John Sayers' Recording Studio Design Forum Archive

Sent another sub test...

Been kinda busy today! Sorry about that.

But yeah, the REW data is looking rather nice, with some caveats. Overall, a nice improvement.

First, however, there seems to be an error in the file you sent yesterday WITHOUT the sub ("cable fixed"): the L and R readings are practically identical: I mean, I know your speakers are really nice, and well matched, and all that... and your soffits are built superbly, and also well matched, but when you can't even see the L curve at all when the R curve is on top of it... well, that makes me wonder... just a bit....
That's supposedly the difference between your speakers, according to what you sent. Dead flat! Not even 1/10th of a dB difference! It would be so nice if that was real, but it ain't... I think you did the R test twice, and named one of them as L. Or vice versa: So please do the full set of 7 measurements and send me that: --S, L--, R--, LR- LS-, RS-, LRS.

("S" stands for "Sub" in the above, obviously.)

Don't worry too much about finessing the sub settings and position yet: we'll get to that. I just wanted to get a general idea of how it is doing, and I'm a bit surprised that it is not doing so well! Here's the thing:
Three curves on there. The purple one is just before the cloud went in, the light green one with the cloud but no sub (just the mains), and the dark green one is cloud with sub. As you can see, the sub is filling in the 60 to 90 Hz region nicely, but look what's happening below 50 Hz: You are losing about 6 dB, all the way down! And below 40 Hz, you are losing about 5Hz on the response as well. That's hard to explain, unless the settings are way off on your sub. Or the position. Or maybe it's just that your sub doesn't go as low as your mains do! Do you have just one sub, or two?

Anyway, tuning the sub location/settings is for later. Right now, lets look at what the cloud did for you...

First, the Frequency Response. Similar to the above, but cleaning up and zooming in for more resolution, and using only one main speaker (to avoid effects from interference patterns between speakers):
Purple is before the cloud, light green is with the cloud in. You can see the general nice very smoothing effect it had on the region between about 150 Hz and 200 Hz, as well as 70-110 Hz, around 300 Hz and 420 Hz., and even some minor stuff at 45 Hz. Quite nice.

Now the waterfalls. First, a series of the EXACT SAME waterfall, from before the cloud went in, showing the aprx. frequencies that I was attempting to target:
Those are the six specific frequencies that I was trying to tune it for. And here's the result:
Some of those frequencies got very nice results, others only budged a little, but the overall effect is clear. Here's both of them on the same graph:
The Spectrogram (this one includes the sub): Self explanatory, and very nice.

The impulse response, out to 250 ms (smoothed to 0.1ms):
RT-60 graph: Your decay time right now is at 210 - 220 ms (depending on how you measure it), but there's a bit of a drop around 250 Hz, which I'm pretty sure is due to your beautiful recliners. We might get some of that back once the desk is in, but not much. However, the good thing is that there's no sudden changes between adjacent frequency bands (specs call for no more than 50 ms: your biggest jumps are about 40ms: all the rest are around 20-30), and the overall time is great! 215 ms for that size room is pretty decent. And it turns out that the CORRECT decay time for that size room just happens to be...... (drum roll please! ) ... 215 ms according to the IEC standard, or 185 ms according to the ITC and EBU standards. So yeah, I'd say that we are very much in the ball park! Very accurately in the ball park, as it turns out! Spot on, in fact.

So overall, I'm quite please with how it is looking!

There's a couple of reflections that we'll have to chase down, but nothing too serious. You can see them on the zoomed-in and normalized impulse response graph:
Around 4.5ms, then again around 9 ms, with some other lesser stuff too around 15 and 17. But no big deal. The "string trick" will help you locate those easily, so we can see how to fix them.

There's only one thing that I'm not too happy about: a phase flip at around 146 Hz:
It wasn't there before, so it has to be the cloud. I'm suspecting the light panel. You mentioned that you have that on a hinge, so maybe you can do a pair of tests: Just LRS, first with the light panel closes, then with it open, to see it that's the culprit. We might need to change the angle of the cloud, or the panel, or both, or do something else... Hmmmm...

Apart from those minor issues, it's looking really good!



- Stuart -

Changing the angle of anything at this point will really suck, so I hope we don't have to....I'll send the new tests shortly...bummer.

Next test sent... didn't change the phase issue. Gonna lay off the sub tests until we sort this out.

Sent another test moving the mic forward....Interesting.

Well,

I did a nose level test with the sheet rock lift..check email. Looks like 91" is the sweet spot for spl, but I think we're stuck with the phase flip, at least in terms of the cloud angle....

On the bright side....The spl curve looks a little nicer at that spot... Don't know if its worth putting it there....


going higher to the ceiling yielded nothing interesting. Maybe the desk will interfere somehow...

I still have the cloud suspended on the lift at the nose, so get back to me as soon as you can. Would rather not go through the hassle of placing all the anchors until you give me feedback. Its a major bitch....

Try this: Tilt the cloud down 4° more, which puts the nose at your 90.5" mark, roughly, and then do the "sub reverse" procedure that I sent you by e-mail.

Also, set up a large piece of wood where the desk will be, roughly the size shown in the most recent model. The desk is going to change things quite a bit, so it's work simulating.

Do a REW test with the extra tilt on the cloud but without the "desk", then another with the "desk". Send me those when you can, and I'll analyze them as soon as I can.

- Stuart -

another test sent.

Wondering about the phase flips at 45 and 147 hz. Worried that those won't be correctable at this point.

Ok to finish the trim and fabric on the cloud then? sending another file with some Universal Audio plugin EQ experimentation. Kinda interesting.

getting some weird sub behavior.Put everything back the way it was including the insulation up above and I can't replicate what I got last night. Sending another file with the anomalies. tried switching subs too. Don't know if I have another set of cables though.

Spent the day resetting the rest of the anchors to the new angle now that you approve. Gonna hold off on doing the reverse sub test until we figure this out....pretty shot after spending the whole day trouble shooting this...

The sub location is the picture I sent previously. I didn't change much so I don't understand the drastic behavior....

It's not cables, so I wouldn't worry about that. I'm guessing one of three things:

1) The subs don't actually go as low as they are advertised as going, or the advertising is misleading (partial truth)

2) Some issue with the location of the subs in the room.

3) Some issue with the control settings.

Or a combination of all three...

For those following the thread: The issue is that with the subs turned ON, the low frequency response gets worse, not better. The roll off moves up substantially, from about 40 Hz to about 54 Hz, then takes a very steep dive, and introduces a large null:
Green is WITHOUT the sub, and purple is WITH the sub...

Strange...

- Stuart -

Especially since I was getting some really cool measurements last night.. Been trouble shooting this all day to no avail. I even wondered if the tension of my replaced anchor points might have changed the resonant frequency of the cloud...Of course cause I only had my imagination at my disposal in this case It is almost Halloween...hmmm.

Mystery solved! A combination of the physical position of the sub in the room, and the "phase-invert" switch on the sub... plus maybe the sub-par crossover circuit in the sub itself.

I got Frank to move the sub physically all the way to the rear wall, then flip the "phase" switch:

First the sub at the rear wall, as compared to the "no sub" test from a few days ago: Green is "no sub", and red is "sub at rear wall, phase normal". You can clearly see that the weird stuff with the low frequency response and huge dip has gone away, and the sub is now doing what it is supposed to do: extending the low range

Now see what happens when Frank flips the phase switch: That red curve is the exact same one as above: "sub at rear wall, phase normal", and the blue curve is "phase flipped". You can clearly see what a huge difference that makes: it kills the entire low end! Everything below about 90 Hz just dies. The roll-off starts at around 55 Hz, instead of 40 Hz. It is even worse than the "no sub" curve"!

So at least we figured that out.

All of this heavily underlines the importance of what many of us keep on stating and repeating on the forum: correctly positioning and configuring your speakers is critically important! No other changes at all were made to Frank's room between these tests; the ONLY thing that changed was the location of the sub in the room, and the setting of the "phase invert" switch.

Now it's time for Frank to play "hunt the best spot for the subs"... a wonderfully boring game, played slowly and painfully...



- Stuart -