ProSoundWeb Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: [1]   Go Down

Author Topic: True Q  (Read 1554 times)

John Roberts {JR}

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 16415
  • Hickory, Mississippi, USA
    • Resotune
True Q
« on: July 22, 2013, 01:15:52 pm »

Even I am growing weary of whining about the lack of concise definition for Q in boost/cut EQ.

Instead of trying to pick a winner from the sundry approaches out there, I propose a different new technology solution to provide an empirical measurement Rosetta Stone of sorts where we can define a new improved Q measurement and then all can work backwards or forwards to map the sundry platforms to agree with mfr's presets.

My suggestion is to use a measurement platform to subtract a baseline signal level measurement, from the equalized signal measurement. This difference signal should contain only the equalization boost (or cut). We then apply the classic Q/bandwidth definition for bandpass sections to this extracted EQ fraction.

This should give us a definitive Q measurement, independent of the platform generating the EQ. The rest is just connecting the sundry dots. 

As long as a manufacturer's EQ  bandwidth presets include information about what DSP platform that applies to. One person somewhere in the world (or the manufacturer) measures that DSP platform at that EQ setting to measure the true Q. Now the end user can use that true Q to adjust his XYZ DSP platform to agree with manufacturer's preset.

or not...  8)

JR
Logged
Don't tune your drums half-ass. Listen to what a properly "cleared" drum sounds like.   http://circularscience.com/

Riley Casey

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1363
Re: True Q
« Reply #1 on: July 22, 2013, 07:40:15 pm »

Gasp!! JR is trying to interfere with someone's Constitutional /Adam Smith / god given right ( you choose ) to have a secret sauce. For shame!   ???


...
Instead of trying to pick a winner from the sundry approaches out there, I propose a different new technology solution to provide an empirical measurement Rosetta Stone of sorts  ...
JR

John Roberts {JR}

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 16415
  • Hickory, Mississippi, USA
    • Resotune
Re: True Q
« Reply #2 on: July 31, 2013, 11:12:33 am »

OK, another idea run up the flagpole that nobody saluted. :-(

Here's one more.   Since measurement systems can save a frequency response sweep to some kind of computer media, it seems a measurement system could also open up a previously saved file. So a manufacturer just needs to publish a readable file of the actual target response they want their speakers to get. Easy peasy...

Of course there probably isn't a standard format for response files, but how many flavors of that are there?

======

Back to my "True Q" new improved empirical measured Q,  I understand that some measurement systems can perform a sweep and then download the data from that sweep into a spreadsheet. If you do before and after (EQ) sweeps of a platform, you can then subtract those two sets of numerical values from each other to extract just the EQ difference signal. This difference signal, even for a small 1dB boost/cut will take the form of a common bandpass, with measurable -3dB points. This extracted EQ component will look like a BPF output and we can determine Q using the classic definition.

For the example of a boost EQ delivering roughly +1dB ( 1.1V) above the flat (1V) nominal signal, the extracted EQ term will peak at approx 0.1V and have -3dB points down around 0.07V.

Since EQ is often layered on top of other filters, this method subtracting two files seems most accurate, but for just measuring EQ relative to flat 0dB one might be able to directly compute it.

JR
Logged
Don't tune your drums half-ass. Listen to what a properly "cleared" drum sounds like.   http://circularscience.com/

ProSoundWeb Community

Re: True Q
« Reply #2 on: July 31, 2013, 11:12:33 am »


Pages: [1]   Go Up
 



Page created in 0.075 seconds with 26 queries.