Theory question

Folks here are two plots that although look very similar, were done at different receiver volumes.



The green plot was done with the volume on my Onkyo 805 at -4.5dB, the volume setting I reached during the Set Levels process for this session. I was trying different subwoofer positions and found a place along the wall where I had to reduce the volume setting of the 805 to -8.5dB to get the blue plot due to clipping of the RS SPL meter.

Sooooooooo, theoretically, what happened? Have I found the place where the room mode is highest? Do I want to pursue subwoofer positions that give me a higher output for a lower volume setting, or look for a smoother curve (I'm not saying the green curve is smoother) even if I need an increased volume setting to achieve a similar output? Would that be akin to avoiding the room modes?

I'm amazed now at how much difference an inch can make (so to speak) in subwoofer placement.

deadhead

[Wayne A. Pflughaupt]

Yeah, I'd try to find a place where response is smoother. You have ~30 dB between your worst peak and depression there, making it essentially unequalizable. I'd like to see that deviation narrowed to something like 15-18 dB, if possible.

Do you have a usable corner?

Regards,
Wayne

Wayne,

You keep asking that corner question... The plot is essentially the same, only the receiver volumes are different.

My room is near square, 14'7" x 14'9". If I understand the physics, the peaks are at the room modes, about 40 and 80 Hz, and the valley at 60 Hz is the cancellation that happens between them. I think I need to wait for my GIK 244s to arrive before I fool around any more.

However, they don't look as bad at 20-130dB. Almost looks like a decent room curve with equalization (IF I understand the theory).

deadhead

deadhead wrote:

Wayne,

You keep asking that corner question... The plot is essentially the same, only the receiver volumes are different.

My room is near square, 14'7" x 14'9". If I understand the physics, the peaks are at the room modes, about 40 and 80 Hz, and the valley at 60 Hz is the cancellation that happens between them. I think I need to wait for my GIK 244s to arrive before I fool around any more.

However, they don't look as bad at 20-130dB. Almost looks like a decent room curve with equalization (IF I understand the theory).

deadhead

You seem to understand the theories fine.... Just my thoughts, on looking, and reading your posts.... I'd agree with you, that the plot you are seeing is the result of the room modes summing and canceling in your space. One of the biggest issues you have, I believe, is the near-square proportions of the room, creating near-identical standing modal frequencies along two different axial room modes (plus their nearly-identical oblique and tangential modes, also.... makes me wonder how high this room is) to sum with each other, and thus, grossly amplifying the modal energy at those frequencies to high levels. So, my thoughts.... if you could re-surface a wall at the end of your room's long dimension, and reduce the room length to 12' from the current 14'9", your summing problem would be much improved. Making this dimensional adjustment would reduce both the modal peaks at 38 and 77 Hz, as well as the cancellation notch near 60 Hz, by likely more than 6 dB each, which would in turn allow you to equalize and/or treat the room to level its response. By setting this new wall surface at an angle of 5 or so degrees off-parallel to the opposite wall, your room's incidence of standing waves at frequencies above 2 kHz will be reduced, also.

A point to consider: if you do add this new wall surface, make sure there is limited free resonant space left behind it (i.e., fill the cavity behind the wall with mineral fibre or glass insulation batting), or you will replace your first problem with a new one.

If you cannot do this, placing some corner bass traps in two opposing room corners, or all four, perhaps, given your room's serious dimensional tendency towards summing modes, and you may be able to reduce some of the low-frequency energy issues you are experiencing, through absorption of the modal frequencies, rather than limiting their source.

So... my two cents' worth....

Maybe I'm doing something fundamentally wrong. I've got an RS 33-2050 SPL. When I "Check Levels" on the soundcard at each session, is there an eyeball correction required when I'm comparing the RS meter to the reading on REW? I'm wondering if I have too much sub because I have too much gain. The gain knob on the sub is already at 9 o'clock. I think it will only go back to about 8 o'clock.

Here is a graph of the mains and the sub taken separately but overlayed.



I have the LPF set to 120 on the Onkyo 805. The crossover is OUT on the VTF2 Mk3. Mains are PSB Century 500i, two-way with an eight inch woofer. Response is 50-20k with LF Cutoff of -10dB @32Hz. Will experimenting with the crossover freq help?

Any suggestions will be attempted, except for adding a wall.

deadhead

[brucek]

Any suggestions

I suggest you don't have the mains crossed at 120Hz for some reason. A setting must be wrong.

Below is what a 120Hz crossover should look like when the signals come from the receiver (until the speakers get hold of it and it's subjected to the room).

Your mains don't look crossed at all.... and the sub looks questionable as to its crossed frequency.



Typical 120Hz crossover


brucek

deadhead wrote:

Maybe I'm doing something fundamentally wrong. I've got an RS 33-2050 SPL. When I "Check Levels" on the soundcard at each session, is there an eyeball correction required when I'm comparing the RS meter to the reading on REW? I'm wondering if I have too much sub because I have too much gain. The gain knob on the sub is already at 9 o'clock. I think it will only go back to about 8 o'clock.

Here is a graph of the mains and the sub taken separately but overlayed.

Attachment 8068

I have the LPF set to 120 on the Onkyo 805. The crossover is OUT on the VTF2 Mk3. Mains are PSB Century 500i, two-way with an eight inch woofer. Response is 50-20k with LF Cutoff of -10dB @32Hz. Will experimenting with the crossover freq help?

Any suggestions will be attempted, except for adding a wall.

deadhead

Sorry about the wall.... but, I believe that change would help your room's response considerably. Nonetheless, looking the last two plots, I'd agree about the crossover not being set correctly on the mains - the mains response is down to just about 40 Hz, which is far below the suggested cross frequency. Hardware? Settings? Choice of inputs/outputs used?

Cheers,

Gents,

My understanding is the Low Pass Filter (LPF) is not the same as a crossover. The crossover in these plots is set at 60Hz. Sorry, I should have mentioned that, but it was early. On the Onkyo 805, I can choose a crossover separately for all the channels (channel pairs), as well as the setting for the LPF. The Audyssey MultEQ XT equalization gurus recommend an 80Hz crossover for the mains and the LPF set at 120Hz. I was playing with the crossover and set it at 60 because the woofers on the mains are pretty good sized. However, if the response is 50-20k Hz with an LF cutoff of -10dB at 32Hz, where would the theoretical correct spot be for the crossover? I can choose from 40Hz all the way to 120Hz.

I have not played with the LPF setting, nor have I played with actually setting a crossover freq on the sub. All I did to plot the mains was turn off the sub. However, I did turn Audyssey off for all these measurements. To get the sub plot, I pulled the banana plugs from the speakers themselves.

I spent the afternoon changing the Phase at the same subwoofer position and measuring the differences, then moving the sub an inch and measured again. Phase 0 moves the dip more towards the high 60s, but doesn't reduce the peak at 80, so they're closer together. It also does weird stuff above 80Hz--lots of spikey valleys not there at Phase 180.

I'm using the front panel AUX 2 input on the 805 to do the measurements.

If I could change rooms I would, but the theater is where it is, and the room is what it is.

Thanks,

deadhead

[brucek]

The crossover in these plots is set at 60Hz. Sorry

Ahhh, that makes a difference. Although you still are a bit off. Here's the 60Hz targets (in a perfect world).

60 Hz crossover target

However, if the response is 50-20k Hz with an LF cutoff of -10dB at 32Hz, where would the theoretical correct spot be for the crossover?

Crossover of 80Hz would be better.....

and the LPF set at 120Hz

Yeah Ok, it really matters not though, most LFE (which I suppose they're concerned with here) is completely gone by 120Hz in the source material anyway..

brucek

[Wayne A. Pflughaupt]

deadhead wrote:

My understanding is the Low Pass Filter (LPF) is not the same as a crossover.

Yes, it is. With an HT receiver, we're dealing with a two-way crossover. One side is a high pass (the mains), the other is a low pass (the sub).

The crossover in these plots is set at 60Hz. Sorry, I should have mentioned that, but it was early. On the Onkyo 805, I can choose a crossover separately for all the channels (channel pairs), as well as the setting for the LPF.

I guess this means you can set the high pass at different frequencies for the main channels, and the low pass seperately for the sub?

Regards,
Wayne

[brucek]

My understanding is the Low Pass Filter (LPF) is not the same as a crossover

In the case of the Onkyo, the LPF (as I referenced above) is an upper limit filter for the LFE channel (read your manual page 88).

The crossover is set individually for left/right mains, center, surround and back pairs. I suggest using 80Hz for them all.

brucek

I had a few minutes to do some experimenting. Here are a series of overlayed plots of the mains only, 1/3 smoothing, changing the crossover setting on the 805. Looking at the 40Hz bump, the top plot is with the crossover at 50Hz, the next is 60Hz, then 70, then 90, then 100, then 120Hz. For some reason, I don't have 40Hz and 80Hz saved, but they fit the pattern.



Is this what you would call a traditional crossover, or it it something else? The LPF of the sub was still set to 120 on the 805. If I ran this series again with the LPF at, say 40Hz, would I see a difference here in the mains, or only on the sub side?

Thanks,

deadhead

[brucek]

Is this what you would call a traditional crossover, or it it something else?

It's not suitable to evaluate a crossover of a receiver with speaker measurements in a room. There is not only the rooms effect on the response, but in adition you have the natural rolloff of the mains speaker itself.

If you want to evaluate your crossover, then use REW to feed the AUX input, while the output is taken from the preamp output and fed back to the line-in of REW. Now you've removed the room and the speaker from the equation.

I'm sure the crossover is the same as any other receiver. My advice would be to set it to 80Hz for all channels.

The LPF of the sub was still set to 120 on the 805. If I ran this series again with the LPF at, say 40Hz, would I see a difference here in the mains, or only on the sub side?

Set the LPF of the receiver for sub to 120Hz. It is not a crossover. It is a special filter to remove information above 120Hz for unwanted information above that frequency. The crossover of the sub is set to 80Hz. If you set the LPF any lower it will interfere with the crossover setting.

The LPF filter is only active for sources that generate LFE information (i.e. 5.1 DD).

From your manual.
With this setting, you can specify the cutoff frequency of the LFE channel’s low-pass filter (LPF), which can be used to filter out unwanted hum. The LPF only applies to sources that use the LFE channel.

brucek

Gents,

Thanks for all your help so far.

Back to theory... Are there any Things that are always or almost always true?

For example, is it likely that if a few test runs with REW indicate a smoother response with the Phase set at 180 rather than 0 (remember I have a huge swing between my highest peak and lowest valley) can I assume no matter where the sub is +/-18 inches or so, I will always be better off with Phase at 180?

Or, if I set the sub crossover to 60Hz (or 70 or 80), should I always set the mains crossover at the same freq, or could there be benefit from experimenting with the main's xo setting?

The reason I'm asking is I set down with REW and started playing around. Sub xo set at A, mains xo set at A, Phase set at 0, sub X far from the side wall and Y far from the rear wall and I ran a measurement. Then I changed the Phase to 180. Ran another measurement. Then I changed the xo on the mains 10Hz and ran another. Then I changed the phase back and ran another. Etc., etc., etc. It quickly became apparent I could be at this for weeks. I seem to have gotten the smoothest response with the mains and sub both set at 60Hz crossovers and Phase at 180, in the sub's current position. Should that combo hold, or might moving the sub half-a-foot cuase me to have to go through the whole process again?

Thanks again,

deadhead

[brucek]

is it likely that if a few test runs with REW indicate a smoother response with the Phase set at 180 rather than 0 (remember I have a huge swing between my highest peak and lowest valley) can I assume no matter where the sub is +/-18 inches or so, I will always be better off with Phase at 180?

No, movement can make quite a difference if it happened to be in a null. Remember that phase does nothing on a sub by itself. It affects the interaction between itself and other speakers at the crossover frequency only. Also not ethat if you only have a single zero and 180 degree switch, then sub distance trim in the receiver can be used as a proxy.

if I set the sub crossover to 60Hz (or 70 or 80), should I always set the mains crossover at the same freq, or could there be benefit from experimenting with the main's xo setting?

You're a bit confused here. The sub itself has its internal low pass filter bypassed. There is no ability in the receiver to use a different crossover on a sub and a different one on mains. You simply set the mains crossover frequency and the sub is automatically using that setting. So if you select 80Hz as the mains crossover, then signals above 80Hz go to the mains and signals below 80Hz are directed to the sub.

Sub xo set at A,

Where are you setting a sub crossover frequency? The only setting you have is Yes and No.......

brucek

The Hsu VTF2 Mk3 has a "Crossover Frequency" knob that runs from 30 to 90Hz. I can also switch it OUT. Right now, I have it set at 60 and the crossover in the 805 set for 60 for the mains. The Hsu manual stated if the bass was boomy, to try switching IN the crossover. Seems to have helped.

deadhead

[brucek]

The Hsu VTF2 Mk3 has a "Crossover Frequency" knob that runs from 30 to 90Hz. I can also switch it OUT. Right now, I have it set at 60 and the crossover in the 805 set for 60 for the mains.

This is a pretty big no-no. The internal low pass filter in the subwoofer itself is for use when there is no bass management available in an older receiver, or when using high level inputs.

When the receiver has bass management capabilities, then you always switch out the filters in the sub itself as they will negatively interfere with the crossover mixing. Using the subs own LPF as a quasi-EQ is not an idea that would have much support, whether HSU suggests it or not.

brucek

brucek,

I agree with everything you've said in this thread. Still, it sounds better this way. REW indicated the big peaks at 40 and 80Hz are much reduced. When I switch the Crossover Freq OUT, the 40Hz peak reappears. The plot is still fairly ugly, but nothing like that in post #1.

This whole process is fairly frustrating, fun at times, but frustrating. I learned my room dimensions suck, and I need treatments. REW is pretty cool, but it doesn't show the names of the saved plots (thumbnails) on the left side or in the "Measured" list as what they are saved as, but rather the time they were taken, so I've gotten into the habit of keeping a list. (Is there a way to change how saved measurements are displayed?) The RS SPL clips often when moving the sub around with my peaky room, and if I raise the range one notch, the plot doesn't make any sense. I've tried probably 50 different positions of the sub--an inch at a time--and when the amplitude is largest, the room modes are probably the reason. It doesn't sound right. I think I have the sub in the least efficient place in the room, but it seems the smoothest. I don't have the BFD in the loop yet (I have it and the cables, but...) and I also have Audyssey MultEQ XT built into the 805. If I run a series of REW measurements to find the best place for smoothest output of the sub, I need to reduce the sub gain to near minimum. Then, when I run Audyssey, the sub cannot be measured because the "ambient noise is too high"--she needs more gain. I experimented for a while doing three Audyssey runs to find out what it thought the gain on the sub should be, then tweaking the gain knob to see if I could make that 0dB on the next series of measurements. I do like what Audyssey does overall, but the bass is finicky.

During the Lobby scene of The Matrix, there is some kind of bass music track playing in the background. I don't know if it's an organ or what, but it's pretty deep. With some of the sub positions I tried, that bass music is almost gone at the listening position. Currently, it sounds pretty good.

I think it's a work in progress, but I'd like to thank everyone here, GIK Acoustics 244s, an Auralex Gramma, and REW. I've got an ECM8000. Xenyx 1002 on the way. I'm not going to mess with the setup again till the Xenyx arrives. Wait... that's tomorrow.

deadhead

[brucek]

REW is pretty cool, but it doesn't show the names of the saved plots (thumbnails) on the left side or in the "Measured" list as what they are saved as, but rather the time they were taken

You simply edit in the name that you like for each measurement as shown below, and also the notes you want as shown below.
This information is saved in the mdat file.

The RS SPL clips often when moving the sub around with my peaky room, and if I raise the range one notch, the plot doesn't make any sense.

If the measure is peaky and clips after you moved to a new position, simply redo the Check Levels routine and set a level at the listening position a bit lower. Takes a second.
A microphone will make things a bit easier in that regard, since it doesn't have a needle to clip, although if you have a measure with a very large peak, then REW can clip its input and you do have to turn the REW input down a bit to take a measure such as that.

brucek

brucek,

Excellent! Worth getting up just for that!

deadhead

[Wayne A. Pflughaupt]

deadhead,

While unusual, it's certainly not unheard of to stagger crossover high and low pass points, if there happens to be a response peak in that region.

Regards,
Wayne

Wayne,

I was hoping that would be one of those "always true" Things...

deadhead