ETF Results from my new room. Gurus, what do you think?

Post and discuss acoustic topics, Studio design, construction, and soundproofing here

ETF Results from my new room. Gurus, what do you think?

Postby MarkEdmonds » Tue Jun 15, 2004 8:15 pm

Having cracked the incredible problem of how to wire in a microphone and connect a mixer to a sound card, I can post my first attempt at ETF results.

Readers of my build diary will know what the room looks like and how it has been treated.

Here is a view that shows the measuring position. It is the listening position but slightly higher so that the ECM8000 is midway between each driver.

Image

The remaining images I will link to so I don't have to reduce the size.

All tests were done using 1 sec MLS, low and full freq as appropriate.

http://www.sonicbedlam.co.uk/ETFPlots/impulse.gif
http://www.sonicbedlam.co.uk/ETFPlots/logfreqfr.gif
http://www.sonicbedlam.co.uk/ETFPlots/etc.gif
http://www.sonicbedlam.co.uk/ETFPlots/logfreq100-1k.gif
http://www.sonicbedlam.co.uk/ETFPlots/logfreq1k-10k.gif
http://www.sonicbedlam.co.uk/ETFPlots/low.gif
http://www.sonicbedlam.co.uk/ETFPlots/low3.gif
http://www.sonicbedlam.co.uk/ETFPlots/low4.gif
http://www.sonicbedlam.co.uk/ETFPlots/phase.gif
http://www.sonicbedlam.co.uk/ETFPlots/rt60.gif
http://www.sonicbedlam.co.uk/ETFPlots/waterfalllow.gif
http://www.sonicbedlam.co.uk/ETFPlots/waterfallfull.gif

Now my knowledge is on the border here so although I have a rough idea of some good and bad limits, I am far from able to see the "big picture".

The full range log freq plot shows a big dip at 102.27Hz. I'm not sure I understand this but is it related to the large quantity of bass trapping I have? This 100Hz drop is my main concern.

I've also got a peak at 188.51Hz which perhaps is a room mode? I would like to clean this up as it is a fairly critical area.

In general, how do these plots look? With a better level of knowledge, is it possible to interpret the data and point out other big areas of concern that I need to address?

I'm scratching around here so any advice on the above is greatly appreciated.

Thanks folks.

Mark
MarkEdmonds
 
Posts: 493
Joined: Sun Feb 08, 2004 9:15 pm
Location: Peterborough, UK

Postby Savant » Tue Jun 15, 2004 8:22 pm

Mark,

1. If I look through your diary, will I find the room dimensions, loudspeaker locations (including height above the floor) and listening (mic) location (also including height)?

2. For maximum accuracy at low frequencies, consider re-running with the 5 second sweep.

Thanks.
---Σοφός---

10*bLog(P²)

"That buzzing noise means something." - Winnie the Pooh
Savant
 
Posts: 753
Joined: Mon Feb 02, 2004 1:52 pm
Location: Kansas

Postby Bob » Tue Jun 15, 2004 8:23 pm

Did you use two speakers for all the tests, or just the phase?
Regards
Bob Golds
"The only thing we regret in life is the love we failed to give."
"Be a rapturist -- the backward of a terrorist. Commit random acts of senseless kindness, whenever possible" - Jake Stonebender
Bob
 
Posts: 4359
Joined: Sun Feb 08, 2004 4:37 am
Location: Oakville, Ontario, Canada

Postby MarkEdmonds » Tue Jun 15, 2004 8:40 pm

Jeff - I think the room dimensions are in there but certainly not the speaker positions. Don't waste your time reading through the whole diary for this. I will repost them below. I assume the room dimensions shuold be quoted prior to adding any treatment?

Bob - All tests were done using both speakers (simultaneously).

Room dimensions:

Front to back: 3.65m
Left to right: 4.06m
Ceiling: 2.62m

Mic location:

From front wall: 1.4m
Height: 1.28m
Midpoint on left to right axis perpendicular to front wall.

Speaker positions measured to midpoint of baffle:

Left to right axis at centre +/-0.86m
Height: 1.24m
From front wall: 0.65m

Thanks,

Mark
MarkEdmonds
 
Posts: 493
Joined: Sun Feb 08, 2004 9:15 pm
Location: Peterborough, UK

Postby notallthere » Tue Jun 15, 2004 11:46 pm

Mark

Nice tests, glad you got things up and running.

Is the impulse.gif a LF or Full frequency test. Would like to compare my LF impulse graph as I have 3 distinct pulses after the main pulse. The timing of the these pulses match room boundry reflections almost to the inch. Which confuses me as I thought room resonance decay was more of a problem below 160 Hz than reflections were.

RJ
notallthere
 
Posts: 23
Joined: Tue Apr 20, 2004 11:10 pm
Location: Ontario, Canada

Postby Paul Woodlock » Wed Jun 16, 2004 12:03 am

Greetings Mark

Some comments :)

Your reverb time looks cool. Quite uniform across the band, and pretty low.

Your log frequency plot looks similar to mine under certain conditions.

The way your curve rolls off ( the top end is about 7 or 8dB down on average compared to the low end. ) is how my cruves look when I measure with two speakers running.

I would repeat the test again measuring each speaker indepentently. I would wager that broad dip at aruond 6kHz would go away when only one speaker measurement, and general response would be flatter. ( except the low modes )

I also have a peak at around 190Hz in my temporary studio .Although it's not quite as broad as yours.


Could you show your Energy time curve on a plot from 0dB-100dB, and from -5 to 500ms. Currently we can only see the first 15ms of action :)


And could you also show the waterfalls for at least 300ms. :)

Cheers mate.

Interesting stuff. I can't wait to give my new studio the ETF treatment! before and after the acoustic treatment.

The quest is to get the bass at +/-3dB eh? I woud imagine ridiculous amounts bass trapping ( or a LARGE ROOM ) is required for this.

It's also worth remembering that the freq response of the speakers themselves is being measured too. Passive crossovers can cause freq response uneveness at and around the x-over freq.

My old Peavey Studio monitors ( Oxymoron ) has a real shit response. They had a 10dB peak at about 3kHz.

If you do an ETF test with the speaker in the middle of the room ( easy in your case ), and the mic less than 2 feet away from the speaker, aimed at the acoustic centre ( usually between the bass and top drivers, but nearer the top.) of the speaker, you can get a clearer picture of the speakers response as the room effects are lessened. Or alternatively ,and probably better, take the speaker outside, and measure when the ambient noise is low. :) :)

Paul
Paul Woodlock
Strange Being
 
Posts: 2808
Joined: Mon Feb 02, 2004 1:32 am
Location: Peterborough UK

Re: ETF Results from my new room. Gurus, what do you think?

Postby Savant » Wed Jun 16, 2004 2:54 pm

Mark,
MarkEdmonds wrote:The full range log freq plot shows a big dip at 102.27Hz. I'm not sure I understand this but is it related to the large quantity of bass trapping I have? This 100Hz drop is my main concern.

Here's what I think is happening at 100 Hz:
• The (2,1,0) and (1,2,0) modes coincide here, both of which produce a null at your microphone position; the (2,1,0) mode more so.
• Your SBIR also produces a big dip, which is deepest between 100 and 130 Hz or so.

To find out which is contributing more - the modes or the speaker/mic locations - you would have to run a bunch more tests at various positions around the mix area and even outside it.

However, more measurement might not be worth it: Some deep corner traps would help with the modes and good broadband absorption around on the walls would help the SBIR. (I haven't calculated all the exact frequencies for the SBIR - just the overall response from Thomas Barefoots calculator.)
I've also got a peak at 188.51Hz which perhaps is a room mode? I would like to clean this up as it is a fairly critical area.

Best I can tell, more tangential modes {particularly the (2,4,0) mode} and a little bit of the SBIR again. Since this is higher up in frequency, anything you do for the lower stuff should carry over to help with this.

Does this help?
---Σοφός---

10*bLog(P²)

"That buzzing noise means something." - Winnie the Pooh
Savant
 
Posts: 753
Joined: Mon Feb 02, 2004 1:52 pm
Location: Kansas

Postby Bob » Wed Jun 16, 2004 4:08 pm

Hi Jeff:

Some deep corner traps would help with the modes and good broadband absorption around on the walls would help the SBIR.


I think Mark is already swimming in deep corner traps.

This is what he's got now:
http://www.bobgolds.com/TrapMark/home.htm (scroll down)
http://forum.studiotips.com/viewtopic.php?t=456
Regards
Bob Golds
"The only thing we regret in life is the love we failed to give."
"Be a rapturist -- the backward of a terrorist. Commit random acts of senseless kindness, whenever possible" - Jake Stonebender
Bob
 
Posts: 4359
Joined: Sun Feb 08, 2004 4:37 am
Location: Oakville, Ontario, Canada

Postby Savant » Wed Jun 16, 2004 4:27 pm

Holy s**t! I don't think I've ever seen someone that wild about Mineral Fiber. Kudos, Mark!

Well, with the corners very obviously taken care of, that leaves the room boundaries and the loudspeakers themselves.

Paul's advice about measuring your LS response is very good. The only thing I would do differently is put the microphone up to the loudspeaker to it's almost touching (>1mm) the dustcap of the cone. You won't get the HF response right on, but the LF/MF will be dead nuts on - almost anechoic. This next part is important:

Run the 5 second sweep and window it as far out as you can without looking at reflections. Your floor will be the most likely culprit. With the mic more or less @ the LS, you should be able to get about 7-8 ms of reflection free data. Unfortunately, that means you will still get significant "room" in the measurement below about 140 Hz. The good thing is that reflection shouldn't contribute too much to the response, so you can trust it down quite low.

Of course, if you take the LS outside and put it on the (not hard) ground, you won't have to worry about any of this.

As for the room boundaries, I am assuming your wall treatments are beefy: 4" thick or more? (It sure does look like it.)
---Σοφός---

10*bLog(P²)

"That buzzing noise means something." - Winnie the Pooh
Savant
 
Posts: 753
Joined: Mon Feb 02, 2004 1:52 pm
Location: Kansas

Postby Paul Woodlock » Thu Jun 17, 2004 3:36 am

Savant wrote:Holy s**t! I don't think I've ever seen someone that wild about Mineral Fiber. Kudos, Mark!

Well, with the corners very obviously taken care of, that leaves the room boundaries and the loudspeakers themselves.


From my experience Speakers can be far from flat, despite extraordinary claims from marketing BS.


Paul's advice about measuring your LS response is very good. The only thing I would do differently is put the microphone up to the loudspeaker to it's almost touching (>1mm) the dustcap of the cone. You won't get the HF response right on, but the LF/MF will be dead nuts on - almost anechoic.


hehe - I once used ETF to measure the response of my headphones. stuck the slim behringer measuring mic inside while I was wearing them. The RT60 was pretty good in there :) But interestingly the modes of the room were still apparent.

This next part is important:

Run the 5 second sweep and window it as far out as you can without looking at reflections. Your floor will be the most likely culprit. With the mic more or less @ the LS, you should be able to get about 7-8 ms of reflection free data. Unfortunately, that means you will still get significant "room" in the measurement below about 140 Hz. The good thing is that reflection shouldn't contribute too much to the response, so you can trust it down quite low.

Of course, if you take the LS outside and put it on the (not hard) ground, you won't have to worry about any of this.


Those with spare slabs of rockwool, could place it on the ground between the speaker and the mic.



As for the room boundaries, I am assuming your wall treatments are beefy: 4" thick or more? (It sure does look like it.)


Mark has a rockwool womb :) :) :)


Paul
Paul Woodlock
Strange Being
 
Posts: 2808
Joined: Mon Feb 02, 2004 1:32 am
Location: Peterborough UK

Postby Eric.Desart » Thu Jun 17, 2004 6:47 am

Hi Mark,

I saw that low frequent waterfall.
What is this hardly decaying ca 40 Hz thing? Looks as a resonance of something.

Note I'm not used to ETF, so I maybe misinterpret some things.

Eric
Eric.Desart
Moderator
 
Posts: 2461
Joined: Tue Feb 10, 2004 4:29 am
Location: Antwerp, Belgium

Postby bert stoltenborg » Thu Jun 17, 2004 2:27 pm

paul,

measuring a headphone can be done by drilling a hole in a piece of mdf, stick the mic trough it till it is flush and put the head phone over the mic; the mdf is closing of the phone.

Same with in ear pieces; you can measure them in a terminated tube.

I don't think you see modes when you measure them this way.

Be sure not to exeed the max input of the mic, or your measuremant result get distorted.

Regards,

Bert
bert stoltenborg
 
Posts: 4569
Joined: Sun Apr 18, 2004 11:03 am
Location: Achterhood, Netherlands

Postby Savant » Thu Jun 17, 2004 2:29 pm

Eric, Mark:
Eric.Desart wrote:I saw that low frequent waterfall.
What is this hardly decaying ca 40 Hz thing? Looks as a resonance of something.

Based on dimensions 3.65m L x 4.06m W x 2.62m H, (1,0,0) is ~47 Hz. I'm guessing he's just close enough with his mic location to get a nice ring there. Even with all that rockwool, 47 Hz is going to be difficult to damp. I ran a prediction with his mic/LS locations and got a bump at 46-48 Hz - i.e., it should be there.

Of course, the loudspeakers I think Mark is using can also be quite "bumpy" right in that range, in my experience.
Mark: Is there a boundary roll-off switch on the back? If so, you might consider using it? (IMO, 40-50 Hz is best left for a matched, dedicated subwoofer.)
---Σοφός---

10*bLog(P²)

"That buzzing noise means something." - Winnie the Pooh
Savant
 
Posts: 753
Joined: Mon Feb 02, 2004 1:52 pm
Location: Kansas

Postby Eric.Desart » Thu Jun 17, 2004 6:42 pm

Jeff,

Thanks.
I wonder in how far the speakers are the cause....

Eric
Eric.Desart
Moderator
 
Posts: 2461
Joined: Tue Feb 10, 2004 4:29 am
Location: Antwerp, Belgium

Postby Savant » Thu Jun 17, 2004 6:48 pm

Eric,

Eric.Desart wrote:Thanks.
I wonder in how far the speakers are the cause....


I think probably more than the loudspeaker company would care to admit! ;-)
---Σοφός---

10*bLog(P²)

"That buzzing noise means something." - Winnie the Pooh
Savant
 
Posts: 753
Joined: Mon Feb 02, 2004 1:52 pm
Location: Kansas

Postby MarkEdmonds » Thu Jun 17, 2004 9:21 pm

Hi chaps - sincere thanks for the continued feedback on the thread.

I've been really short of time over the last couple of days for testing but have established that the 100Hz and 47Hz problems are features of the room, not the speakers.

I have done on-axis testing at 1m distance which shows the response to be pretty flat above 300Hz.

I have done testing with the mic just off the cone for the low end and that shows a smooth but accelerating roll-off at about 3db/oct from ~120Hz downwards.

Hope this isn't a silly question but when you put the mic so it is almost touching the cone, could you introduce irregularities through magnetic interference between the mic and the driver?

On Saturday I am going to repeat the tests to check consistent measurements and will post the results. I'll also be posting questions concerning a longer term view on monitor selection for the room. Plenty to come but just not enough time in the day when the day job takes up 12 hours of it :((.

Mark
MarkEdmonds
 
Posts: 493
Joined: Sun Feb 08, 2004 9:15 pm
Location: Peterborough, UK

Postby Savant » Thu Jun 17, 2004 9:29 pm

Mark,

MarkEdmonds wrote:Hope this isn't a silly question but when you put the mic so it is almost touching the cone, could you introduce irregularities through magnetic interference between the mic and the driver?


Not a silly question. But the even with the mic near the dustcap, the LS magnet should be far enough away not to introduce any funny business. At least, I've never seen any.

Easy check: Turn the LS off and run the test. If the magnet is going to affect anything, it will do so whether there's signal running through the LS or not. If it's quiet enough, you should be able to see any weird spikes caused by something like that. (And please - let us know what you see! :-)
---Σοφός---

10*bLog(P²)

"That buzzing noise means something." - Winnie the Pooh
Savant
 
Posts: 753
Joined: Mon Feb 02, 2004 1:52 pm
Location: Kansas

Postby MarkEdmonds » Fri Jun 18, 2004 9:59 pm

Hehe, just been having great fun. In preparation for the series of measurements I am doing tomorrow, I've been doing walkabout tests in the room.

ETF highlighted a major problem @ 102Hz so I set a sine wave up at that freq and wandered around. If ever there was proof that this was a room problem, this was it.

I wish I knew where my level meter is so I could have watched the needle vary but depending on where I stood, I could get incredible null spots. At the worse points, the level dropped so much it was almost as if the speakers were disappearing and I also got that strange unfocused sensation in the ears which I associate with something being out of phase.

The pattern of nulls was roughly along the room diagonals.

I experimented with lots of speaker positions and also found that if you plotted an imaginary line from the speakers to the room corners, the nulls seemed to follow this.

Interestingly, at the points of deepest trapping, if I put my head right up against the traps, there was a very local boost. My brain can't work that one out.

I also experimented with adjacent frequencies and it is a narrow Q (correct expression?) round 102Hz.

Finally, I also found that on the sides of the room that have more corner trapping (can't be helped by door positions), the freq response tailed off more quickly than the side with less trapping. Roughly, the left side where the trapping is less went by 25Hz but the right side with big trapping went at 30Hz. Maybe my crazy idea of building triangular traps on castors for the door areas isn't such a bad idea afterall.

Anyway, very amusing experiment which everyone should try in an untreated room just to hear how bad things can get.

Current speaker position after these tests "sounds right" to my ears so I will use that as the base point for the more accurate measuring.
MarkEdmonds
 
Posts: 493
Joined: Sun Feb 08, 2004 9:15 pm
Location: Peterborough, UK

Postby Scott R. Foster » Sat Jun 19, 2004 10:50 am

Hmm:

Must be a 1/4 wave null off the back wall... LOL.

But seriously - 102 Hz = a wave of about 11 feet... got a pair of surfaces 11' apart [or 22', or 33', or 44'....].

What is the distance from your floor to the bottom of the ceiling [not the mineral wool but the boundary behind the wool]?

If you move the speaker does the phenomena go away or change frequency?
SRF
Scott R. Foster
 
Posts: 3854
Joined: Mon Feb 02, 2004 12:41 pm
Location: Jacksonville, FL USA

Postby MarkEdmonds » Sat Jun 19, 2004 3:54 pm

Scott, yes, moving the speakers does change the level of the null.

Ceiling height from floor to boundary is 9' 6".

I'm going to be starting on the ETF session very soon.

I will be moving the speakers around the room on a grid to record the results. I was planning on a six inch grid. Is this about right or should I go smaller or larger?

Thanks...

MArk
MarkEdmonds
 
Posts: 493
Joined: Sun Feb 08, 2004 9:15 pm
Location: Peterborough, UK

Next

Return to Discussion

Who is online

Users browsing this forum: Bing [Bot] and 8 guests