using a spl meter for a sub?

[Scp53]

ok, today i went and bought a rs digital spl meter(finally i got one!!).

heres the prob, i hook my sub(dayton titanic mk3 10"- supposed to be flat and non boomy) to my computer with a 1/8 mini to l/r rca into the subs rca inputs. i turn the phase to 0(doesnt matter though because im playing no other speakers), set the gain low, and have the crossover cranked all the way up.

i have testtone generator 3.92 on my computer and ran some test from 30hz to 80 or so. seems that the meter fluctuates +/-6 db for my subs frq response. the hump is in 50-80 region and has lower spl after 50. so basically my question is, where should i place the sub to test? in the middle of the room or in a corner? im mad as hell right now because i think the sub is boomy. the meter was varied in distance from the driver and the dips and rises seem to be the same. wtf. someone please recommend how i should set my sub up to graph is frq response. does windows xp have a boomy output on the volume control(ive read about this prob)? i went into advanced options and lowered the bass slider on windows vol control(but who knows what freq's thats adjusts). all help is greatly appreciated. and sorry for the profanity, im just venting steam.

thanks, scp53

[Scp53]

ok, i must say that now im much happier. i ve tweaked some more and its much better. i guess i was just blowing off some steam. heck, i put the sub like 4in away on both sides in the corner and the freq response is better(this is a different room though). its reading 58db(c wt-ing) or so at 32hz and the same at 60hz. at 50hz it takes a real low dip though, like below 50db. and then its level up from 60hz for the most part.

i can say right now that im glad i got the meter(and i didnt buy into that extended warrenty crap). its actually kind of fun to use. however, id still like some help with putting it to the best use and placing the sub. tomorrow(after i ski) im gonna check the phase of the speakers with the sub. should i use music or a sine wave? and i take back that i think the sub is boomy, it was just a placement thing. and will the frq reponse stay the same at one volume as say a volume of 20 more db(i think i know the answer to this)? thanks,

scp53

[michael hurd]

The radio shack meter is non-linear, meaning it does not measure low frequencies properly. I do not know if anyone has come up with a correction chart for the digital meter, but there is one available for the analog meter. I can post a link for the correction chart for the analog meter, but am unsure if the digital one would require the same values.

Using the meter with a correction chart will get you a lot closer, than using the meter itself. Doing so, would net you a frequncy response that is definately not flat.

[michael hurd]

At 10Hz add 20dBs to the meter's readings and at ...

12Hz add 16.5dB

16Hz add 11.5dB

20Hz add 7.5dB

25Hz add 5dB

31.5Hz add 3dB

40Hz add 2.5dB

50Hz add 1.5dB

63Hz add 1.5dB

80Hz add 1.5dB

100Hz add 2dB

125Hz add .5dB

[Scp53]

aren't those corrections for using pink noise at a c-weighting? whats best to calibrate speakers to? pink noise or sine wave? or some warble tone or somethin? im still searching the net for corrections for the digital one. maybe some one on the forum can make one if none exists?

thanks, scp53

[Scp53]

i found this link, http://www.hometheaterforum.com/htforum/showthread.php?s=&threadid=26728

looks to be where you got your correction chart. as i read it, i think the reviewer/tester person said that the digital and analog one have the same corrections. now, what do i use for the sound? pink noise, sine wave, etc? im not sure at all. please help me with this. thanks

scp53

[WMcD]

You might find the same subject discussed from time to time on the forum.

There is nothing wrong with your meter or sub or amp.

The room itself creates great variations in sound pressure from one location to another. Varying the location of the sub or yourself will alter things.

You should find a CD with constant test tones in the bass or a freeware computer program so you can connect your sound card or headphone out to the amp. Start with about 50 Hz.

Walking around the room you'll hear amazing variations of bass level. At some locations, it will sound as if the sub is not turned on at all. Move a few feet, and the sound is back.

As I've said before, few will believe just how much the room affects the bass until they hear it themselves like this. Almost parlor magic. It it, though, the first step to understanding the magnitude of the problem created by the room itself.

Gil

[Scp53]

hey Gil, i have done a lot in terms of walking around the room and listening to the sound at different spots with different frequencies. I KNOW for a fact that the room has a big affect on sound and in this case, my sub.

if you didnt read above, i have already started using a computer program to generate the tones. my question(s) are- what kinds of sound should i use? i ran pink noise just a little while ago through my sub and response seemed very level. however, i think pink noise is not the best sound to use- it has to much upper frequency in it(i read a definition of pink noise but am too tired to really understand it.). when i use a sine wave response its quite different. the response has many more dips and peaks,etc. so im hoping Gil, Mike, or one of the many other experts on here is able to answer what "sound" to use. and what are the corrections for rs digital meter using that "sound". thanks

scp53

[WMcD]

Sorry about not noticing your initial description of using the computer.

There is a bit of contradiction in measuring bass response in rooms. If you put the mike one place with a sine wave, you do get the measurment for that place.

But we see that the dips in response are different in different places. Also, if you change the frequency by a few percent, the locations of the nodes (dips) move to other places in the room.

If you need to measure just the speaker without the room, you have to take the speaker outside. But that doesn't give a fair report of what is going on when it is in the specific room. Placement near walls and corners do increase bass.

So how to measure? It is really an age old problem.

One thing you can do is have an outside wall/corner intersection of some size. Klipsch has one in Indy and probably Hope too. You'll notice that in the Jubilee article, measurement of the Jubilee and K-Horn were made in such conditions.

But again, this is with only three surfaces and not the six of a room. OTOH, it is a reproducable condition in that all outdoor corners should be about the same if of equal size. We don't have to worry about how near or far the other three walls are situated because they're not present outdoors.

Back to a real room.

One thing which PWK did was to place the mike in several locations and run the sound curves on the same piece of paper in the plotter. Then he took only the top-most trace of any as the actual response.

This honors the principle that dips because of interfering reflections can go down to zero dB. E.g. average level may measure 80 dB but the dips can be off the scale low, say 30 dB. I.e. a 50 dB trough.

OTOH, peaks because of additive reflections can only go up to 3 dB with two of them adding. Maybe we can get 10 dB at some freq. in good room at a low freq.

There several other ways of doing this in room measurement of the bass.

One is to use a warble tone which oscillates in frequency over a given range. 1/3rd octave is pretty common. That way the meter is reading an average response over 1/3rd octave. The meter will bounce around and you estimate the average.

The other is to use band limited pink noise.

First we have to discuss the definition of white noise. It has the same energy content at every given frequency point. The name is a reference to sunlight by analogy.

But that is not how our ears work. They report balanced sound when the energy content is equal within ocataves or sub octaves.

Look at octaves. An octave width is the band in which frequency doubles. It is like going from middle C to the next C on the piano, up or down.

Without picking specific musical tones, let's look at octave widths. In the bass 40 Hz to 80 Hz is one octave and has 40 frequency point. But moving to the mid, an octave is 400 Hz to 800 Hz. There are 400 frequency points. Moving to the treble, an octave is 4000 Hz to 8000 Hz. There are 4000 frequency points.

If we put in white noise with equal energy at every frequency, there is increasing energy per octave even though it is constant with frequency points.

If we want equal energy within each octave, we need to cut or roll off the high frequencies. By analogy, if we did that with white light with a filter, the light would have more in the red end and less at the blue. The light would be more pink, or at least red-er. (Redder?)

So noise with equal energy per octave, or fractions of octaves is called pink noise. It is doing about the same thing as the warble tone. It can be limited to an octave or 1/3rd octave, or even more narrow. You pick out the frequency points (actually the average given) and plot them.

The beauty of the warble tone, or band limited pink noise is that the source is doing all the averaging. On the measuring end all we need is mike and a read out. But you're taking the measurement at one spot in the room, typically.

There are other ways of averaging. With stuff like the LMS system you can take one curve and average over 1/3rd octave by computer processing. Or you can take the measurement in several locations and average those. Or you can do both. The same could be done with pink noise and an RS meter but you have to do the math.

There are some other techniques which pretty much don't work well for bass measurments. In those, a burst of sound is measured and the measurement is stopped before a reflection from a wall (with a longer path) comes to the mike.

Bell Labs was doing this in the 1930's by recording the burst on a film track and then only considering the earliest squiggle, and not the later one from the reflection.

Overall, you can see that averaging by any of the above is really allowing us to smooth out data and look for trends. OTOH, averaging is also ignoring data.

Best,

Gil

Best,

Gil

[colterphoto1]

Nice discussion guys. Especially appreciate Mr. McD's comments. DrWho is coming to my place next Saturday to pick the ChorusII I got for him in FL, I also have my Cornwalls and RB5's here. He has some type of spreadsheet program for plotting our data, so we were planning on going through this process in my listening room with several types of speakers. Your 'how-to' guide is very timely and helpful.

Michael

[formica]

Scp53 what would you like to measure room response or sub response?

In either case, most people who aren't using a RTA (real time analyser) use individual sine waves of single frequency width. Set your meter to "C" weighed, which doesn't modify the response to a human hearings. The "C" weighed is "supposed" to be a flatter response curve, and the RS SPL correction values posted are based on flattening this mode. Fast or slow response should not make a difference if you are using constant sine waves. The values you read are then plotted on a logarithmic scale using a program like Excel.

If you measure your sub indoor this will give you room response combined with the sub's response. Outside you'll get just the sub's response if you place the sub away from all boundaries. The difference of the two is your aprox. room's response... which you already realise is the most significant of the two.

BTW, the RS Digital and Analog meter have similar frequency deficiencies... so the correction values you have will be close in either case. Remember these are just "general" correction values... and the RS SPL is far from being a calibrated mic. For bass frequencies, this will get you close enough.

Hope that helps...

Rob

[Scp53]

are you sure its A weighting? look at this link http://www.hometheaterforum.com/htforum/showthread.php?s=&threadid=26728

maybe that persons wrong, i dunno? and he also says those corrections are for useing the pink noise? or is it also forsine wave,etc? someone really needs to clear this up for me(and maybe others too).

scp53

[sfogg]

" Set your meter to "A" weighed, which doesn't modify the response to a human hearings. The "A" weighed is supposed to be a flat response curve,"

That isn't correct. 'A' weighting rolls off the bass response far more then 'C' weighting does. That is why you often see signal to noise readings made referenced to 'A' weighting. They are cheating as 'A' weighting will really roll off 60hz power line noise in the signal.

If you don't want a modified reading you need to measure flat... which the RS meter can't do.

"and the RS SPL correction values posted are based on flattening this mode. "

Nope, they are the inverse of 'C' weighting. All they do is compensate for the weighting of 'C' mode. It doesn't 'correct' for any difficiencies in the meter itself.

If at all possible when measuring things like this if you are going to do it by hand with a SPL meter try and measure every 1hz. Room modes can be very narrow and without measuring to this precise of a level you might miss them.

Additionally, it is important to be aware that when measuring via SPL meter (instead of an RTA) the numbers you get may or may not be all that accurate for the frequency you are measuring. The SPL meter is giving you the sum total of all noise in the room, it isn't just looking at that specific frequency area (like was a very fine scale RTA would do) so if you sub is doubling or you have extraneous noises in the room it can be throwing off your readings.

Shawn

[Scp53]

i agree with going up every 1hz or so. ive tested in my room and spl can drasically drop or peak with just a slight change in frequency. so what sound should i use- pink noise, sine wave, or warble tone? and where can i get a tone generator for the computer? i have one but its only good for sine and pink noise. its Test Tone generator 3.92 evaluation version.

scp53

[sfogg]

" what sound should i use- pink noise, sine wave, or warble tone? "

What are you trying to measure?

If you are looking for the subs response within the room and you are doing this with a SPL meter you would need to use sine waves 1hz at a time. If you use an RTA you can do it with pink noise and get all the data in one measurement.

"and where can i get a tone generator for the computer? i have one but its only good for sine and pink noise."

That is really all you need. Warble tones are sort of useful if you can set start and stop frequency and the time it takes to move between them. Then by listening to that you can sometimes hear when one area either jumps up much louder or gets much software.

Another decent function generator is this one:

http://www.marchandelec.com/fg.html

Shawn

[Scp53]

thanks for the fast reply. so, do those corrections correspond to using a certain tone or what? should i add those values when im doing test with a sine wave? according to that link i posted, the corrections are only good for pink noise. I'd appreciate you clearing this up for me. thanks again,

scp53

[sfogg]

" do those corrections correspond to using a certain tone or what?"

The corrections apply to the frequency of the sine wave you are measuring.

"according to that link i posted, the corrections are only good for pink noise."

That isn't correct. Pink noise is pretty much by definition a full frequency range signal. If you are measuring pink noise (via a SPL meter) you therefor are measuing over a very wide bandwidth..... so you can't yourself apply the corrections to those frequencies as you don't have a measurement at those specific frequencies. Where you could apply them when using pink noise is if you measured using something like a 1/24 octave RTA.

Shawn

[formica]

Should have reread my post... as Shawn is correct that it should have read "C" weighed... I have corrected my original post, and it should now make more sense.

Rob

[Scp53]

so for frequencies with no correction just take the measurement right off the meter?

scp53

[sfogg]

"C" weighting is rolling off the response too.

Like I said if you want flat response you don't apply weighting to the signal at all. Flat measurement is something the R/S meter can't do. The 'correction' posted above is the inverse of the weighting of C so it in effect gives you flat measurement. Not including any errors in the meter itself of course.

Years ago I saw a site that listed how to remove the weigthing from a R/S meter to give flat response.

Scroll down slightly on this page:

http://www.askmar.com/Noise/Noise%20FAQ/FAQ_Text.html#weighting

and you can see 'C' weighting vs 'A' weighting. Neither is flat.....

Shawn