Welcome, Guest
Username: Password: Remember me
  • Page:
  • 1
  • 2

TOPIC: Re: Feature requests and suggestions

Feature requests and suggestions 4 years 10 months ago #21235

  • pos
  • pos's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 350
  • Thank you received: 196
  • Karma: 39
Hello planetti,

Thank you.

"Mixing" corrections is more complicated than it seems when you take windowing and centering into account.
As of now the only way to achieve this is either to chain two convolutions (for example a DRC on the inputs, and a crossover on the outputs), or to use an offline convolution tool such as sox to actually mix both correction into a single FIR with convolving one correction with the other (the order does not matter).
In theory the resulting FIR length (taps) should be the addtions of the two FIR lengths. With sox I don't think this is the case because of the FFT convolution, and that could be a problem. One has to load the impulse into a measurement software such as HOLM to see what the result looks like.
One thing I could do is to include a direct offfline convolution tool into rephase to do such things, in a special "tool" menu.
This would let the user either "mix" two correction, or do an offline convolution of a song (in wav format...).
I could also add other tools in that menu, such a taking two mono FIR and pan-potting thel L/R in a stereo FIR...
https//wavetracing.com | rephase.org
The administrator has disabled public write access.

Feature requests and suggestions 4 years 10 months ago #21246

  • planetti
  • planetti's Avatar
  • Offline
  • Senior Boarder
  • Posts: 45
  • Thank you received: 7
  • Karma: -1
Yes, you already got the point! :woohoo:

And sorry, I'm not a FIR-FFT-calculation expert, nor a programmer. :whistle:

My simple thinking is: rePhase can illustrate graphs of frequency and phase, the user can manipulate the data manually, and then the FIR-filter can be generated. Therefore it may be possible to reverse the calculation?

The idea of the function in rePhase could be, for example:
1) Import any FIR-filter A (count the number of taps, read or ask for the frequency basis, determine the peak, type and necessary parameters)
2) illustrate the graphs of frequency and phase
3) the user does his tuning
4) generating of the user tuned FIR-filter B (which contains the tuning difference only?) in the same format/parameters as filter A
5) convolving A*B

re 1) Import formats like .wav, .txt or miniDSP format [b0=n.nnnnnnn, b1=nnnnnnn, ...]
re 5) As you already mentioned, the convolution function could also be used for offline convolution of a song

That would raise the extraordinary rePhase to the giant killer tool ... ;)
The administrator has disabled public write access.

Feature requests and suggestions 4 years 10 months ago #21257

  • pos
  • pos's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 350
  • Thank you received: 196
  • Karma: 39
These are all good ideas, but unfortunately unlikely to happen as these things take a lot of time... :/
I will see if I can add the offline convolution tool, but that will probably be about it.
https//wavetracing.com | rephase.org
The administrator has disabled public write access.
The following user(s) said Thank You: planetti

Feature requests and suggestions 4 years 9 months ago #21761

  • Richard
  • Richard's Avatar
  • Offline
  • Expert Boarder
  • Posts: 132
  • Thank you received: 71
  • Karma: 13
Hi pos,

rePhase is absolutely brilliant, and we all love it, as you know already. Thank you so much!
Here's a couple of feature suggestions from me...

1.) "Inverse total amplitude & phase shape” reciprocal button
In one quick button click, swap all the “minus –dB” data points into identical “plus +dB” values (and vice versa +dB becomes -dB) for the amplitude curve, and likewise, swap all “negative degree” phase values to “positive degree” phase values (and vice versa.)
This way the entire transfer function can be made into its own reciprocal correction filter in one click! (Or two clicks if “Invert Amplitude” and “Invert Phase” were two independent button functions – maybe also handy!)
Obviously that’s totally different from basic inverting the signal polarity and putting 180 degree phase shift at all frequencies – which I know already exists in rePhase.

I can perform this job already in HOLM using “Manipulation > Inversion (B=1/A)” which creates the reciprocal function in a new trace, and I could just export that trace as another .txt and import directly into rePhase. But it would be great to have the reciprocal built-in as a quick function in rePhase, handy for comparison view back and forth while you’re working in rePhase, without having to keep re-importing another measurement or having running several rePhase windows concurrently.

2.) “Generate FIR automatically from measurement data” button
I seem to import my speaker measurements into rePhase, then spend hours trying to tweak dozens of paragraphic EQ filters and phase warps to bend the shape back to flat again… If I “Bypass” the measurement data, I can see my filter shape is getting closer to the reciprocal of the speaker curve (in another window, or HOLM) but it’s still only a bendy approximation. Not perfect match, even after hours of tweaking and fiddling! If I’ve got the measurement data reciprocal right there imported already, ie. I’m looking at my holy grail target function (speaker amplitude & phase reciprocal) right in front of me on the screen in rePhase (once I’ve added some sensible upper / lower freq bandwidth limits, of course - I’m not trying to correct down to 1Hz, obviously!!!) why can’t I just automatically generate an FIR filter in rePhase that has that target function? Can rePhase calculate and derive the required FIR coefficients directly from the measurement data? This would be equivalent to drawing any curve shape on the screen, as such, – by importing .txt file of whatever, and saying “Okay rePhase, give me that thing right there on the screen! Make a FIR filter that does that!! And it can just go through maximal iterations and voila! Is that possible? Or does it need to know the mathematical formula definition and parameters for every specific EQ band boost and cut, and every filter slope formula to be able to calculate and define the final resulting FIR coefficients? Which is why it can only generate FIRs from its own known filter types and settings….?
Could it perhaps re-convert HOLM’s imported measurement 65536 point FFT data into the 6144 taps equivalent nearest approximation by numerical value, and then output that series as the target FIR filter. …or something?

3.) “Filter Tab Section Bypass” buttons
Filters Linearization – ACTIVE / BYPASS button
Linear Phase Filters – ACTIVE / BYPASS button
Minimum Phase Filters – ACTIVE / BYPASS button
Paragraphic Phase EQ – ACTIVE / BYPASS button
Paragraphic Gain EQ – ACTIVE / BYPASS button

This way you can quickly isolate what you’re working on without seeing all the other stuff you’ve already tweaked or individually switching bands off and forgetting where you had them! If you fiddle with EQ you can view comparison to see if it was better before / after EQ, with the crossover corrections in place, etc. Just handy for convenience…

4.) Paragraphic EQ Gain range
Please include +/-15 dB (same as Behringer DCX2496) or maybe +/-18dB as these just seem to be more comfortable ranges to work with, comparable with most hardware. I know you can still get the values from +/-24dB range, but when I’m peering closely at the graph screen and move slider up to max, I want it to stop on +15dB if that’s the limit I’m used to working from. Maybe it’s just me…?

5.) Phase graph a different colour to amplitude graph
They’re both dark blue, it gets a bit messy and confusing. How about red for phase curve?

6.) Screen capture for graph button, quickly save jpeg output.

Thank you for your brilliant software. Love it.
Richard
The administrator has disabled public write access.

Feature requests and suggestions 4 years 9 months ago #21899

  • mr-marlen
  • mr-marlen's Avatar
  • Offline
  • Senior Boarder
  • Posts: 50
  • Thank you received: 7
  • Karma: 2
I think it will be very good if we can export Freq+Phase of resulting curve after applying FIR filter (Red curve in Rephase).
The administrator has disabled public write access.

Feature requests and suggestions 4 years 9 months ago #21964

  • mr-marlen
  • mr-marlen's Avatar
  • Offline
  • Senior Boarder
  • Posts: 50
  • Thank you received: 7
  • Karma: 2
It will be also very good if You will add to last version +-360 grad ranges for phase.
Last Edit: 4 years 9 months ago by mr-marlen.
The administrator has disabled public write access.

Feature requests and suggestions 4 years 9 months ago #21984

  • pos
  • pos's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 350
  • Thank you received: 196
  • Karma: 39
Hi Richard
Richard wrote:
1.) "Inverse total amplitude & phase shape” reciprocal button
In one quick button click, swap all the “minus –dB” data points into identical “plus +dB” values (and vice versa +dB becomes -dB) for the amplitude curve, and likewise, swap all “negative degree” phase values to “positive degree” phase values (and vice versa.)
This way the entire transfer function can be made into its own reciprocal correction filter in one click! (Or two clicks if “Invert Amplitude” and “Invert Phase” were two independent button functions – maybe also handy!)
Obviously that’s totally different from basic inverting the signal polarity and putting 180 degree phase shift at all frequencies – which I know already exists in rePhase.

I can perform this job already in HOLM using “Manipulation > Inversion (B=1/A)” which creates the reciprocal function in a new trace, and I could just export that trace as another .txt and import directly into rePhase. But it would be great to have the reciprocal built-in as a quick function in rePhase, handy for comparison view back and forth while you’re working in rePhase, without having to keep re-importing another measurement or having running several rePhase windows concurrently.

You mean inversion of the correction, not the imported measurement, right?
That would be easy enough to do. What would be your use case?
2.) “Generate FIR automatically from measurement data” button
I seem to import my speaker measurements into rePhase, then spend hours trying to tweak dozens of paragraphic EQ filters and phase warps to bend the shape back to flat again… If I “Bypass” the measurement data, I can see my filter shape is getting closer to the reciprocal of the speaker curve (in another window, or HOLM) but it’s still only a bendy approximation. Not perfect match, even after hours of tweaking and fiddling! If I’ve got the measurement data reciprocal right there imported already, ie. I’m looking at my holy grail target function (speaker amplitude & phase reciprocal) right in front of me on the screen in rePhase (once I’ve added some sensible upper / lower freq bandwidth limits, of course - I’m not trying to correct down to 1Hz, obviously!!!) why can’t I just automatically generate an FIR filter in rePhase that has that target function? Can rePhase calculate and derive the required FIR coefficients directly from the measurement data? This would be equivalent to drawing any curve shape on the screen, as such, – by importing .txt file of whatever, and saying “Okay rePhase, give me that thing right there on the screen! Make a FIR filter that does that!! And it can just go through maximal iterations and voila! Is that possible? Or does it need to know the mathematical formula definition and parameters for every specific EQ band boost and cut, and every filter slope formula to be able to calculate and define the final resulting FIR coefficients? Which is why it can only generate FIRs from its own known filter types and settings….?
Could it perhaps re-convert HOLM’s imported measurement 65536 point FFT data into the 6144 taps equivalent nearest approximation by numerical value, and then output that series as the target FIR filter. …or something?

This is unlikely to happen as I really want to keep rePhase a manual tool.
There are many gotcha when doing impulse inversion, and some dedicated tools exists such as Align2 (DRC-FIR) that will take all the needed precautions to make sure everything is correctly made...

That said I did consider adding an "automated EQ" button in the paragraphic tab, something similar to what REW does, but this is quite different to an impulse inversion, and would let the user modify the settings afterward.
3.) “Filter Tab Section Bypass” buttons
Filters Linearization – ACTIVE / BYPASS button
Linear Phase Filters – ACTIVE / BYPASS button
Minimum Phase Filters – ACTIVE / BYPASS button
Paragraphic Phase EQ – ACTIVE / BYPASS button
Paragraphic Gain EQ – ACTIVE / BYPASS button

This way you can quickly isolate what you’re working on without seeing all the other stuff you’ve already tweaked or individually switching bands off and forgetting where you had them! If you fiddle with EQ you can view comparison to see if it was better before / after EQ, with the crossover corrections in place, etc. Just handy for convenience…
Yes , that one is on the todo list (and has been for a long time).
I plan to add a bypass button on each tab, and possibly each EQ bank.
I would alos like to add a "solo" button that would only show the current tab/bank.
4.) Paragraphic EQ Gain range
Please include +/-15 dB (same as Behringer DCX2496) or maybe +/-18dB as these just seem to be more comfortable ranges to work with, comparable with most hardware. I know you can still get the values from +/-24dB range, but when I’m peering closely at the graph screen and move slider up to max, I want it to stop on +15dB if that’s the limit I’m used to working from. Maybe it’s just me…?
OK
5.) Phase graph a different colour to amplitude graph
They’re both dark blue, it gets a bit messy and confusing. How about red for phase curve?
Red is reserved for the result curves.
What about a different shade of blue?
6.) Screen capture for graph button, quickly save jpeg output.
I will see what can be done!
I think I already looked into that in the past, but I am not sure the GUI library I am using will let me do such a thing.

Thank you for all those good suggestions!
https//wavetracing.com | rephase.org
The administrator has disabled public write access.
The following user(s) said Thank You: Richard

Feature requests and suggestions 4 years 9 months ago #21985

  • pos
  • pos's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 350
  • Thank you received: 196
  • Karma: 39
mr-marlen wrote:
I think it will be very good if we can export Freq+Phase of resulting curve after applying FIR filter (Red curve in Rephase).
You mean blue curves with an optional imported measurement? That would be an interesting feature indeed.
I can see a use for exporting the blue (target) curves, but not the red (result) ones?
What is your use case?
mr-marlen wrote:
It will be also very good if You will add to last version +-360 grad ranges for phase.
I removed these ranges because they tended to promote a bad way of using phase EQs.
You should only need slight phase EQ, and rely on the filter linearization tab for the heavy work.
Phase EQ is only there to polish the end result (and then it should not even be needed at all when doing an all active crossover correction)
https//wavetracing.com | rephase.org
The administrator has disabled public write access.

Feature requests and suggestions 4 years 9 months ago #22019

  • mr-marlen
  • mr-marlen's Avatar
  • Offline
  • Senior Boarder
  • Posts: 50
  • Thank you received: 7
  • Karma: 2
pos wrote:
mr-marlen wrote:
I think it will be very good if we can export Freq+Phase of resulting curve after applying FIR filter (Red curve in Rephase).
You mean blue curves with an optional imported measurement? That would be an interesting feature indeed.
I can see a use for exporting the blue (target) curves, but not the red (result) ones?
What is your use case?
mr-marlen wrote:
It will be also very good if You will add to last version +-360 grad ranges for phase.
I removed these ranges because they tended to promote a bad way of using phase EQs.
You should only need slight phase EQ, and rely on the filter linearization tab for the heavy work.
Phase EQ is only there to polish the end result (and then it should not even be needed at all when doing an all active crossover correction)
My use case too see what we got after accepting FIR filters to drivers.
My simulation consits of the next steps:
1) Measuring all drivers in cabinet from 0,6-1m on tweeter axis, using 2 channel MLS measure with real time delay for each driver (reference channel from additional DAC output);
2) Exporting Freq+Phase form same starting cursor.
3) Recalculating Freq+Phase of all drivers from 0,6m of measure to listening point (1,8-2,5m) considering geometry of the cabinet (dy; dx; dz);
4) Exporting recalculated Freq+Phase for FIR simulation.
For now i am using some soft to multiple Freq+Phase recovered from FIR impulse response (blue curve) with Freq+phase of driver.
Also it will be very good option if there will be additional window with option for adding additional group delay amount to exporting curve (i need it when WF FIR is 2048taps; TW FIR is 698; delta=2048/2 - 698/2 = 675 taps = 14,06 ms )
For now i can not get good resaults with my simulation of FIR filter because there is difficulties with delay time in miniSHARC soft. Some pics aded with 256dB/oct LR phase linear filters.
Attachments:
The administrator has disabled public write access.

Feature requests and suggestions 4 years 9 months ago #22023

  • mr-marlen
  • mr-marlen's Avatar
  • Offline
  • Senior Boarder
  • Posts: 50
  • Thank you received: 7
  • Karma: 2
Hmmm maybe i should just invert mid FIR filter impulse...
I have been checked how it will be if i will invert mid polarity (pictures in this post are simulation graphs).
Attachments:
Last Edit: 4 years 9 months ago by mr-marlen.
The administrator has disabled public write access.

Feature requests and suggestions 4 years 9 months ago #22108

  • pos
  • pos's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 350
  • Thank you received: 196
  • Karma: 39
Exporting the target curve (plus optional measurement) should give you what you are looking for, as the result curves are supposed to be close "enough" the the target ones anyway (if that is not the case then you need more taps for your application).
mr-marlen wrote:
Also it will be very good option if there will be additional window with option for adding additional group delay amount to exporting curve
If you have imported a measurement then you can play with the time offset setting.
If you did not import a measurement then you can make up a flat one (two lines, 0Hz and 100kHz, 0 magnitude and phase...) to be able to use this option...
https//wavetracing.com | rephase.org
The administrator has disabled public write access.
The following user(s) said Thank You: devteam

Feature requests and suggestions 4 years 7 months ago #22826

  • akras
  • akras's Avatar
  • Offline
  • Fresh Boarder
  • Posts: 14
  • Thank you received: 1
  • Karma: 0
I couldn't find any versions of rePhase that would run on Mac computers... Do they exist? If not, would it be difficult to make a version of rePhase that could run on OS X? Macs are rather popular, so I think there would be enough interest to justify the effort to make an OS X compatible version.
Thank you-
The administrator has disabled public write access.

Feature requests and suggestions 4 years 7 months ago #22842

  • pos
  • pos's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 350
  • Thank you received: 196
  • Karma: 39
The code could certainly run on macos and linux, but it would require some adaptation and I would need to test it before each release. Unfortunately I don't have a macos machine at hand, nor do I havve the spare time to test and adapt everything.
Maybe that will happen in the future, don't hold your breath though...
In the meantime you can probably use a windows emulator?
https//wavetracing.com | rephase.org
The administrator has disabled public write access.

Feature requests and suggestions 4 years 7 months ago #22875

  • Xcel
  • Xcel's Avatar
  • Offline
  • Fresh Boarder
  • Posts: 11
  • Thank you received: 3
  • Karma: 3
I use VMWare Fusion on my MacBook Pro (OS X Yosemite 10.10.5) to run rePhase and it works just fine.

Greetings

Xcel
The administrator has disabled public write access.
  • Page:
  • 1
  • 2
Moderators: devteam