REW Update

Help Support GroupDIY:

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,662
Location
Marcelland
Defining it is easy.
I don't think so. Q is a mathematical abstraction, that correlates only distantly with the very object of equalization. I think BW is much more evident. But it can't use the historic definition of BW either.
The audio community has more or less settle on a definition of boost/cut for shelving EQ's; it is based on a number of assumptions, but it works. Why couldn't it settle on a imilar agreement for BW in peak EQ's. The only people I've seen using the academic definition of Q in peaking EQ's are DSP programmers, who generally end up with a total cock-up, that needs to be corrected by people who understand audio.
Justifying the definition is the hard part. Let's say 3dB.
Why choose the value that just doesn't work?
 

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
13,544
Location
Norfolk - UK
I don't think so. Q is a mathematical abstraction, that correlates only distantly with the very object of equalization. I think BW is much more evident. But it can't use the historic definition of BW either.
The audio community has more or less settle on a definition of boost/cut for shelving EQ's; it is based on a number of assumptions, but it works. Why couldn't it settle on a imilar agreement for BW in peak EQ's. The only people I've seen using the academic definition of Q in peaking EQ's are DSP programmers, who generally end up with a total cock-up, that needs to be corrected by people who understand audio.

Why choose the value that just doesn't work?
I think we are talking at cross purposes. You asked me to define significant in the context of how far past the normal -3dB point that is used to define Q does the filter response need to go. I said define significant as 3dB - so the response would have to change by a total of 6dB from the peak. I also said justifying it would be the hard part. I agree it would be nice if there was a generally agreed definition of the parameters of a peaking EQ.

Cheers

Ian
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,662
Location
Marcelland
so the response would have to change by a total of 6dB from the peak.
OK, now I understand. Actually, it's close to the definition I have used for years, that is defining the BW from teh points situated at an arbitrary fraction of the boost or cut. For example, for a 1/3 octave graphic EQ with a boost of 12dB, I used to take the points at +3dB, because it made sense in terms of perception. Using the (inappropriate) academic definition would take the 9dB points resulting in a misleading figure suggesting the filters were much narrower.
This method does not solve all issues since the same EQ set at 6 dB boost would be perceived as narrower because the audibility of the skirts is lesser at 6db than at 12dB.
I also said justifying it would be the hard part.
That's correct. It has to be arbitrary, so it requires a will to reconcile different opinions. I believe that's why the AES tentative committee didn't achieve its task.
 

trobbins

Well-known member
Joined
Jul 1, 2011
Messages
261
Location
Australia
Further to my post #200, I have prepared an article detailing the measurement setups and results I made from 3 contemporary Williamson amp output transformers. For those transformers I did have to go beyond the automated REW capabilities, but in general I found it a breeze to use REW to quickly make an impedance plot and a frequency spectrum plot of any output transformer that is on my bench, and then to interrogate those results for particular parameters like primary winding and leakage inductances, shunt capacitance, resonances, nominal bandwidth including gain/phase dips and bumps that would impact feedback stability margins, and core distortion.
https://www.dalmura.com.au/static/Williamson output transformer measurements.pdf
 

rock soderstrom

⚽️⚽️⚽️⚽️⚽️
Joined
Oct 14, 2009
Messages
1,158
Location
Berlin
Further to my post #200, I have prepared an article detailing the measurement setups and results I made from 3 contemporary Williamson amp output transformers. For those transformers I did have to go beyond the automated REW capabilities, but in general I found it a breeze to use REW to quickly make an impedance plot and a frequency spectrum plot of any output transformer that is on my bench, and then to interrogate those results for particular parameters like primary winding and leakage inductances, shunt capacitance, resonances, nominal bandwidth including gain/phase dips and bumps that would impact feedback stability margins, and core distortion.
https://www.dalmura.com.au/static/Williamson output transformer measurements.pdf
I have only briefly skimmed your "thesis", but I will study it in more detail later. Many thanks for your detailed work.👍
 

warpie

Well-known member
Joined
Feb 7, 2009
Messages
1,218
Is it possible to use dual input/output on REW? Can't see such option in the "preference" tab.
 

trobbins

Well-known member
Joined
Jul 1, 2011
Messages
261
Location
Australia
What do you want to do? The impedance function uses 2 inputs and 1 output. There is a pro version that allows multiple input channels. The help section and some searches on the REW forum are a good idea.

With respect to equalisers and 'Q', there is a new update that includes that capability.
 

warpie

Well-known member
Joined
Feb 7, 2009
Messages
1,218
I would like to simultaneously test the frequency response of my stereo/ dual channel gear. Is this only available on the pro version?
 

warpie

Well-known member
Joined
Feb 7, 2009
Messages
1,218
I think so, but perhaps search through the REW forum, or post on that forum.
https://www.avnirvana.com/forums/official-rew-room-eq-wizard-support-forum.10/

Is there really a need for simultaneous measurement for a frequency response? I can see that performance measurement of things like cross-talk may be interesting.

Well, frequency plots is only an example.

After searching a bit on the avnirvana I think it's not possible. At least on mac and without the pro version.
 

trobbins

Well-known member
Joined
Jul 1, 2011
Messages
261
Location
Australia
It's worth noting that REW can be pretty processor intensive (or perhaps comms intensive if using USB2) even for one channel, especially for RTA with long FFT lengths or for measurements with long sweep times. It would almost be worth using 2 PC's if time or refresh rate was a key issue, and that could also side-step the need for running two channels on the one instance of REW.
 

Latest posts

Top