Graph of SPL difference between two setups?

Max Rockbin

Registered
Thread Starter
Joined
Aug 22, 2021
Messages
4
More  
Preamp, Processor or Receiver
miniDSP DDRC-24
Main Amp
pioneer sc-91 AVR
Additional Amp
active speakers
Front Speakers
Celestion SL600
Subwoofers
Hsu VTF-1
Other Speakers or Equipment
JBL LSR305
I know a lot of people make small changes to their setup and then compare the graphs in REW to see if it made things better.

I'd like to know if there is a way in REW to see a graph of just the difference between two SPL graphs. I think that'd be a really useful quick way to see what actually changed.
For example, here are graphs of before (green) and after (red) moving the location of one of my bass traps and moving my sub (just have one) close to the trap.
It'd be great if, like the generate average button, there was a generate difference button.

For reference - this is of a Genelec 8030c right speaker + sub. Mic in main listening position 4 1/2' from speaker. The bass trap is 2' x 4' x 17" of rockwool positioned in the corner of the room.
old position - flat against the back wall, oriented horizontally - long side on the floor. New position, diagonally in the back left corner, creating a triangular cavity behind the trap - about 16" on a side (top of the triangle & trap covered with wood.) I had read that by creating a cavity like that, you greatly increase the effective thickness of the trap, making it much better at absorbing low bass.
20210912 moved bass trap graph 8030CwSub rt spkr.jpg

1/24 smoothing:
 

sm52

Member
Joined
Mar 14, 2019
Messages
863
There is such a button. ALL SPL tab, Controls, Trace Arithmetic, A: Graph 1, B: Graph 2, at the bottom select A / B, and Generate.
 

Max Rockbin

Registered
Thread Starter
Joined
Aug 22, 2021
Messages
4
More  
Preamp, Processor or Receiver
miniDSP DDRC-24
Main Amp
pioneer sc-91 AVR
Additional Amp
active speakers
Front Speakers
Celestion SL600
Subwoofers
Hsu VTF-1
Other Speakers or Equipment
JBL LSR305
Thanks! One thing that confuses me...
Can you tell me why I want A/B instead of A-B? It seems to work. A - B gives a graph that looks sort of like one of the original graph with values between about 60 - 90db. A/B gives values mostly +/- 10dB, which is more like what I'd expect for A-B.

I'm going to guess it's because dB are logarithmic, so subtraction is really division? So A/B refers to the absolute (not logarithmic) sound pressure ratio and the result is expressed in dB which is logarithmic. But just intuitively I would have thought A - B of two graphs expressed in dB would be the dB difference, so I'm not sure that's right. Purple is A/B. Blue is A-B for the graphs in my previous post.
45126
 
Last edited:

Max Rockbin

Registered
Thread Starter
Joined
Aug 22, 2021
Messages
4
More  
Preamp, Processor or Receiver
miniDSP DDRC-24
Main Amp
pioneer sc-91 AVR
Additional Amp
active speakers
Front Speakers
Celestion SL600
Subwoofers
Hsu VTF-1
Other Speakers or Equipment
JBL LSR305
I found the answer to my exact question here
Where
All trace arithmetic is done on the linearly scaled values, with conversion back to dB afterwards. Multiplication is analogous to cascading transfer functions - for example, multiplying a drive unit measurement by a filter response will show how the response would look with the filter applied. Addition is analogous to combining sources, as Matthew notes - for example adding response measurements from a sub and a main speaker to see the combined result.
Which is what I kinda guessed. It's not intuitive (to me), but people seem to find it more useful for making filters and such.
 

thothsong

Member
Joined
Jun 29, 2021
Messages
388
I'm still confused about this. Here's a chart of two measurements that are fairly uniformly ~12 dB apart above ~120 Hz, but A/B is showing larger differences above 2 KHz, for example as shown at 9 KHz, 19 dB versus 11.7 dB. Why the difference? (REW 5.20.1)

45482
 

thothsong

Member
Joined
Jun 29, 2021
Messages
388
Different phase responses.
Ah, OK. Can you point to things I should read to better understand that? And in the meantime, to get the difference ignoring phase, is the right thing to do to a) select each measurement individually in turn, b) click Average the Responses to get a magnitude, c) smooth the two resulting magnitudes as desired, and then d) generate A/B using those magnitudes?
 
Top Bottom