[Nick]

I am hoping for guidance about how to create a smooth freq. response and how to understand what should be changed to improve a waterfall graph. The speakers are full range two channel for music only. If someone has an opinion about if I can get a flatter response by integrating a sub, I would appreciate the input please.

The following graph shows 1) No EQ in gold and 2) two attempts at creating a smooth curve using 1/6. The waterfalls are both with EQ. I can't seem to lower one peak without creating a null or a different peak. This EQ was done with five filters. I would very much appreciate any help regarding what I might try next.

Thank you,
Nick

 

[glaufman]

Well, the very first thing I would suggest is removing all smoothing from your curves. It shouldn't be needed in this frequency range, and simply obscures a lot of what's going on. Reposting your curves without smoothing will help people formulate suggestions to help you.

 

[Nick]

Okay, thank you for your response Glaufman . Do I need to do something with the waterfalls also? Here is the unsmoothed chart

 

[glaufman]

OK that looks better. Now I would suggest going back to no EQ and scanning the two speakers separately. That way we can see if you need to play with placement options before EQ,

 

[Nick]

Okay, here is the right speaker in red, left in green:

 

[Wayne A. Pflughaupt]

I am hoping for guidance about how to create a smooth freq. response and how to understand what should be changed to improve a waterfall graph.

Primarily, waterfalls can only be improved with low frequency absorbers like bass traps. If you have a peak that’s caused by a room mode, EQ can improve the ringing at that frequency, but only to the point of bringing it back in line with what is the norm for the room.

The speakers are full range two channel for music only. If someone has an opinion about if I can get a flatter response by integrating a sub, I would appreciate the input please.

Generally adding a sub isn’t really about achieving flatter response, but better extension. That said, in many situations better response can be achieved because the sub can be placed in a location that delivers better bass response. You usually can’t place full range speakers in the location for best bass, because locating them to achieve the best sound stage is a higher priority.

The following graph shows 1) No EQ in gold and 2) two attempts at creating a smooth curve using 1/6. The waterfalls are both with EQ.

What kind of equalization are you using, something built in to a receiver or pre-pro, or outboard?

I can't seem to lower one peak without creating a null or a different peak.

If you don’t have the ability to EQ each channel separately, then yes this is what’s going to happen. That may not matter in the grand scheme of things, however. Typically the best way to equalize separate bass sources is to generate a response reading with them both operating and equalizing as if they were a single unit. But like Greg said, it's always best to try experiment with placement first, if that option is available to you.

Regards,
Wayne

 

[Nick]

Hi Wayne,

Thank you for your response.

Primarily, waterfalls can only be improved with low frequency absorbers like bass traps. If you have a peak that’s caused by a room mode, EQ can improve the ringing at that frequency, but only to the point of bringing it back in line with what is the norm for the room.

I do have bass traps that are very helpful. When I understand what ringing looks like, I'll try measuring with and without the traps to see the difference.

What kind of equalization are you using, something built in to a receiver or pre-pro, or outboard?

I am using the FBQ, so I can EQ each channel separately. I may switch to the DSP because it allows for settings to be saved at different seating locations. I don't think the FBQ does that, but I also don't know if the DSP has right/left capability. But, next, I'll try doing separate channel EQ with the FBQ.

I have found that what REW predicts the response will be actually measures at much lower levels than what was predicted -- is this perhaps due to a need to change the levels after finding the peaks?

I have been moving the speakers around and found a spot that minimizes some of the nulls. I don't think I can do much more via placement, though I haven't tried moving them closer together yet (they are about nine feet apart and just over two feet from the side walls and five feet from the back wall).

I have a sub I can use, but I'm not sure I need it. The EQ appears to be increasing the low bass -- I bought the sub thinking that I was only going to be able to attenuate peaks and would need the sub for from about 50hz on down.

 

[brucek]

I have found that what REW predicts the response will be actually measures at much lower levels than what was predicted -- is this perhaps due to a need to change the levels after finding the peaks?

Yes, once you remeasure after the filters are introduced, you have to redo Check Levels and Calibrate SPL routines.

brucek

 

[glaufman]

Are these individual speaker graphs in you original placement or the better one you found? Is the EQ on or off?

 

[Wayne A. Pflughaupt]

I am using the FBQ, so I can EQ each channel separately. I may switch to the DSP because it allows for settings to be saved at different seating locations. I don't think the FBQ does that, but I also don't know if the DSP has right/left capability. But, next, I'll try doing separate channel EQ with the FBQ.

The DSP1124 isn’t the best choice for using on the mains, as it’s a bit noisy. It’s fine as a dedicated subwoofer EQ, though, and it does have separate L/R capability.

As for equalizing each channel separately, keep in mind that above about 400 Hz L/R filters need to be matching. If not, it can do weird things to the imaging. Below that point it’s fine to use separate L/R filters, especially for the bass frequencies. However, many people find that trying to EQ two bass sources to be an exercise in frustration. In the end it may be best to EQ the bass frequencies the same way as the upper frequencies, with matching filters.

Regards,
Wayne

 

[Nick]

Are these individual speaker graphs in you original placement or the better one you found? Is the EQ on or off?

Yes, the graph in red (right) and green is each individual speaker measured without EQ after moving them around some. I find that each one wants to be in a bit different place. This may be due to a large opening on the left wall about 2/3 of the way back and a five foot square skylight in the room.

As for equalizing each channel separately, keep in mind that above about 400 Hz L/R filters need to be matching. If not, it can do weird things to the imaging. Below that point it’s fine to use separate L/R filters, especially for the bass frequencies. However, many people find that trying to EQ two bass sources to be an exercise in frustration. In the end it may be best to EQ the bass frequencies the same way as the upper frequencies, with matching filters.

My intention is to only use the EQ in the under 200hz frequency range, so any left/right EQ would be done in this range. The speakers are bi-amped, so I'm hoping any noise is minimized in the m/h frequencies, though the speakers cross over at 90hz (they are a 1st order crossover, so there is overlap).

I was thinking about adding a sub, but it would require a Y-adapter to be shared with the bass output for the speakers. It sounds like a might create a problem because then I would be equalizing the lower mains and the sub together.

Thanks to each of you for your input.

Nick

 

[Wayne A. Pflughaupt]

I was thinking about adding a sub, but it would require a Y-adapter to be shared with the bass output for the speakers. It sounds like a might create a problem because then I would be equalizing the lower mains and the sub together.

Well, if you add the sub, you'd want to limit the main's bass so that they don't overlap (read duplicate) with what the sub is doing.

Regards,
Wayne

 

[Nick]

Well, if you add the sub, you'd want to limit the main's bass so that they don't overlap (read duplicate) with what the sub is doing.

Wayne, I have found that if I use a filter in the 28hz area the speakers will produce bass that emulates your hard-knee house curve down to 20hz. So, it appears that adding a sub is not necessary to increase extension. I was thinking I would need the sub to get good bass below 40hz, thinking EQ was only practical for attenuating peaks.

I have a second listening position in the room and I plan to do measurements to optimize the sound at the spot. As far as I can tell, with the FBQ I would have to set a different batch of filters and then turn the gain to 0db for the seating position I was not using. This would mean I would have to reset the filters for each spot when I changed seating locations. I read in the house curve section in the forum that the DSP1124P will save inputs for different settings that allow one to switch to one group of filters or another without having to reset the gain each time. Can you please verify if this feature is available on the DSP1124?

Thank you,
Nick

 

[Wayne A. Pflughaupt]

I read in the house curve section in the forum that the DSP1124P will save inputs for different settings that allow one to switch to one group of filters or another without having to reset the gain each time. Can you please verify if this feature is available on the DSP1124?

That’s partially correct. The BFD does have memories, but there is no overall gain control.

To explain, traditionally analog equalizers had a level control that was meant as an amplitude compensation tool. So, if your equalizing resulted in an overall hotter signal at the output compared to the original input signal, the level control would allow you to reduce the output signal to match the input. And the same for equalizing that resulted in the output being lower than the input.

The BFD does not have this feature. If your EQing results in an overall increase or increase in output, then that’s what you have. There’s no compensation. So in your situation, wanting to use a separate bank of filters for different locations, you may well end up by default changing the sub’s volume level when you switch the BFD to a different memory setting.

Regards,
Wayne

 

[Nick]

I spent a few hours today moving the my speakers around. The result seemed to be that they wanted to be in different spots as far as reducing nulls. I found a compromise spot, but it doesn't appear to be much different than what I had before. Here are the right (red) and left measurements in the final spot I settled on:



The third graph shows the results from last week (pink) done in stereo and today's result with EQ done separately in each channel. Included in the result are my attempts at adding levels where I had nulls. I'm not clear on if there is a recommended way to go about altering what REW recommends, particularly with regard to all the changes that occur when one tries to raise levels?

I guess I've come away with the feeling that I just altered stuff from before but didn't really improve upon it. So I'm hoping again for what folks think I ought to be trying to do next? Below is the new waterfall as compared to those from the earlier post (sorry the graphs seem to be getting rearranged out of order).

Nick

 

[Wayne A. Pflughaupt]

Sorry for not mentioning this previously, probably would have saved you a lot of time and trouble. :hide: Since you have the capability of equalizing your main channels, I wouldn't worry about locating the speakers for best frequency response. What I'd suggest instead is locating the speakers where they give you the best imaging, and then equalize whatever response you get that is less than ideal.

Regards,
Wayne

 

[Nick]

Thanks, Wayne. I didn't mind moving them around, it was interesting to me. The imaging is pretty good mostly wherever I have placed them, so I am happy with the imaging. I would like to become more adept at setting filters, though! When I look at the first waterfall graph vs. the new one, it seems that the first one looks 'better'...

 

[Wayne A. Pflughaupt]

I spent a few hours today moving the my speakers around. The result seemed to be that they wanted to be in different spots as far as reducing nulls. I found a compromise spot, but it doesn't appear to be much different than what I had before. Here are the right (red) and left measurements in the final spot I settled on:

Hmm... Well, it looks identical to your first graph:

I would like to become more adept at setting filters, though! When I look at the first waterfall graph vs. the new one, it seems that the first one looks 'better'...

Some of my colleagues here have a different view, but I don’t take much stock in time domain equalizing (i.e. equalizing for waterfalls). See here for some thoughts I recently posted on the subject. IMO The goal of equalizing is to improve sound quality. If you are concerned about reducing signal decay times, you need some bass traps.

I would like to become more adept at setting filters, though!

Have you tried using REW’s RTA feature? You can adjust filters and see the changes in response in real time.

Regards,
Wayne