First attempt - what's next? :)

I have gone through all the steps to measure my setup properly, including buying an ECM8000 mic and a Novation Speedio external sound... card? unit? Whatever. It seems to give a good frequency response even uncalibrated anyway (but I have of course applied calibration).

I have a setup consisting of B&W XT2 front speakers (satellite), roof mounted due to various reasons, and a Rel Storm sub in one of the front corners. The room is large, 6.5 x 6m with openings at the back towards kitchen and dining area, giving a total of approx 60 square metres. The sub corner is 2 metres to the left of the left speaker.

My initial goals with this exercise are:
- Integrate the sub with the XT2 speakers
- Try to get a good flat subwoofer response
- See if I need to get a bigger subwoofer

Since this is our living room, there are limitations in what I can do to the room, but I thought I'd give it a try and see what comes up.

Now, I'm a bit curious as to what to do now with the great program. The graphs look pretty and all but I need some pointers to what direction I should go - what would my next steps be? Also, I'd love to hear any comments regarding my frequency response - is there hope?

The graph is measured from the right speaker + subwoofer, cutoff frequency in my processor at 80Hz. Mic pointing upwards, full range signal used, calibrated at 75dB, 1/3 octave smoothing applied.



Any comments most appreciated!

Oh well, here's the room as well....



(Edit: Ignore the stand-mounted speakers at front, the mic is placed in the sofa and it's those black things in the roof that's measured)

Oh well, here's the unsmoothed graph as well, for your sheer enjoyment.

First: welcome to the Shack! Nice to see other people from Scandinavia on this forum

Please remember to use the scales recommended in the REW guidelines (Up to 105dB if I remember well). Take a llok at the first threads in this section and how other people do it. It will make it easier for everyone to compare your results...

Otherwise, if you want to get a good flat subwoofer response then you should test your sub alone. As far as I can see, your curves are for the two front speakers and the sub, right?

I'm not an expert but I would say that you have a good starting point here. Do you use any form of room correction system?

---

Martin Logan Vista | Swans F1.1C | SVS PC-Ultra | SlimDevices Transporter | Bladelius Tyr | Yamaha RX-V1800 | Denon DVD-2930 | Klipsch RS-52 | Sharp XV-Z3000 | Adeo Greyvision Plano 16:9 100"

Thanks This seem to be a very good forum, but way too much information to grasp at once. I'll read as much as I can...

Hm and I thought I used the correct scaling. Must have lost it somewhere.. the measurement is with just the right speaker and the sub, no room correction at all. That dip at 46Hz seem very interesting....

I'm reading about waterfalls, and those are really ugly. At 1s I still have lots of sound. There is probably lots to do, I just don't know where to start

Hey CalleG, good thread. I've got a similar room with similar questions. I'll be listening. Thanks for bringing it up.
-Steve

As previously suggested, the graph standard is vertical 45dB-105dB and horizontal 15Hz-200Hz.
Turn off the smoothing, it's reserved for higher frequencies and not to be used with sub measures up to 200Hz.
Use a 75dB target and a subwoofer test signal from REW (not full range).

Anyway, your response is quite good. You won't likely hear the sharp cancellation at ~45Hz, so you can move the sub a bit or ignore it.

You have a bit of a peak at ~30Hz that could be equalized away and as such would allow you to turn up the wholesale level of the sub a bit to bring out the bottom end more.

Yeah, we usually recommend taking a sub only measure (with the mains disconnected) to get a feel for the equalization that it requires. The mains are then added after equalization to allow the adjustment of the sub phase control to smooth the crossover area.

brucek

OK so here goes... I pulled the plug to the poweramp, leaving just the sub. This time I triple-checked the scales

Also, I did a waterfall - note that the scale is 600ms

The sub needs to be in precisely that corner, as can be seen from the room picture. I've placed it close to both walls - is that something I should experiment further with?

Seems like equalization would come in handy, surprise eh? Is the 1124 most suited for this or is there another option?

Are there any other obvious things I should do or check or refurniture?

Attachments

	subonly.jpg (45.3 KB, 79 views)
	waterfall-subonly.jpg (99.6 KB, 77 views)

A little EQ at the 30Hz peak would be useful, but its not a huge deal - our hearing is actually quite insensitive to long decay times below 50Hz. You can ignore the very low frequency stuff at 15Hz or so. It would be worth moving the mic around a bit to see how that affects the 46Hz notch, otherwise your response looks good.

So are you saying I don't really need an eq? I was kinda looking forward to getting a new toy..
Why should the lowest region be ignored? What would it take to get notable lower bass response?

A bigger sub? Your response already extends below 20Hz though, does it sound lacking? You have a good level match to the mains currently, which should give very good results for music, but can always simply turn the sub up a few dB to get a greater low bass effect for movies.

You right on the edge man. It's a pretty good response. A lot of people would love to have it so good.

The BFD is a cool toy, and you could bring down the 30Hz area a bit, but there are level setup hassles with a digital device that you might not enjoy so much.

Here's the skinny. If I had that response, I'd remove my BFD.

brucek

Well, I'm actually pretty satisfied with the sound, but am missing that bass that can be felt. Since my current sub isn't all that small, I was wondering what it would take to really make a difference - the answer seem to be "a lot".

I've been trying to visualize a Stentor or a large Velodyne of some kind, but perhaps it won't make much difference then...

Wow, that actually makes me proud over my fine room... thanks I didn't think the graphs looked all that good, I'm glad I came here for advice.

Is there some acoustic treatment I could sneak in the room without my better half noticing, that would help with that 30Hz peak? And yes, I do think she'd notice if I tucked a helmholtz in front of the fireplace.

Thanks a lot for all helpful comments and suggestions! I'll be back again soon with a full frequency graph.

I'm definately no expert but looking at your graph with the peak at 30 Hz and your comments that you're lacking the bass you can "feel" I'm thinking that some EQ just might do you some good. Then again the response is much better than most people get with no treatment, so you certainly can't complain about what you've got.

It looks like you're trying to calibrate for 75dB and you're right on from 18Hz to the low 20's. However, the output then rises until the 30 Hz peak at 89dB (over 2.5x louder than your 75dB target) which in my opinion is probably drowning out the low "tangible" bass that you wanted.

If I were going to take a wild swing at it I'd probably go with a filter at 32+2 with a gain of -10 and a bandwidth of 15/60. This should reduce your 30 Hz peak by 10 dB to about 78dB while at 24.5 Hz and 43 Hz you'd be only 3dB down from what you've measured. That's roughly the right bandwidth I think to kill the whole peak without bringing down the surrounding area much. A second filter would be something like 63-6 with a gain of -8 and a bandwidth of 8/60. With that filter the points that would be down 3dB from your current measurement are 46 Hz and 68 Hz. I would leave it with just those 2 filters. Hopefully if all goes well you'd be very flat from ~17-92 Hz and you'd have skillfully avoided making those two dips much worse.

Anyways, like I and others said, your existing response is very good, and may not be worth buying a BFD over. On the other hand, if you want to blow $100 on a new toy and have some fun you can probably turn it into a phenomenal graph with no more than two filters. You should play around with the filter settings in REW though and simulate the expected response as if you had the BFD. That could give you an idea of what to expect and if you feel it's worth it. If you really wanted to humor me you could even plug in my two filters and post the new graph for us to see

---

You can see the sub in the left corner, that tall rosewood thing. I'll try to get some new pictures.

Also, I'll do my assigned homework and play with those filter recommendations, as soon as I figure out how. Cool! And thanks!

The placement of the sub is completely free, as long as it's in that corner. Not much space to move around in I'm afraid.. silly room from that aspect. I'll try experimenting a bit with that as well.

I'm assuming that you saved the measurement that you posted here as an .mdat file? If not you'll have to measure it again and then use Ctrl-S or File->Save Measurement so you can load it back in later. To load a saved file you can use Ctrl-O or File->Open Measurement File.

Either way once your graph is on the screen you can either click the EQ Filters button on the right side of the top toolbar, or just press Ctrl-Shift-Q. I guess I should mention before you do that, check the Equalizer menu and make sure that BFD Pro DSP1124P is selected. You can pick a different equalizer if that's what you'd intend to buy, but the filter settings I gave you were intended for this one.

Anyways from the EQ Filters window you should play around with adding filters and see what they do to your graph. The two I suggested might be a reasonable start, and with some fine tuning you'll probably tweak in something better. Make sure it's set to Manual Control and PA type and just have some fun. REW will show a dashed line estimate of the filter's effects as long as you have the "corrected" line checked below the graph so you should post a shot of how it turns out once you're happy.

I can remember in my college days trying to calculate appropriate filter bandwidths by hand after having spent hours running test tones at 1 Hz intervals and plotting it all in Excel. It almost makes you want to cry seeing how easy this whole process has become.