REW Update

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
trobbins said:
Having some fun with REW ... with the aim of making almost distortionless sinewave test signal by adding in harmonics to null out harmonics generated in the line output of a soundcard (or even an amplifier under test).

My EMU0404 USB provides about a raw 0.0009% THD at 2Vrms from its line output.  REW can be used to null the harmonics in that test signal to lower THD to around 0.00004% using a simple loopback test.
I will have to give REW a try again but my preferred method for audio analysis for the last few years has been to simply record a stimulus through the DUT into a hi res wav file and then analyze and plot with GNU Octave.

At one point I had a procedure to make a loopback recording and then subtract that from each DUT measurement. The idea was that this would cancel all of the imperfections of the test rig in one pass. Unfortunately I don't have any Octave code for calculating phase and THD and other such things so I don't know if it was as effective as getting down to quad-zero THD. That is impressive.
 
All you REW users I need your help :)

I gave it a go today and I like it so far. I only run some basic frequency responses though.

Do you guys know if it's possible to somehow "lock" the cursor to the plot itself? I find pretty hard, annoying and time consuming trying to manually position the cursor on the plot line. Is there any way to lock the cursor on the x axis?

I'm not sure how else to describe what I'm looking for. I hope it makes some sense but if not I'll try to explain it a bit better.
 
I do not know of any way of doing this. It would be nice if the cursor followed the response like it does on LTspice.

Cheers

ian
Shame. I took for granted that there is the option.

Another thing I don't like is that you have to go through a few clicks in order to save an existing project.
 
warpie, not sure what you mean by 'project'?

Iif you are wanting to save a 'measurement' then the floppy disk symbol in the LHS listing of measurements will save the full details, as compared to using the camera symbol 'capture' which only saves what is displayed. And the File menu has a save all function.

In a Measure plot, I have not found a way to turn off the cursor lines/values - for example if I want a clean plot without cursors showing.

One hassle that may show up over time is that you may need the REW version used to make the data files to open the data files - or at least a later version.
 
warpie, not sure what you mean by 'project'?

Iif you are wanting to save a 'measurement' then the floppy disk symbol in the LHS listing of measurements will save the full details, as compared to using the camera symbol 'capture' which only saves what is displayed. And the File menu has a save all function.

In a Measure plot, I have not found a way to turn off the cursor lines/values - for example if I want a clean plot without cursors showing.

One hassle that may show up over time is that you may need the REW version used to make the data files to open the data files - or at least a later version.

Yes I mean the "save all" floppy icon. First it asks you whether you want to save any notes, then you have to choose the file you want to overwrite, then it asks you twice whether you want to overwrite the existing .mdat file. That's four clicks in order to overwite an existing file while in most software is simply ctr+s.
 
One day I hope REW adds :

- calibrate-able dBu unit for RTA amplitude
- intermodulation distortion test measures
- a second 'measurement cursor' with the difference between them displayed ... like my Tektronix CRO :)
I have found John to be extremely receptive to suggestions. Right after I mentioned unpractical limitations to the stepped sine measurements and lack of dBu measurements in the Level panel, he implemented it.
 
warpie, the floppy icon just relates to a single measurement. The 'save all' in the file menu saves more than one measurement, so imho it is quite appropriate to ask for notes to save time in a year or two when you've forgotten what the full set of measurements was for. If you then want to save to an existing file name then again imho that is quite appropriate to 'save' some newb from overwriting a lot of work. I don't get two pop ups about over-writing - are you using the latest beta software?

Best you use the REW forum to suggest any possible function tweak, such as regards cursor operation - but first search the forum.
 
warpie, the floppy icon just relates to a single measurement. The 'save all' in the file menu saves more than one measurement, so imho it is quite appropriate to ask for notes to save time in a year or two when you've forgotten what the full set of measurements was for. If you then want to save to an existing file name then again imho that is quite appropriate to 'save' some newb from overwriting a lot of work. I don't get two pop ups about over-writing - are you using the latest beta software?
Best you use the REW forum to suggest any possible function tweak, such as regards cursor operation - but first search the forum.

I'm using 5.19 which I believe it's not the latest version. I'm on high sierra.
 
REW is unlike pretty much any other advanced app that I've come across, as it not only has an active forum to bring up issues and bugs and possible enhancements, but John is full on with continuous improvement not matter what the use or feature (out of literally hundreds) and before you know it a forum suggestion or bug gets turned around in the next quick release. For my post #200, John gave me a link to a draft version that goes down to 0.1Hz, so I could push a measurement below the normal 2Hz limit and hence characterise an inductance measurement better.
 
It's amazing that it's freeware. You don't see such level of support even in paid software.
 
Oh, and automatic Q measurement of EQ plots - and Bode plots with phase response as well as amplitude.

Cheers

Ian
It's a very good idea Ian, the phase diagram requires that there are two inputs to be compared by the software, one directly from the generator and one from the DUT, normally we need a splitter and a piece of software that can read the phase difference from two inputs (the reference and the signal coming from the DUT) ... REW would need to read the signal arriving on the sound card input (from the DUT) and the signal generated by the sound card (I think directly from the USB port) and sincerely I don't know if this is possible (I admit my ignorance). A small calibration routine will also be needed for phase zeroing. We could ask this question on the official REW forum.

Cheers
JM
 
"ruffrecords said:
Oh, and automatic Q measurement of EQ plots "


For that it would be necessary first to define what is Q in an EQ.
The usual definition of Q is adequate only for filters that slope to minus infinite. An EQ slopes to unity.
In other words, Q is an inappropriate measure of an EQ's selectivity. I much prefer a definition based on perceived effect.
 
"ruffrecords said:
Oh, and automatic Q measurement of EQ plots "

For that it would be necessary first to define what is Q in an EQ.
The usual definition of Q is adequate only for filters that slope to minus infinite. An EQ slopes to unity.
In other words, Q is an inappropriate measure of an EQ's selectivity. I much prefer a definition based on perceived effect.
I would define Q as the inverse fractional bandwidth at the half power points. This does not requires the filter to slope to infinity. It only has meaning if the filter response falls significantly either side of the half power points so it can only be sensibly measured at maximum boost or cut. Easy enough to find in an LTspice sim so should be possible in REW.

Cheers

Ian
 

Attachments

  • Screenshot from 2021-05-07 21-56-48.png
    Screenshot from 2021-05-07 21-56-48.png
    36.7 KB · Views: 8
Which results in Q=0 for an EQ that has a max boost of 3dB; I can only guess what Q is when the max boost is less than 3dB.
Of course, which is why I added the caveat that it only has meaning if the filter response falls significantly either side of the half power points. A filter with a max boost of 3dB clearly does not meet this criteria.

Cheers

Ian
 

Latest posts

Back
Top