I finally have all the bits in place to do the REW and BFD thing and thought I would start this thread to document my journey and to ask for help along the way.

Equipment:
SVS 25-31 pc+ subwoofer
Paradigm Atoms all around
Paradigm C170 center
Yamaha HTR-5640 receiver


Mitch

[Mitch G]

The sub allows for different tunes. So, ran sweeps with both the stock 25Hz tune and the 20Hz tune.
The graphs are below.

Note the Yamaha has a hard-coded 90Hz cutoff. The big hump around 43Hz is not a surprise since I noticed it when I did the Avia calibration.

I think I'm going to stay with the 20Hz tune (bottom graph).


Mitch

[Mitch G]

So, does this look reasonably BFD-able?

I assume I'm going to put some cuts (is that the right term?) for that 43Hz hump.

But, how about that dip around 56Hz? How do you think I should address that?


Thanks for reading,


Mitch

[Gard]

Mitch G wrote:

So, does this look reasonably BFD-able?

I assume I'm going to put some cuts (is that the right term?) for that 43Hz hump.

But, how about that dip around 56Hz? How do you think I should address that?


Thanks for reading,


Mitch

I would raise the level so the dip hits your target curve and then filter down the rest.

[Wayne A. Pflughaupt]

So, does this look reasonably BFD-able?

Piece of cake. Shouldn’t take more than 3-4 filters.

Regards,
Wayne

[Mitch G]

A few questions:

After posting those charts, I went through another REW sequence where I basically started from scratch. This second time around, the target response that was calculated was pushed up to something like 86dB from the 83 shown in the graphs.
What is going on with that? If I increase the sub level again in response to Gard's suggestion, won't this number keep going up because things will be recalibrated to this higher SPL?

Also, if I go with Gard's suggestion of raising the level and then filtering down the peaks, do I increase the level on the receiver or the sub?
BTW, it might be worth noting that in my attempt to set up the input level to the BFD, I had a of a time causing more than 3 LEDs to light up. I used Darla Tapping (Nemo), Sock Explosion (Monsters), Pod Rising (WoTW) and Omaha Beach (Saving Private Ryan).
But, I do have a bunch of "room" left on the sub's gain control, so I'm guessing I should increase the level there.

BTW, I did some quick sweeps in different listening positions and I was surprised at how deep that dip at 63-ish Hz got in some places.


Mitch

[brucek]

What is going on with that? If I increase the sub level again in response to Gard's suggestion, won't this number keep going up because things will be recalibrated to this higher SPL?

Turn up the sub, calibrate REW and lastly reduce your "target" manually. This will put the target further below the actual measurement and will engage more filter recommendations.

Also, if I go with Gard's suggestion of raising the level and then filtering down the peaks, do I increase the level on the receiver or the sub?

Doesn't really matter at this point, you're just trying to set filters, not calibrate the final system levels.

BTW, it might be worth noting that in my attempt to set up the input level to the BFD, I had a of a time causing more than 3 LEDs to light up.

Were you monitoring the input level to the BFD?
Did you play a bass heavy movie?

Remember, the input level setup for the BFD is simply adjusting the trim level of the receivers subwoofer output in relation to the rest of the speaker levels. The main influence on the input level is of course the receivers main volume control that is at the mercy of how loud or soft you listen to your system.

Setting the input level to the BFD is simply an exercise to ensure you don't have an input level feeding the BFD that clips it when listening to a movie at its loudest level, and conversly checking that you have as high a level as possible feeding the BFD when listening to a movie at its loudest level. This simply means moving the trim level around a bit to have the most efficient level feeding the BFD.

BTW, I did some quick sweeps in different listening positions and I was surprised at how deep that dip at 63-ish Hz got in some places.

Wait on that one and see if your mains clean up the dip.........

brucek

[Wayne A. Pflughaupt]

But, I do have a bunch of "room" left on the sub's gain control, so I'm guessing I should increase the level there. BTW, it might be worth noting that in my attempt to set up the input level to the BFD, I had a of a time causing more than 3 LEDs to light up.

What brucek said. Since the level you’re getting on the BFD meter is so low, reduce the subwoofer’s level control and crank up the receiver’s sub output. The BFD doesn’t have an input level adustment like other equalizers, so you have to send it a hotter signal from the receiver.

Regards,
Wayne A. Pflughaupt

[Mitch G]

Wayne A. Pflughaupt wrote:

What brucek said. Since the level you’re getting on the BFD meter is so low, reduce the subwoofer’s level control and crank up the receiver’s sub output. The BFD doesn’t have an input level adustment like other equalizers, so you have to send it a hotter signal from the receiver.

Regards,
Wayne A. Pflughaupt

I'll be the first to admit that I don't know nothing about this stuff - so bear with me here.
If I have room on the sub's gain dial to go up or down at this point, do I really need to increase the receiver's output to the BFD? I guess I was thinking that pumping up the output from the receiver to the BFD would only add possible distortion to the signal. And, then what if I can't turn down the sub's gain control any further when it comes time to recalibrate and make sure everything is hunky dory?

Am I making any sense?


Mitch

[Wayne A. Pflughaupt]

Hey Mitch,

If I have room on the sub's gain dial to go up or down at this point, do I really need to increase the receiver's output to the BFD?

Well, it’s not exactly a “horrible things are going to happen” scenario, but it will maximize signal-to-noise ratio. Here’s how our digital circuits guru brucek explained it in another thread on this forum:

The BFD1124 is a 24bit device. This means the device theoretically can resolve 16777216 discrete voltage steps (translates to ~ = 6.125x24=147db). Well, that's the theory, but another 1124 spec is that the noise unweighted is >94dB. That's a little better than 15 bits of resolution. So about 9 bits (of the 24 bits) are lost in the noise - you have no contol of this and it's fixed. That leaves ~15 bits to define the input voltage levels or a possibility of 32768 voltage steps as long as you supply the input with a full scale signal.

You can't supply that full input signal if you've introduced gain in the BFD [i.e. boosting filters] or you'll clip the signal at the frequency where you added the gain, so you must subsequently turn down the master input level (of all frequencies) to compensate and so lose dynamic range - and so on with more increased gain. Remember that the voltage signal feeding the BFD isn't subject to all those peaks and troughs you see on the REW screen. The level is perfect into the BFD.

brucek was talking about adding boosting filters, but the same applies with the input signal in general: i.e., you want to get it as high as possible to optimize the BFD’s performance. That said...

I guess I was thinking that pumping up the output from the receiver to the BFD would only add possible distortion to the signal.

... You don’t want it so high that the red clip light at the top of the meter is hitting. Below that point you won’t get any distortion. Actually, you might want to keep it a little further down, to the –6 or –3 dB light, as my crude experiments have shown that the meter probably isn’t too accurate with the kind of broad-band low frequency signals that action flicks generate. See my posts on that thread linked above for more details.

And, then what if I can't turn down the sub's gain control any further when it comes time to recalibrate and make sure everything is hunky dory?

Well, that would be unusual, but obviously you’d have to reduce the receiver’s sub out level.

Regards,
Wayne A. Pflughaupt

[Mitch G]

Wayne A. Pflughaupt wrote:

Hey <snipped>

... You don’t want it so high that the red clip light at the top of the meter is hitting. Below that point you won’t get any distortion. Actually, you might want to keep it a little further down, to the –6 or –3 dB light, as my crude experiments have shown that the meter probably isn’t too accurate with the kind of broad-band low frequency signals that action flicks generate. See my posts on that thread linked above for more details.

The distortion I was (am?) concerned about would be that caused by my receiver being asked to pump out as high a sub volume as would be necessary to get the BFD to show some yellow (or red) LEDs during bassy moments.
In other words, I'm worried about distortion introduced by the receiver working at the higher end of it's operating range.

But, maybe that's not a valid assumption on my part? Do receivers not introduce more distortion at higher volumes?


Mitch

[Wayne A. Pflughaupt]

AFAIK there should be no distortion from a pre-amp output signal. A pre-amp isn't overdriven as easly as an amplifier. A pre-amp has a maximum voltage output capability (given as 0-dBV, “V” being volts). The volume control is actually an attenuator that reduces the signal from maximum. That’s why their settings typically are given as a negative figure – i.e., so many dBV below maximum output.

That’s the theory anyway. In practice, a pre-amp can be overloaded by its input source, which theoretically could happen with a bass-heavy DVD. I can’t say I’ve ever heard anyone say they were having a problem with it, so I’m assuming that receivers have enough headroom built in. Come to think of it, that may well account for the overall lower level you get with DVDs compared to other sources, to make sure the pre amp sections don’t clip – which is the last place you want clipping to happen. If the amplifiers overdrive, you can simply turn down the volume. If the source clips the pre-amps, it will sound nasty at any volume.

Bottom line, there’s no danger from clipping from the sub outs. brucek knows a lot more about electronic circuitry than I do – maybe he can chime in here if I’ve missed it somewhere.

Regards,
Wayne A. Pflughaupt

[Mitch G]

That makes sense. Thanks.

So, I'll basically start over and go through the following steps:

1) Set subwoofer volume on receiver as per the BFD guide using bassy movie scenes - even if it means I have the receiver set for maximum subwoofer volume.

2) Do a basic audio Avia-type calibration - just so I know the slate is clean, so to speak.

3) Do the set input volume, measurement level, meter calibration stuff in REW.

4) Manually set my target response curve for 80dB. (I'm not going to play much with house curves, etc. yet.)

5) Do a sweep. Enjoy the pretty picture.

6) Increase the subwoofer's gain on the subwoofer to push the frequency response curver over my target response curve as necessary so that (non-boosting) filters can be implemented to drive a measured curve that matches the target response curve.

7) Implement the filters.

8) Sweep again and marvel at the near perfect measured response and tweak the filters as necessary.

9) Get all and try to make the measured response curve perfect by adding more and more filters until everything is totally messed up; realize I'm being obsessive and go back to step 1. j/k

10) Skip step 9 and run measurements that include the mains. (I'll have more questions when I ge to this step, I'm sure.)


Mitch

[brucek]

The distortion I was (am?) concerned about would be that caused by my receiver being asked to pump out as high a sub volume as would be necessary to get the BFD to show some yellow (or red) LEDs during bassy moments

Are you sure you have the BFD operating level switch in the -10dBV position and not +4dBuU position?

I see no problem with increasing the trim on the Sub out channel on the receiver. It has an allowable operating range that they provide for you to use...

brucek

[Mitch G]

brucek wrote:

Are you sure you have the BFD operating level switch in the -10dBV position and not +4dBuU position?

brucek

To paraphrase Mr. Conductor, "you are a very clever engine."

Those buttons were not pushed in. They are now. Can someone explain in layman's terms the differences between the two settings?


Mitch

[Wayne A. Pflughaupt]

Pro gear operates at the higher +4 dBu level. Home gear uses the lower –10 dBV level, so the switch is compensation for the kind of system you’re using it in. Is that layman enough for ya?

Regards,
Wayne

[Mitch G]

So, with the +4 dBU setting selected, the BFD was essentially expecting a hotter input signal. Since it was getting a wimpier "home" signal, only a couple of LEDs would light up.


Mitch

[Mitch G]

Time for some graphs.
I went through all the steps again.

Here is the measured without any filters plus the REW predicted corrections.

[Mitch G]

Here is the measured response after applying the filters - sub only.
I still have that dip around 55Hz.

[Mitch G]

Here is measured with sub + the 2 front speakers.
(I can't find a way to add in the center or surrounds short of rewiring the system.)
BTW, I had to turn down the overall system volume by 5dB to move the measured line down to the target response line.

[Mitch G]

So, I think it looks halfway decent for a first time.
But, I don't know what to do about that dip around 55Hz.
Should I worry about it? Should I increase the sub volume to bring that dip up to the line and then filter down the rest?

Am I seeing in these results indications that my precious little Atoms are not quite up to the job?

Thanks,


Mitch

[brucek]

Here is measured with sub + the 2 front speakers.
(I can't find a way to add in the center or surrounds short of rewiring the system.)

Looks good. You don't add center and surrounds. We are only interested in the integration to the mains.

But, I don't know what to do about that dip around 55Hz.

That's not a dip, it's a pimple that you needn't concern yourself with.

Really, it looks quite good.

Now, you can use your receiver test tones to do any speaker level matching - but if you want to adjust the sub up or down, use the subwoofer amplifier controls instead of the receiver trim.

brucek

[Mitch G]

Bruce: Thanks for the input. I won't concern myself with that "pimple," then.

Wayne and Gard: Thanks for following this thread and providing your input.

Now, I guess I should start thinking about that there house curve stuff.

Thanks again,


Mitch

[Mitch G]

I forgot to ask about something the other day.
I ran some sweeps that went up to 5KHz, for example, and noticed the graph went all crazy (crazy = not smooth at all) at the higher frequencies.
Is this just a limitation of the SPL meter? A fact of life for higher frequencies?

I don't think I saved any of those graphs, but it looked like a seismograph readout for a pretty big earthquake.


Mitch

[JohnM]

Mitch G wrote:

I forgot to ask about something the other day.
I ran some sweeps that went up to 5KHz, for example, and noticed the graph went all crazy (crazy = not smooth at all) at the higher frequencies.
Is this just a limitation of the SPL meter? A fact of life for higher frequencies?

I don't think I saved any of those graphs, but it looked like a seismograph readout for a pretty big earthquake.

Mitch

That's how the raw response actually looks due to the effects of reflections from the room's surfaces causing comb filtering. The notches are mostly too narrow to be perceived, for something closer to what you hear try applying 1/3 octave smoothing.