Trace Arithmetic Issue

Thread Starter
Joined
Sep 2, 2020
Posts
49
More  
Preamp, Processor or Receiver
Denon AVR-4700
Main Amp
Neurochrome Modulus 686 for music
Additional Amp
Topping Pre90 preamp for music
DAC
SMSL M500
Computer Audio
Roon
Universal / Blu-ray / CD Player
Panasonic DP UB-820; Oppo UDP-203
Streaming Equipment
Nvidia Shield Pro; Roopiee Raspberry Pi
Streaming Subscriptions
Tidal
Front Speakers
PSB Imagine X2T, Magnepan 1.7i (for music)
Center Channel Speaker
PSB Imagine XC
Surround Speakers
PSB Alpha
Surround Back Speakers
PSB Alpha Mini
Front Height Speakers
Totem KIN IC6 Slim (Front O/H)
Rear Height Speakers
Totem KIN IC6 Slim (Rear O/H))
Subwoofers
4 sealed subs with 12" drivers custom built boxes
Other Speakers
Mini DSP 2x4 Balanced; HD Fury Diva; MSO sub tunin
Screen
EluneVision 96" Acoustically Transparent
Video Display Device
JVC DLA X500
Remote Control
Harmony Elite
Other Equipment
Thorens TD-160 Super, SME tonearm, Shure V15 IV with Jico SAS stylus, Arcam rPhono preamp
Hi, I have seen an issue with trace arithmetic. As per a previous post, I have been using |A|/|B| to normalize a speaker response. I have been using the most recent Beta to correct for an import error on sweep import. I recently had a crash when trying to save some jpeg images, to the point where the icon was removed, the executable file appeared to be removed, and I had to re-install from the previously downloaded file.

Seemed to be working OK, but today when I went to generate some new equalized curves, I went through the procedure of subtracting the 1kHz level to get a 0dB normalized curve, then using this to apply |A|/|B| to the measured curve. Instead of making no difference to the curve at 1kHz, it subtracted the original value of the curve. See attached data file. I have seen this behavior before sporatically and relaunched REW. THis time seems more persistent. Any input appreciated.
 
I don't understand what you consider to be wrong about that result? Dividing a response by something that is 15 dB louder will give a result around -15 dB. If you want a result around 0 dB align the levels of the responses before dividing them or shift the result to be at 0 dB in your chosen region.
 
Thanks, my mistake. I was sure I had subtracted the 1kHz value...but not. All good now.
 
Back
Top