Help! 4 subs, REW, minidsp, ARC ; odd results

[Sven777]

Recently replaced 2 Klipsch R-115SW with 4 Rythmik FV18. Gain matched them by REW/UMIK1/SPL measurements all in the same location. Set gain offsets in minidsp.

Then followed general process outlined here to incrementally add all the subs to ensure that each addition resulted in net positive SPL:

Here are my measured results:

Equalized/flattened the response to 76db in REW, imported into minidsp PEQ, results (top line is unequalized full 4subs, bottom is equalized):

Ran ARC and followed steps here to ensure a good house curve and that the final eq'd curves hug the target curves as much as possible:

ARC file after getting audio measurements and before I did any fiddling:

rythmik_initial_setup_post_audio_measurements_before_adjustments.arc3

Adjusted/final ARC file:

rythmik_fully_adjusted.arc3

End result is that new subs sound less boomy/bassy than old subs in exact same room - very "meh". I didn't do any measuring/empirical testing with the old subs when I put them in 5 years ago - just kind of subcrawled a bit and then fiddled with phases until I got something pleasing-ish.

Can anyone spot what I've done wrong? It seems like "on paper" with the empirical tests I've done I should be getting great results, but I don't yet understand.

Another strange thing is that the frequency sweep graphs in ARC for the subs don't look like my REW frequency sweep graphs. Can anyone explain that? My fronts are Founder 120Hs - do their frequency graphs look correct? They seem to have this huge 94db cliff from 30-50hz and then tapers down to around 75db by around 100hz and then stays there. I'm not sure if that's normal.

Thanks for any assistance anyone can provide!

 

[neRok]

Can anyone spot what I've done wrong? It seems like "on paper" with the empirical tests I've done I should be getting great results, but I don't yet understand.

You have probably EQ'd too aggressively, but also you may not have time aligned them properly (and I don't know what ARC actually does). If you have 4 subs all the same, then the fact they don't have the same SPL in your room is because they are activating different room modes in different ways. Check out the first few gifs in the thread Visualizing How Different Loudspeaker LF Directivity Patterns Couple to Room Modes. It shows how a sub being a different distance from a wall activates the same room mode in different ways. Also the plots show how the listening position changes the listeners relationship to the room mode.

Room modes and SBIR cause peaks and nulls in the frequency response. I would suggest using REW Room Sim to find out where your modes are "theoretically". Start with your speakers and your listening position, and see what is happening. Some problems are purely LP related (distance to back wall being the main problem), so try identify those issues. Then turn off the speakers and add a single sub, and see what peaks and nulls that sub has. Maybe tweak its position slightly, to get it as flat as possible. Do the same with the other subs, 1 at a time. The less worse you can get their individual response compared to the listening position, the better time you will then have when time aligning and applying EQ.

 

[Sven777]

but also you may not have time aligned them properly

Hi neRok - you may have missed where I described how I used REW and minidsp to properly time-align the subs to ensure positive FR summation while adding each sub. The first graph in my original post shows the progression of adding each sub with their ideal delays for maximum SPL output. The line at the top is the final FR aggregate of all 4 subs.

I've never gone through a formal calibration process before with FR measurements, but I'm pretty sure I did it correctly. The previous subs I just subcrawled and tuned-by-ear (no room correction) until I got something I liked. I was expecting that when the ARC'ing and house curve was done that the final result would have appreciable bass when watching a movie at low-to-medium listening levels, but it feels like I need to crank up the main volume a fair bit before I get any kind of boomy response. I'm hearing in other places that I just need to increase the sub levels to-taste - which I can certainly do - but I wasn't expecting these results since the house curve is already gradual +7db from 100hz to 20hz. Here are the target curves for ARC - the final EQ'd results (not shown here) hug the curves pretty closely:

When setting up new rooms, is there a normal range of "you'll typically increase sub by ??-to-?? db to make it sound good"? It seems I need to add at least 7db to make it start sounding appreciably "moving" and that just seems like a lot of extra db for "normal home theater results".

 

[neRok]

The first graph in my original post shows the progression of adding each sub with their ideal delays for maximum SPL output. The line at the top is the final FR aggregate of all 4 subs.

So I've seen a "problem" with REW's phase alignment tool whereby it wants to align the "peak energy moments". Thus if one (or more) sources is affected by a reflection (room mode) and you attempt to align that frequencies "peak energy", then it can suggest delaying one source by a great amount. But if you look at what is happening at nearby frequencies, you would see that it has caused a very bad time shift that would be heard as "smeared bass". I show this problem in effect here: https://www.audiosciencereview.com/...ubwoofer-s-to-mains-how-to.15269/post-1701271

So to really get an understanding of what problems each of your subwoofers is "suffering", then you should sweep them individually from the listening position, and with an acoustic reference. Then you can study their responses (Filtered IR, Decay plot, GD plot) to see where in the freq response each subwoofer is having problems. If your room is rather rectangular, the problems should also be revealed by using REW's Room Sim. You might then find that moving 1 subwoofer slightly will fix a problem it has enough to blend with the others better. You should also be able to determine where they have excess delay problem and work around that.

In aid of the above, instead of using the "phase alignment" feature, try the apparently new "impulse alignment" feature that I outlined here (scroll down to the bolded bit): https://www.audiosciencereview.com/...ubwoofer-s-to-mains-how-to.15269/post-1700278

When setting up new rooms, is there a normal range of "you'll typically increase sub by ??-to-?? db to make it sound good"? It seems I need to add at least 7db to make it start sounding appreciably "moving" and that just seems like a lot of extra db for "normal home theater results".

There are "room targets" and the like. They mostly have a bit of a bass boost like you have. The "ramps" start point and slope is usually in the same ballpark. I wouldn't think you need it higher. Your bass may not sound so good because your alignment has made it "slow", for the reasons I outlined above.

Are you listening at really quiet volumes though, like <70dB? I saw something (plugin/app/etc, can't remember) that turned up the bass as the volume master volume got lower, because apparently the bass seems quieter (relatively) if you don't. So maybe if you are listening really quiet then you will want some more bass boost, but as the volume gets louder, you might not want that much boost? Dunno. I would check all your IR's like I outlined above before worrying about that.

If you upload some measurements (mdat file), I'll take a look. 1 measure for each sub with acoustic ref. List the final delays and gains you used too (and crossovers), so I can combine them the same way and see what's what. I'm trying to learn more about multi-sub, so this would be helpful for me too.

 

[ZolaIII]

You need to get the equal loudness normalisation (ISO 226 2003) or set the curve to it at desired SPL (or cuple of them for usual day, night, chill time and store them so that you can easily change them accordingly).

This is ISO 226 2003 from JRiver as measured to use for reference.
As much as I can see you are getting pretty good bass FR output. Would like to see your waterfal plots and RT60 one's.
Edit: one more thing ARC tends to show nicer graphs then actual measured response is so trust REW more and eventually do additional corrections afterwards with REW calculated PEQ's if need be.

 

[Sven777]

Thanks all - I ended up just increasing the gain to the subs. My hangup was that I felt like I may have been increasing the gain "too much" after RC but essentially all I've heard is "ymmv" in regards to trying to get a feel for what "too much" might be.
Thanks again for the thoughtful replies

 

[Alice of Old Vincennes]

I certainly understand your resignation. This why I despise subs. I solved my problem by unplugging my subs. If I need low hz I wait for thunderstorms.

 

[Sven777]

I certainly understand your resignation. This why I despise subs. I solved my problem by unplugging my subs. If I need low hz I wait for thunderstorms.

lol - in addition to gain increase I also added a ButtKicker which added a lot to the overall experience for me - so much so that I put in a preorder for a Crowson system. Maybe in your case something like that would be a decent sub replacement

 

[ozzy9832001]

I would think EQing to 76dB would be too quiet for most people. I usually EQ to 86dB, even if the mains are playing as low at 65dB I still find the bass pleasing.

Sometimes, we get hung up on this idea of a ruler flat frequency response, and that's just not how things really work.

Adding too many filters can cause the sound to be very unnatural and weird. I don't know what the XO is on the subs, but it looks like between 100hz and 120hz, which is what mine are crossed over at. All you should do is boost the gain to get 100hz region to 85dB and then shelf the peaks below 50hz, since it's relatively flat, just loud.

The way the bass interacts with a room is never easy to define unless the room has been built from the ground up.

 

[Sven777]

I would think EQing to 76dB would be too quiet for most people. I usually EQ to 86dB, even if the mains are playing as low at 65dB I still find the bass pleasing.

Sometimes, we get hung up on this idea of a ruler flat frequency response, and that's just not how things really work.

Adding too many filters can cause the sound to be very unnatural and weird. I don't know what the XO is on the subs, but it looks like between 100hz and 120hz, which is what mine are crossed over at. All you should do is boost the gain to get 100hz region to 85dB and then shelf the peaks below 50hz, since it's relatively flat, just loud.

The way the bass interacts with a room is never easy to define unless the room has been built from the ground up.

That's a very practical approach and resonates with me - tyvm!

 

[OCA]

I don't have Anthem and never used ARC but I've written a browser script (platform free) to extract measurements from Genesis .arc3 files and export them to REW. It could be useful for you guys.

ExtractAllMeasurements_arc3.html

drive.google.com

It's also linked in the Audyssey ART video description.