Speaker/Room calibration

[rbbert]

I have and have used the Radio Shack meter for just what you are doing, but have been told that the RS meter has a significantly uneven frequency response, and even the "corrections" found on the Internet aren't accurate enough for this type of use, especially under 300 and over 5000 Hz?

 

[Bruce B]

1) can you post a picture of the controls on the back of the speakers, so we can see what is adjustable.
2) did you originally have the system tuned when the were originally set up?
3)Do you think you can eek out the last 1-3% with the controls?
4)What is your goal?
Wendell

Thanks... the ultimate goal is to have as flat of response as I can below 1k by just using good positioning techniques and speaker controls without having to resort to outboard digital or analog EQ. This is what the consumer is left with. Once they purchase a speaker, the dealer may do the initial setup, but what happens after 1000hr. of breakin.. or if you insert a new piece of equipment, or if something in the room changes?
The speakers were originally set up by the dealer by just using a tape measure and listening. All controls have remained at factory default.
I'll post a pic this evening.

I have and have used the Radio Shack meter for just what you are doing, but have been told that the RS meter has a significantly uneven frequency response, and even the "corrections" found on the Internet aren't accurate enough for this type of use, especially under 300 and over 5000 Hz?

Though this test would suggest the RS meter is fairly accurate, except for below 31.5Hz where the corrections are from +5dB to +20dB as stated in both of these links.

Audio Innovation

Subwoofer Builder

Which according to my graph, the tracing below 30Hz is way off the charts!

This is why I wanted to use my Earthworks M50. I'm sure a few can atest to the accuracy of this mic. I'll try to get it working this evening.

 

[theguesswho]

Hi again Bruce, I think the response looks great. You have one small peak at 75hz which you could reduce with one filter from an equalizer, but why corrupt the entire audio signal for one peak?
The low freq is no problem as many people like a house curve on the bottom end, though you did say you wanted it flat.
But lots of people say flat does not sound good anyway. Besides there is very little musical info below 30 hz. So why worry about it?
You said your system sounds great, so whats the problem?
I tune my system by ear and it sounds great, I have a reference album with deep base that is increadible. I guess you could try moving your speakers or chair, that helps. You allredy got a great room and speakers to die for, what else could it be?
Personally I would not trust any of those measuring devices, trust your ears. Your ears are what you are trying to please, and they are more senisitve than any screwey test instrument!

Jack
Oh by the way only my mother calls me Wendell
Thanks,
Jack D.R.

 

[rhbblb1]

I hate to be the wet noodle, but I see 3 major problems. There is a 15dB null which is very wide and extends from 25Hz to 55Hz. This includes some pretty important musical information. You also have a 10dB null between 90Hz and 150Hz and a 12dB peak at 75Hz. All 3 of these abnormalities will be audible. The great bass response below 20Hz looks impressive, but is meaningless for music. OK for home theater.

Why don't you try decreasing the bass filter to 50 and increase the bass level to +6dB. Then fiddle with the bass extension (whatever that means). Probably the Q. If so, you may want less bass extension to increase output between 20 and 60Hz and have less below 20Hz.
Please try these changes and post your results. I'm trying to get more bass extension between 25 and 55Hz and and less at 75Hz. These peaks and nulls are room induced and may be difficult to eliminate with the speaker controls, but worth a try.

 

[DonH50]

Is the 40 Hz null a room mode? That is high enough to be annoying... Low string on a guitar is around there.

I use an old M30 and cheap m-audio preamp. It has enough gain for my system but not enough for "normal" use (I have other preamps for regular recording).

The RS meter, like virtually all inexpensive meters, is C-weighted and thus rolls off below 30 Hz. You have to spend some money to get a flat SPL meter, but I would take the money and just get a different preamp. It is not going to come close to matching your M50, correction curves or not (too much unit-to-unit variation among SPL meters, for one thing).

I have seen changes in bass response as a speaker breaks in, almost nothing measurable in the mid/treble. YMMV! - Don

 

[Bruce B]

I hate to be the wet noodle, but I see 3 major problems. There is a 15dB null which is very wide and extends from 25Hz to 55Hz. This includes some pretty important musical information. You also have a 10dB null between 90Hz and 150Hz and a 12dB peak at 75Hz. All 3 of these abnormalities will be audible. The great bass response below 20Hz looks impressive, but is meaningless for music. .

Thank-you! This is the feedback I need. I will try the settings tonight. I'm trying to learn all I can about room acoustics and how to fix problems.

Is the 40 Hz null a room mode? That is high enough to be annoying... Low string on a guitar is around there.- Don

The 40Hz null is a room mode. I can change it from 38 - 44Hz by moving the mic forward/back.

I can change my listening position by about 3-4'

 

[microstrip]

(...) The 40Hz null is a room mode. I can change it from 38 - 44Hz by moving the mic forward/back.

I can change my listening position by about 3-4'

The 40Hz null is probably increased by the first width mode and does not seem easy to solve without eq. compensation.

But I would not loose too much time with results coming from a RS meter - see the notes and curves at the Realtraps site and compare them with the values at several sources. There is too much disagreement on the net on these curves and the compensation factors, suggesting great variation between units.

http://www.realtraps.com/art_spl.htm

I have two Behringer ECM8000s bought at different times. Curiously they agree between them within 1dB, but they disagree strongly with my RS sound meter.

 

[Bruce B]

I will add that there is a short tutorial on Room EQ Wizard on YouTube that I'm going by.

Part 1

Part 2

Part 3

 

[RUR]

The 40Hz null is probably increased by the first width mode and does not seem easy to solve without eq. compensation.

Or an additional, properly placed transducer, which may not be in the cards.

 

[DonH50]

Nulls are notoriously hard to treat because they can be very deep. It would be easier to move the listening position if you can...
Peaks you can usually EQ and tame.

I would trust the ECM8000's over the RS meter...

 

[theguesswho]

Don't forget you can't eq dips in the response, as suggested by the other posters. You have mentioned this several times yourself and I agree. Also you dont hear dips in the response as the ear fills in the missing freq.
So just that one Peak to cut by 3db or so.
It's still too early to Tell but perhaps it's the room. I have never been a fan of treated rooms they always sound dead to me. Are you able to remove some of the treatment, this would not be the first time this has been necessary.
You liked the sound prior to measuring and I am sure you still do so why worry?

Jack
POE143
If the first octave sucks why continue?

 

[Kal Rubinson]

Nulls are notoriously hard to treat because they can be very deep.

It is not their depth that is the issue. It is the underlying mechanism of cancellation that thwarts correction.

Don't forget you can't eq dips in the response, as suggested by the other posters.

Sure, you can. You cannot EQ nulls but you can EQ dips or troughs.

 

[theguesswho]

Hi kal, I believe that Bruce said he would
Not use any gain/pasative filters if he were to use an EQ. As it is now with the adjustments available on the back of the speakers limited to what he has, how can he boost a null?

Jack
POE143

 

[Kal Rubinson]

Hi kal, I believe that Bruce said he would
Not use any gain/pasative filters if he were to use an EQ. As it is now with the adjustments available on the back of the speakers limited to what he has, how can he boost a null?

My point was more general. You cannot boost a null with EQ. You can boost a dip or a trough even if Bruce chooses not to do so.

 

[theguesswho]

Hi Kal, OK got it and agree.

Jack
POE143
"If the first octave sucks why continue?"

 

[Bruce B]

Here's the controls on the back of the speaker. There is also a tweeter attenuator at the top out of the picture that goes from 0dB to -6dB

 

[jpv]

Bruce,
I have been using REW for about 5 years. I used the RS digital meter and found it fairly accurate when compared to my calibrate mic. Don't think using it will throw off you measurements that much. This will be interesting to see how you fix the null/peaks in the bass using just the controls on the back othe speakers. I have the feeling you can't but thats where DRC comes in. If you can could you post your REW MDAT so we can see wants really going on.
John

 

[Bruce B]

O'kay.... had to scrap the M-audio and use the good stuff. I'll have to troubleshoot that later.

Used my Earthworks mic with Millennia pre into DAD AX24 A-D converter.

Everything set at default and here is the new graph.

 

[DonH50]

Kal: Should I have said "null due to cancellation"? I think we both understand the issue, just want to clear up the semantics...

Bruce: There's a lot going on in that plot... It would be interesting (but time-consuming) to do runs with each control individually set to their min/max positions to see the effect compared to your nominal plot, WITHOUT MOVING THE MIC! Not something you'de want to post, but then you could see the impact each control has on the FR (frequency response).

For comparison, here's is the response in my room after using MCACC and tweaking the response manually a little, at 1/12th octave and full-range (sorry about the size; I have a small version but it is not very legible and I don't have one in-between, and this data is over a year old). N.B. I am NOT holding this up as the perfect room! Just an example after treatment and running my AVR's calibration, followed by some manual tweaking. My overall excursion (min/max peaks) is a bit less than what you have going on...

 

[Bruce B]

Here is what I ended up with by moving my listening position about 2' closer, changing the bass filter to 150 and bass level down 2dB

Any more suggestions?