TOPIC:

I have recently purchased a miniDSP and have been using REW for about 4 weeks and both are excellent products. I have a Dayton Audio calibration mic , balanced <-> USB preamp and both mic and soundcard calibration files have been added. I have generated a Pink Noise ( periodic noise ) CD so I can test with the RTA ( with PN full range ) or a log sweep.

Test A: Test with log sweep.
Test B: Test with RTA and PN

Each test gives very different results above about 2 kHz.

Is there any reason for this. Some articles I have read suggest Pink Noise ( periodic noise ) should be used for room EQ and not log or linear sweeps.

Any advice is appreciated.
Thanks,
Roger This email address is being protected from spambots. You need JavaScript enabled to view it.
Toowong Australia.

Go for the log sweep.

With PN you have to make sure the clocks (playback and recording) are correct and the FFT's should be equal length and synchronized with the recording (one FFT for one PN sequence).

Some Background: Transfer function measurement with sweeps

>Go for the log sweep.
>With PN you have to make sure the clocks (playback and recording) are
>correct and the FFT's should be equal length and synchronized with the
>recording (one FFT for one PN sequence).

I wish to do a confidence test of REW on my laptop (same mic and speaker) with either log sweep or pink periodic noise. The only thing which changes is the source. The first source is a log sweep. The second source is pink period noise recorded to a CD with an FFT of 65536 ( generated with REW ) which is the same the REW RTA listens with , yet the db vs freq graphs are very different below 2 k HZ. So I beleive the FFT when generated and FFT when RTA listens is the same however there is at least 20 db difference in the result when compared to a log sweep.

I was hoping that a log sweep and RTA with pink period noise all within REW on the same laptop mic and speaker would achieve the same result but the differnce is about 20db below 2 kHz.

I would be pleased if you could explain further.

With appreciation.
Roger

If you want an RTA and a sweep measurement to have the same displayed levels you need to tick the "Adjust RTA Levels" box in the RTA graph controls. RTA levels depend on RTA resolution, the higher the resolution, the narrower each RTA bin, and the less energy there is in that bin hence the lower the displayed level. If you tick the box REW will automatically compensate for the bin width effect and show levels that correspond to sweep measurement levels under similar conditions.

I have to use white noise to get a 'flat' response in a FFT display.
Pink noise gives 3dB/oct roll-off when analysed with a FFT.
Is this the effect you see?

I do not see the same as you,

From what i understand ...
white noise has the same spl or voltage across the freq spectrum
pink noise has the same energy or watts across the freq spectrum

so if a cd plays white noise and a dummy load is connected to the amplifier it should produce a linear amplification for all frequencies ie .1 Vpp x 20 gain = 2 Vpp from 20 to 20,000 Hz (in theory)

however our ears behave logarithmicly to the average energy level

so if a cd plays pink noise , is amplified and played through speakers (not dummy load)
and "listened" to with a calibration mic and analysed with FFT / REW then we want to see a flat response which matches the behaviour of our ears.

Is this your interpretation too ?

cheers,

Roger.

It all depends on how you measure the SPL or power/energy.
If you look with the same bandwidth at each frequency (like a FFT), white noise will give you the same level/SPL/power/energy for each frequency.
If you measure with 1/n octave bandwidth bands e.g. 20-40Hz, 40-80Hz, 80-160Hz etc. then pink noise will give you the same value for each frequency(band).

I am really sorry for reviving this old topic,
I just have been trying to tackle the bright stage I get with PN vs Sweeps, and never found a mention of this till now.
I have spent months trying to tune my Car with PN, I followed many tutorials . I equalized each component with DSP rew and UMIK1, according to a house curve ( downward sloped ), the results with PN were to bright to my ears and lacking lower end KHz.
But whenever I tuned with sweeps, the end result was just great, really linear, much more linear.
I don't know how to explain it.

After I tuned with Sweeps in REW, I tried to measure with PN to see how I should have tuned with PN, and the result was much more sloped measure.
Flat curve tuned with Sweeps, was really sloped with PN only above a certain frequency ( they were the same up to somewhere between 2-3khz)
In short, with PN i always got really high dbs above 2-3khz,

since then I tune with Sweeps, and the results are consistent ( sweeps are less comfortable..... but end result worths it )