RTA mode Spectrum vs RTA and changes to SPL.

[Lanion]

I am trying to use a pink noise generator and the RTA feature of REW to make some EQs. I am confused about the difference between the spectrum mode and the RTA mode in the real time analyzer.

In the REW graph attached the green line is taken with the mode set to 'RTA 1/24 octave'. The blue line is with the mode set to 'spectrum' with 1/24th smoothing enabled. The green line is a much flatter frequency response, so clearly it matters which one I pick for the purposes of EQ. Which mode should I use and why?

 

[John Mulcahy]

Use the RTA with pink noise or pink periodic noise (pink periodic noise will provide more stable low frequency levels). Pink noise has constant energy per octave, with an RTA (which has a fractional octave frequency scale) it will have a nominally flat response, its spectrum (which has a linear frequency scale) falls at 10 dB per decade.

 

[Lanion]

Use the RTA with pink noise or pink periodic noise (pink periodic noise will provide more stable low frequency levels). Pink noise has constant energy per octave, with an RTA (which has a fractional octave frequency scale) it will have a nominally flat response, its spectrum (which has a linear frequency scale) falls at 10 dB per decade.

ok, thanks!

 

[bisesik]

Hello! Happy New Year!


Is it possible to change the appearance of the graph which was captured in the RTA Window with the Spectrum mode to view of RTA 1/x octave mode for the existing measurement file? I have mistakenly captured many measurements in the past using pink noise in the Spectrum mode instead of white noise. Any chance?

Thank you!
Ivan.

 

[John Mulcahy]

Is it possible to change the appearance of the graph which was captured in the RTA Window with the Spectrum mode to view of RTA 1/x octave mode for the existing measurement file?

No, not after the RTA capture has been saved as a measurement. You could multiply the data by an inverse pink response to get something similar to an RTA result, using trace arithmetic A*B. Examples to use for multiplication for 48 kHz and 96 kHz data attached.

 

[bisesik]

No, not after the RTA capture has been saved as a measurement. You could multiply the data by an inverse pink response to get something similar to an RTA result, using trace arithmetic A*B. Examples to use for multiplication for 48 kHz and 96 kHz data attached.

Thank you John! It works!

You did my day!
Thanks a lot!