A few questions on speaker calibration and REW results

[HereToAskQuestions]

Hello all, thank you in advance for your consideration and help! I am trying to calibrate my 7.1 home theater system using REW and a Umik-1. The audio pathway is PC via HDMI to NuForce A/V Processor 18 to ADA TPM 1645 amp for speakers and Behringer A800 for subwoofer. I am in a small NYC apartment living room.

1. I have taken my first measurements from the speakers. On REW I'm using the asio4all output to isolate each speaker channel individually. The AV Processor is supposed to have a 24db high pass filter at 80 Hz for these speakers. When running the measurement I used the range 20-20K Hz which based on REW's manual says it sweeps from 10 to 40K I believe. Anyhow, The resulting measurement for all the speakers has a large SPL peak at 50 Hz (see attached screenshot). This is unexpected to me as to my understanding the digital high pass filter (i.e., crossover) in the AV Processor should stop signals from coming through at that low a frequency (or at least heavily attenuate them). When a use REW to generate EQ filters I plan to program back into the AV Processor (has built in 11 band PEQ for each channel) most of the filters are used to decrease this large 50 Hz SPL peak leaving less for correcting signal within the frequency range the speaker is actually supposed to carry (80 Hz +). Questions:
   1. Why are my speakers producing a large SPL at 50 Hz even though they supposed to be high pass filtered at 80 Hz, is this some room/speaker resonance that's amplyifying what should be an attenuated output at that frequency?
   2. I wouldn't think the speaker could even do that low a frequency (only 6 inch woofer) so is there any chance some of the signal in the test is leaking into the subwoofer?
   3. When I run a calibration sweep, is it correct to sweep 20 to 20Khz for the speakers or given that they're high pass filtered at 80 hz should I use that as the bottom range.
   4. When fitting the PEQ filters to the measurement, should I try to match the full 20 to 20Khz range or should I try to match only the 80 to 20 Khz range the speakers are intended to play (or maybe even a smaller range than that to focus on a specific type of sound source)?
2. Second question is in regards to how the number of speakers vs the number of subwoofers is factored into setting SPL levels for each speaker/subwoofer. I know common practice is to set all speakers and subwoofers each at 75 dB. However, given that I have 7 speakers, and the subwoofer needs to carry the bass audio for all 7 of those speakers at once, shouldn't the subwoofer SPL need to be increased to match the increase SPL in the high frequency range produced by have mutiple speakers playing at the same time. To clarify my thought process let me propose a situation with mono audio to be simple: if I have one speaker and one subwoofer with a high and low pass crossover respectively, both are 75 dBs and therefore the response is flat from bottom to top. Now if I add another speaker, with that speaker individually being 75 dB, now the two speakers together put out 78 dB (+3dB for doubling amount of speakers) in this setup then should the subwoofer not be increased the same +3 dB so to maintain the flat ratio of SPL between low and high end.
   1. The practical implication I'm wondeirng about is this: when I listen to music on my system I put it into "all stereo" mode where all the left speakers play the left channel, all right speakers play right channel, and center speaker plays mix of L/R. In this mode, does my AV Processor automatically increase the volume of the subwoofer signal to account for the fact that the left and right high pass speaker channels are now louder because there's at least 3 speakers playing the same signal for each channel (vs 1 sub playing the low end for both L and R channels). OR do i need to set my subwofer SPL level higher than 75 dB to account for the mutiple 75 dB speakers it's paired with.
   2. I'm hoping the answer is "calibrate each individually to 75 dB and then your audio/video Processor does the magic of calibrating SPL levels to sub and speakers based on the playback mode to account for all this.
3. When creating PEQ filters how much boost is allowable? My amp can put 100 watts into each speaker, whereas the speakers max RMS is 60 watts, so that menas i have a lot of headroom left? In practical terms, the AV Processor is never turned above 40% volume as it'd be unbearably loud in the apt above that given small space and decently high sensitivity speakers. With that in mind, is more PEQ boost allowable for my situation or should I still limit it to 3 db (or another value)?

Really appreciate any insight and help anyone can offer. Feel free to just answer any part individually!

 

[HereToAskQuestions]

4. One more question, should i just do the process of measure / create filters once. Or should I be repeating it mutliple to check the impact of my filters and verify they're working?

 

[Wayne A. Pflughaupt]

Why are my speakers producing a large SPL at 50 Hz even though they supposed to be high pass filtered at 80 Hz, is this some room/speaker resonance that's amplyifying what should be an attenuated output at that frequency?

The attenuating is for the main channel speakers. The attenuated frequencies (below 80 Hz) are sent to your subwoofer. Thus, the 50 Hz peak is in the subwoofers, not the main speakers.

When a use REW to generate EQ filters I plan to program back into the AV Processor (has built in 11 band PEQ for each channel) most of the filters are used to decrease this large 50 Hz SPL peak

It sounds like you are using the main-channel equalizers to reduce the 50 Hz peak. It’s not coming from the main speakers. Thus, eliminate all those filters and re-program a filter in the subwoofer channel only.

Second question is in regards to how the number of speakers vs the number of subwoofers is factored into setting SPL levels for each speaker/subwoofer. I know common practice is to set all speakers and subwoofers each at 75 dB.

Actually, the sub is typically calibrated 10 dB higher than the main channels. But that’s just a “hard and fast” rule. In reality, you should adjust the sub level with music for what sounds like a natural blend with the main L/R speakers.

if I have one speaker and one subwoofer with a high and low pass crossover respectively, both are 75 dBs and therefore the response is flat from bottom to top. Now if I add another speaker, with that speaker individually being 75 dB, now the two speakers together put out 78 dB (+3dB for doubling amount of speakers) in this setup then should the subwoofer not be increased the same +3 dB so to maintain the flat ratio of SPL between low and high end.

This is functionally correct. I have noticed with music that when I switch from regular stereo to all-channel stereo, the sub sounds weaker, for the reasons you stated.

However, people typically calibrate their sub levels based on stereo music.

Movies are a different story. Typically, movies don’t play all channels at the same level simultaneously. But when they do, the producers boost the sub signal accordingly. Thus, no compensation is needed on your part.

But if you listen to music in 5-channel stereo a lot, you need to calibrate your system / subwoofers for that. Your processor is not going to do it for you.

When creating PEQ filters how much boost is allowable?

The myth that EQ boosting is bad just won’t die. In reality, EQ cuts often end up taxing the system as much as positive-gain filters. Read the article linked below for more details. I typically don’t like to see more than about 8 dB in total EQ adjustment (the total of the filter boosted the most + the one cut the most), but the short answer is, “as much as you have headroom for.”

should i just do the process of measure / create filters once. Or should I be repeating it mutliple to check the impact of my filters and verify they're working?

If you’re manually adjusting, it’s best to check each filter with another sweep. Manually adjusting is tricky, especially if you don’t have an equalizer that is supported by REW.

What I typically do after my initial sweep is to identify a peak or depression I want to address, and note the “outer limits” (i.e. where the response starts to deviate up or down). Typically, it’s not terribly hard to determine the amount of boost or cut – the graph tells you that. The trick is figuring the appropriate bandwidth. The filter bandwidth is determined at half the “width” of the depression or peak.

For instance, if the total bandwidth of the problem area is 2/3-octave wide (in the example of a peak, from the point where response starts to rise, to the point where it comes “back in line”), then the filter bandwidth will be 1/3-octave. Thus a 1-octave problem uses a 1/2-octave filter, and a half-octave problem requires a 1/4-octave filter. You probably won’t want to use filters tighter than 1/4-octave.

So – apply your filter, take a measurement and take a look at it, and pay attention to those outer parameters. If the “outer limits” have changed after EQ, then the filter is too wide. Re-adjust the bandwidth, take another reading. Rinse and repeat as needed, then move on to the next issue.

Assuming speakers of good quality, you shouldn’t need more than 2-3 filters for each channel (the sub may require a couple more). Don’t be chasing every little ripple in response. Fixing the worst problems will get the most audible result. Typically the worst problems are the ones that are wide and/or deep.

After EQing, do some listening with your favorite reference music program, and if your processor allows, switch each filter in and out. If removing a filter gets no audible improvement, there’s no reason to keep it. At the end of the day, trust your ears, not the graph. Can’t tell you how many times I’ve seen that certain filters did not make an audible improvement, or even worse, sounded – well, worse – than with it eliminated.

I do make allowances for the center channel that I don’t for the main L/R. For instance, shelving EQ I can switch in or out for program material with insufficient or exaggerated vocal sibilants, the goal being better intelligibility of voices. I also typically high-pass the center channel rather high, say 200 Hz, because often male voices are unnaturally and overly boomy.

Myths and Legends Concerning Positive-Gain Filters when Equalizing Subwoofers

Myths and Legends Regarding Positive-Gain Filters when Equalizing Subwoofers Disclaimer: The following relates to the use of outboard, stand-alone equalizers such as the miniDSP and other equalizers and digital speaker processors from the pro audio field. Reduced Dynamic Range and...

www.avnirvana.com

Regards,
Wayne

 

[ddude003]

Love them old Neve and Manley 3 and 4 band passive analogue EQs... Point being that less is sometimes more... And then there is math and convolution where some is good, more is better and too much is just right...