Jump to content
Sign in to follow this  
SwissBear

Using REW and rephase to generate amplitude and time domain corrections

Rate this topic

Recommended Posts

22 hours ago, pkane2001 said:

I just looked at your measurements file. It appears that you've imported the data as an impulse response into REW? Can you try importing as a frequency response?

 

When I use the export measurement function in Audiolense, the measurements are exported in .pcm format. From the REW Help section it looked like this was ideal:

 

"Full range ETF measurements allow the impulse response to be exported in a .pcm raw data format, using the File -> Write Impulse As *.pcm option. Use REW's File - Import Impulse Response command to load these files. Importing the impulse response allows REW's Spectral Decay and Waterfall plots to be generated from the data (this is not possible when importing frequency response text files)."

 

I didn't try importing as frequency response instead since I couldn't see how to get Audiolense to export in any other format than .pcm.

 

I tried a couple more run throughs with SwissBear's procedure and I always get the flat excess phase. After thinking about it more, I believe the flat excess phase is ok. The flat line is showing the amplitude of the excess phase, not the phase. I believe there shouldn't be any difference in amplitude between the excess phase and my amplitude corrected measurement, thus the flat line I'm seeing. The difference between the two is phase, which is what I'm trying to adjust for in rePhase.

 

On 11/21/2017 at 7:27 AM, lasker98 said:

3) once I have the excess phase corrected for phase in rePhase, I then load in my original measurement to apply the saved eq amplitude filter. As soon as I load the measurement, before applying the REW filters, the corrected phase in rePhase shifts up, maybe 90 to 180 degrees. Does that sound normal? Should I be readjusting the phase to correct for that shift before generating the final impulse?

 

As for the phase shifting once I load in the original measurement for final correction, that also seems normal. It would seem the shift is because the phase corrections from the previous excess phase correction are still loaded in rephase, which are now being applied to the original measurement, which is exactly what is intended to remove the excess phase from the original measurement. As for my last sentence about readjusting for that shift, I would say definitely not, since that would be adding excess phase back in.

 

SwissBear's guide is showing how to do PEQ on the original measurement in REW, then how to correct the original measurement for excess phase only in rePhase, then applying that excess phase correction and the REW PEQ to the original measurement in rephase to create the impulse used for convolution. It's quite an elegant approach imo, since instead of looking at the phase of the original measurement and trying to figure out how and what to correct, we're working only on the excess phase, which is exactly the phase component we want to eliminate to end up with as close to minimum phase as possible.

 

Thanks again to all for your help.

Share this post


Link to post
Share on other sites
5 hours ago, lasker98 said:

I tried a couple more run throughs with SwissBear's procedure and I always get the flat excess phase. After thinking about it more, I believe the flat excess phase is ok. The flat line is showing the amplitude of the excess phase, not the phase. I believe there shouldn't be any difference in amplitude between the excess phase and my amplitude corrected measurement, thus the flat line I'm seeing. The difference between the two is phase, which is what I'm trying to adjust for in rePhase.

 

 

Excess phase is the difference between phase of the amplitude corrected curve and minimum phase, in degrees. If excess phase appears as a flat line at zero, that means your amplitude corrections already produced a minimum phase response and no more adjustment is necessary. I'll try to go through the steps with your data to see if I get the same flat line as you.

 

 

 

Share this post


Link to post
Share on other sites
21 minutes ago, lasker98 said:

That would be great. Thank you.

 

I'm seeing non-zero phase, and it's definitely not flat. How are you generating your PEQ filters? Are you doing this in REW?

 

Here's what excess phase looks like, when I generate it from your data,  loaded in rePhase. I let REW pick the PEQ filters automatically. You can see that it's not at all flat:

 

phase.thumb.JPG.5e3ee67ed61476553f941f16b9fe5dd8.JPG

 

Share this post


Link to post
Share on other sites

Very odd. That looks more like eq'd measurement before anything in rePhase. Here's screenshots of my full REW graphs window. You can kind of see the shapes in the thumbnails in the left sidebar. I've also included the excess phase as imported into rePhase prior to any corrections being applied (phase or frequency). I'm using 6 cycles FDW and psychoacoustic smoothing for the REW EQ.

 

5a15f05f8e4d0__09measurements.thumb.jpg.ad02e2aedf186bd4630af5b5e3a6c8e6.jpg 

 

5a15f06fb7db0__09EP_rePhase.thumb.jpg.28d91eb6137eb317da186cd275e3e46b.jpg

Share this post


Link to post
Share on other sites
4 minutes ago, lasker98 said:

Very odd. That looks more like eq'd measurement before anything in rePhase. Here's screenshots of my full REW graphs window. You can kind of see the shapes in the thumbnails in the left sidebar. I've also included the excess phase as imported into rePhase prior to any corrections being applied (phase or frequency). I'm using 6 cycles FDW and psychoacoustic smoothing for the REW EQ.

 

5a15f05f8e4d0__09measurements.thumb.jpg.ad02e2aedf186bd4630af5b5e3a6c8e6.jpg 

 

5a15f06fb7db0__09EP_rePhase.thumb.jpg.28d91eb6137eb317da186cd275e3e46b.jpg

Can you upload the above measurements file including excess phase ?

Share this post


Link to post
Share on other sites

Just a note for clarification on those measurements. The Left_original and Right_original are the imported measurements from Audiolense that were in the previously posted .mdat measurement file. They have no windowing or smoothing when imported, that's applied after importing into REW. 

 

In the .mdat file I just uploaded, those have the 6 cycles FDW and Psychoacoustic smoothing applied, as they were used in auto generating the REW EQ filters.

Share this post


Link to post
Share on other sites
1 hour ago, lasker98 said:

Just a note for clarification on those measurements. The Left_original and Right_original are the imported measurements from Audiolense that were in the previously posted .mdat measurement file. They have no windowing or smoothing when imported, that's applied after importing into REW. 

 

In the .mdat file I just uploaded, those have the 6 cycles FDW and Psychoacoustic smoothing applied, as they were used in auto generating the REW EQ filters.

 

OK, now I'm getting a result much closer to the one you posted. Although it's not perfectly flat, it's close. I think that's OK. After all, you are correcting phase and not amplitude in rePhase (I think that's what you said earlier :))

 

phase_r.thumb.JPG.4f3dcca8f7060e86f09c5e646d0238c6.JPG

 

Share this post


Link to post
Share on other sites

That's much closer for sure. I appreciate you taking the time to look at my measurements. Would you mind posting the measurements you did to end up with the excess phase you posted? I'm very curious about what you did to get the one dip around 270 Hz.

Thanks.

Share this post


Link to post
Share on other sites
6 hours ago, lasker98 said:

That's much closer for sure. I appreciate you taking the time to look at my measurements. Would you mind posting the measurements you did to end up with the excess phase you posted? I'm very curious about what you did to get the one dip around 270 Hz.

Thanks.

 

That dip surprised me, but it was there no matter what I tried. I'll have to reproduce the measurements, as I don't think I saved them. I'll post them as soon as I get a chance (and recover from food coma).

 

Share this post


Link to post
Share on other sites
13 hours ago, lasker98 said:

That's much closer for sure. I appreciate you taking the time to look at my measurements. Would you mind posting the measurements you did to end up with the excess phase you posted? I'm very curious about what you did to get the one dip around 270 Hz.

Thanks.

 

You know, I think there might be a bug in REW. Just one time I opened your measurements file and was able to generate a completely flat excess phase amplitude plot, just like the one you posted. But never again.

 

I wanted to repeat the steps to make sure I did everything right, and every single time after this, I got the same dip that I saw before. I could swear that I repeated all the same steps, used only your measurements data, and yet, I couldn't reproduce the completely flat line again. Here's the measurements file with the dip:

 

dip.mdat.zip

 

Share this post


Link to post
Share on other sites

There's something odd going on. I load your measurements and see the dip on the "A times B-EP" ( I renamed it to "R_A tmes B-EP" so I was able to make new ones), although the dip is much less than on the screenshot you posted Wednesday. I generate a new "A times B" using the exact same Right_original and R-amplitude_09 and then from that generate a new excess phase version and the result is completely flat. I have no idea why we're getting different results using the same measurements. I'm using REW V5.19 Beta 7 in case that's different from the version you're on. I'm also running Windows 10.

 

I'm attaching my measurements which are the ones you uploaded plus the newly generated A times B and A times B-EP which were created the same as you originally did on the ones I renamed.

 

pkane_dip_02.mdat

Share this post


Link to post
Share on other sites
2 hours ago, lasker98 said:

I'm using REW V5.19 Beta 7 in case that's different from the version you're on. I'm also running Windows 10.

 

That's probably it. I'm using a Mac and REW 5.19 Beta 7a. The difference might be between the Windows and Mac versions. 

Share this post


Link to post
Share on other sites

Guys, could you please explain what to do starting from the end of p. 3 of the guide:

 

"The excess phase, which has been generated, can now be exported as .txt to rePhase,
where it will serve to make the time domain correction."

 

I load the excess phase (I received it as an almost flat line with a small dip on 200hz) to rePhase.. then what? What do I have to do in p.4? I'm completely lost there, could you please tell me next steps, in what buttons to press manner?

Share this post


Link to post
Share on other sites
On 5/2/2018 at 3:50 PM, RAzZin said:

Guys, could you please explain what to do starting from the end of p. 3 of the guide:

 

"The excess phase, which has been generated, can now be exported as .txt to rePhase,
where it will serve to make the time domain correction."

 

I load the excess phase (I received it as an almost flat line with a small dip on 200hz) to rePhase.. then what? What do I have to do in p.4? I'm completely lost there, could you please tell me next steps, in what buttons to press manner?

 

If your excess phase is at zero degrees and flat, there's no real correction needed. If excess phase deviates from flat, you'll then need to use Rephase filter banks to adjust it so that it's nearly flat and near zero, load your frequency EQ filters into the same session and then generate an impulse response file from Rephase that has both, frequency and phase corrections.

 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×
×
  • Create New...