Need for understanding something and request for implementation of a feature

[anaguma.san]

Hello,

First of all, I'd like to express my gratitude to Mr John Mulcahy for making REW such a great tool! I know it's a lot of effort - I simply love it!

Recently I've been using REW to capture and perform analysis of IR's. One of the things I've been looking into is how the IR will affect the audio when using it in convolution.

Question 1: There are two units: dBFS and dBr. From my findings, dBr describes how much signal will be attenuated. I.e., if we load the ideal dirac impulse with an amplitude of 1, it will be 0 dBr over the whole frequency range. Loading any other filter generated in rePhase will show the correct frequency response using dBr. dBFS seems to be 3 dB lower (as in crest factor for sine wave). What is the actual difference between dBFS and dBr, am I right to think that dBr will always tell me how much IR will change each frequency if it is applied as a filter to the signal?

I know that John is very busy and he is developing REW in his spare time. If this is something he thinks is useful I will highly appreciate these changes.

Feature request 1: Would it be possible to change the gain of IR in REW? When adding the offset, of course it doesn't affect the IR itself and dBFS/dBr figures stay the same. Having an option to apply gain for example in "Impulse" view would be very useful. For now, I've been saving IR's in highest accuracy and boosting them outside of REW using sox (I've made silly mistakes on many occasions). Another use case would be when a few measurements were done and some of them had PAD enabled. IR could be then boosted to compensate for this mistake.

Feature request 2: In "SPL & Phase" view, I can use Shift+LMB to calculate the metrics (Average/Min/Max) and parameters of slope for frequency response. Would it be possible to add the total RMS energy for such frequency range? For example, selecting the whole measurement would result in the total SPL (Leq)? Also in case of "Impulse" window a similar metrics view could be added with RMS/Max/Average energy. A use case I have for it is for example when estimating the Impulse peak to noise ratio for IR. Being able to select the time range and obtain the energies would be really helpful.

Kind regards!

 

[John Mulcahy]

All decibel measurements are relative to something. For dBFS that something is full scale, the AES defines 0 dBFS as the rms level of a sine wave whose peaks are at full scale. That differs from a more mathematical definition for such a signal which would be - 3.01 dBFS, on that scale a square wave at full scale would be 0 dBFS whilst the AES would say that is +3.01 dBFS. REW has an option to choose which definition to use.

dBr typically means dB relative to the peak of a signal, whatever that peak may happen to be.

Neither helps much when exporting an IR for use in convolution, best export the IR as 32-bit float and scale as needed for unity gain at the required frequency.

I'm away at the moment but will look at your feature suggestions when I get back.