Home Theater Forum and Systems banner

1 - 8 of 8 Posts

5 Posts
Discussion Starter #1
Hello REW forum

I am a newbie to accoustic testing, a musician setting up my room with some basic treatment and I have been hovering around some of the forums digesting as much as possible. I have a number of questions and first I will start with my problems I am having with REW.

I keep getting error messages stating the Impulse Response is not in the correct place and that my signal level is too low to get an accurate measurement. I posted over on Gear Sluts and I was told that Apple is not confroming to the Java protocol which can cause these issues and to buss the mic signal to the internal audio of my mac which could be better for REW. So while i have data that seems appropriate when I cross reference the numbers from REW to a modal calculator I have doubts about whether its accurate or if there is a user error on my part. and I can't obtain more than 2 tests without getting errors so I cant move the mics to keep testing or move some of the treatment/ monitors and retest etc.

3 Questions are: Is there anything I can do to correct the messages? Does the data from REW seem consistent with the room dimentions so can I trust the results and keep testing? and if so, most other waterfalls I have seen posted have much shorter decay times so I think I making a mistake ? If i have missed some tutorials please direct me.

my testing setup

I am using a matched pair of Josephson c42 in xy pointed at the monitors and routed to from the pres to the line in of the mac ( as advised) and testing L then R. My mic levels are as close to clipping on input as I can get within a few db (and I still get the message that my levels are too low - REW reports levels as low as -79 db and the highest at -37 db even with consistent levels from the mic pre's). i'm running the test in full range and my monitor system is 2.1. The output level from REW is set to -20 and the line out of the card is set to nominal on my audio interface ( output from my Fireface 800 and input to mac via Fireface sends) Mac Os 10.5.8 mac pro 2.8 ghz

AFAIK my JRE/ Java is the latest released.

Here are the a basics of the room as a cross reference for anyone looking at the waterfall plots

L1 221" - 18'5"
L2 266" - 22' 2' there is a closet / alcove in the left rear of the room 24" w
W 94" - 7'10"
H 79" - 6'7"


100 (1) 25.5 51 76.5 102 127
100 (2) 30 61 92 123 153
010 72 144 216
001 85 171 256

110 (1) 76 153 230
110 (2) 78 156 235
101 (1) 89 179 268
101 (2) 91 182 273
011 112 224 336


Apologies in advance if I have totally missed some details in the manual or here.



516 Posts
Just a couple of comments.

First, there is an issue regarding Apple and Java and the use of multi-channel I/O as well as Firewire issues.

That being said...to the best of my knowledge, REW is a single channel FFT system - raising issues as to how you have configured the microphones.

And modal calculators are are by no means references as they are based upon grossly overly-simplified ideal assumptions - and they are in all cases trumped by measurements that measure what is ACTUALLY occurring. Any agreement, especially without knowing anything about the nature of the bounded system is coincidental.

What is the point of using an x-y stereo mic configuration? If you want to do inter-aural cross correlation (IACC) measurements, use a program such as ARTA that actually supports such measurements. Simply employing unsupported techniques in the assumption that they are somehow better is not best practice. A simple accurate omni-directional mic is the preferred configuration.

Also, the C42 is a cardioid mic, not an omni mic. You want to use a single omni mic. A stereo pair is not an appropriate substitute as they do not address the fundamental issue of destructive superposition..

Now to the actual configuration and the tests themselves.

You provide no information as to what sound card you are using and how it interfaces with the computer - single or dual channel, USB or Firewire.... How the 2 mics are configured...in addition to wanting to use a single omni microphone.

You mention the mics are pointed at the monitors - plural.

Exactly what "monitors" are you driving? A sub, multiple subs? 2 full range speakers? 2 full range in addition to 1-2 subs? Are one or both being driven?

There are lots of options, but you have tossed quite a bit at us without many pertinent details, and simply run on ahead wanting interpretations of results without addressing and resolving fundamental issues. (But one thing I can tell you now is that the z-axis 'time slice' over which you are attempting to examine the day is insufficient - you would want to open the time scale up to at least 1000ms...)
Quite a bit of help is available, but first we need to slow down a bit and address fundamental configuration issues, and then proceed with using best practice measurement techniques.

Then one can better move to interpreting results.

5 Posts
Discussion Starter #4

Thanks for the detailed explanation.

1) im using cardiod mics because that is what i have - i don't have any omni mics. The x y configuration was chosen so the mics position was "fixed" in space ( i read somewhere on one of the primers that pointing the mics at the monitor cone was ideal positioning so i chose xy placement ( its not a perfect x/y placement actually slightly spaced so each mic was pointed at the cone) at mix position so i could leave the mics in a position on the stand and run the test 2x 1 for th left channel then the right. The only tests i have run are with the mics placed at ear height at mix position. Eventually I will move the mic stand around in the room ( marking the placements each time and adjusting mic positon to be directed at the cone of the chosen channel) obtain some results then move the monitors postion retest etc and try to extrapolate as much information as best as i can given my resources. Is that logic overlooking or misunderstanding something? I was not using both inputs at the same time on the mics - sorry i was not clear

2) The recording chain ( i touched on but sorry for being unclear) The output is bussed from REW to the fireface -> to monitors which are 2.1 configuration ( specifically a krk rp 10 sub, internally crossover 90 hz-> HHB circle 5 active monitors). The sub is on an Auralex "gramma" decoupling pad. The monitors are on quick lock stands and placed in an isosoles triangle with the mix position at 82" from the front wall ( approx 3/8 distance from from wall relative to full room lenght. The mics are using the preamps from the fireface and then bussed to the Apple internal audio i/o via an unbalanced 1/4 inch: to rca cable pair and a mini jack adapter and the Apple audio is set to full level via Audio midi setup and REW has the apple audio return selected as its input source ( the REW output souce is set to default as neither the specific devices audio devices show up - im sure iwill need to change this to the apple internal audio by chosing this in the audio midi setup to eliminate the multichannel firewire interface from the computer setup given the information regarding the java issue)

REW is set to Full Sweep 20 - 20000, files are left 1 is the left channel recorded and right is the right.

3) The first posted waterfalls are set to 1000 ms but I added the second ones to the next post as the instructions for adding waterfall here on the forum recommended a shorter time.

4) yes the modal calculator was used as a reference point to check my measurements. The tests i did were yeilding such varied results so I was looking at that data to see if there were correlations that were consistent between tests and theoretical. I kept getting error messages from REW so I was unsure whether the tests were in error or natural variations due to the level changes i was making or a user error in terms of the program or methodology. I would get a number of messages a) levels are too low so i would increase the level as much as possible without clipping anything on mic pres or interface or REW then get b) clipping is occurring - i would adjust again then get the message c) impulse response is not alligned ... etc etc.

The modall results and the test calculations are posted in case there are glaring errors in the data set That someoone more knowledgeable may instantly recognize.

Hopefully I have cleared up my ealier post.

Thanks again for your time


Premium Member
1,400 Posts
> You should post a screen shot of your ( adhoc , 2 sound-card ) calibration .

> Loop the output of the "FireFace 800" back to the input of your "FireFace PreAmp / Mac Soundcard" kludge ( & then run the calibration ).

> Don't expect anything useful to be derived from REW until you have achieved a reasonably clean "loopback" .

> A decent SC calibration needs to resemble one of the following 2 examples ;


5 Posts
Discussion Starter #6

i will do that in future but in the short term it seems that fine tuning the system by calibrating my i/o is putting the cart before the horse considering im not sure if my measurement files are corrupted. or my logic has errors.
Correct? or do you suspect something else is up there... Hm ok ill do that after the following.

i will try to test by just using the apple audio i/O an see if i still get errors.

Thanks for the tips

REW Author
6,711 Posts
The plots you attached are not valid measurements. The idea of the loopback test is not so much to fine tune the corrections but it is much easier to see whether the result is valid and so track down whatever problems may exist in the audio and soundcard configuration.
1 - 8 of 8 Posts