Announcement

Collapse
No announcement yet.

VituixCAD v2 released

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Originally posted by kimmosto View Post
    ^You can load and enable mic cal file also in ARTA, but that does not have any effect if frequency responses are finally exported with VituixCAD Convert IR to FR tool.
    I guess this is where I get stuck.

    I've taken measurements in ARTA (via the semi-dual method), and had loaded and enabled the mic calibration file for all my raw driver measurements. I then saved all files in ARTA's native *.pir format.

    Followed all the directions for VituixCAD's new "Preparing of response measurements" and did all the processing for near/far-field measurements as recommended in the document.

    When I model a crossover, the VituixCAD simulated response DOES NOT match the measured response when measuring with ARTA with the calibration file enabled. The simulated response DOES match measured response if I don't load or enable the mic calibration file in ARTA.

    Is this what you meant when you said: "does not have any effect if frequency responses are finally exported with VituixCAD"?

    See the attached image.

    RED = Simulated Response from VituixCAD
    GREEN = Measured Response from ARTA with Mic Calibration File Enabled
    BLUE = Measured Response from ARTA with Mic Calibration File Disabled.

    As you can see, the RED and GREEN are nearly on top of one another (disregard response below 1.8K).

    Does this mean that the approach detailed in the "Preparation of response measurements" will not work for any mic that requires a calibration file?


    Also - in the previous Preparation document, there was this one additional step where one was supposed to view the tweeter's measured phase and minimum phase, and add the delay necessary that would cause the measured phase to match as close as possible to minimum phase. Then that same delay value would need to be updated for every driver in the the speaker (i.e., mid-woofer, woofer). Is that step no longer necessary?


    Click image for larger version

Name:	Comparison.jpg
Views:	1
Size:	132.4 KB
ID:	1398840

    Comment


    • Originally posted by dkalsi View Post
      The simulated response DOES match measured response if I don't load or enable the mic calibration file in ARTA.
      That's correct. If you wan't to look or compare measured frequency responses with ARTA, you need to load and enable calibration file also in ARTA.

      PS. This kind of comparison procedure is not included in my measurement prep. instruction because simulation is typically more accurate within 20-20kHz than home measurement of complete speaker. No harm to make some sanity checks and verify that basic things such as polarities and levels are equal to simulation.

      Originally posted by dkalsi View Post
      Does this mean that the approach detailed in the "Preparation of response measurements" will not work for any mic that requires a calibration file?
      No. See previous answer.

      Originally posted by dkalsi View Post
      ...there was this one additional step where one was supposed to view the tweeter's measured phase and minimum phase... Is that step no longer necessary?
      Not needed anymore. Clicking of Far 1 button (and possible fine tuning of Reference time in ms) in Convert IR to FR tool makes the same trick ie sets Reference time to peak of tweeter's IR.
      Last edited by kimmosto; 12-30-2018, 11:33 AM.

      Comment


      • Originally posted by kimmosto View Post

        That's correct. If you wan't to look or compare measured frequency responses with ARTA, you need to load and enable calibration file also in ARTA.

        PS. This kind of comparison procedure is not included in my measurement prep. instruction because simulation is typically more accurate within 20-20kHz than home measurement of complete speaker. No harm to make some sanity checks and verify that basic things such as polarities and levels are equal to simulation.
        Apologies Kimmosto - I still don't quite understand.

        You mention above that if I want to look or compare measured frequency response with ARTA, I need to load and enable calibration file with ARTA.

        I'm actually trying to compare VituixCAD simulated frequency response with ARTA's measured response.

        When I load the mic calibration file in ARTA before taking all measurements and then also load the mic calibration file in VituixCAD......... it seems that VituixCAD's simulated response is not capturing the impact of the mic calibration file.

        Am I supposed to load in one application and not the other (i.e., in ARTA but not VituixCAD, or in VituixCAD but not ARTA)?

        When measuring with ARTA, if I DONT load the mic calibration file - the resulting FR theoretically caputures how that specific mic hears the speaker (but not how the speaker actually sounds).
        When measuring with ARTA, and I DO load the mic calibration file - then the resulting FR theoretically captures how the Speaker actually sounds.

        If the ultimate goal of the VituixCAD is to simulate how the speaker is most likely expected to sound - then at some point in its processing chain it needs to capture the connection needed in the mic.

        BTW - when I used VituixCAD v1 - and followed your instructions on processing files - VituixCAD v1 simulation matched near perfectly with actual measured FR. In fact, even the off-axis simulations matched measured response.

        I'm just trying to get to the same level of success with VituixCAD 2, and the new method of processing files. Again, I have no knowledge of speaker design or acoustic principals - but I'm just trying to get the measurements correct before I move to the next step :-)

        Comment


        • Originally posted by dkalsi View Post
          ...it seems that VituixCAD's simulated response is not capturing the impact of the mic calibration file.
          Could you send calibration file to my e-mail? I'd like to check is there some compatibility issue with the file, or is the problem just some trivial misunderstanding.

          One thing we haven't verified yet is how you compare simulation to measurement with ARTA? To be more specific, how you transfer simulated total SPL from VituixCAD to ARTA that you can compare possible differences?
          I would do that kind of tasks with scale function of VituixCAD Calculator. That is absolutely free from effects of cal/mic files and offers neutral zone for comparisons.

          Comment


          • Originally posted by kimmosto View Post

            Could you send calibration file to my e-mail? I'd like to check is there some compatibility issue with the file, or is the problem just some trivial misunderstanding.

            One thing we haven't verified yet is how you compare simulation to measurement with ARTA? To be more specific, how you transfer simulated total SPL from VituixCAD to ARTA that you can compare possible differences?
            I would do that kind of tasks with scale function of VituixCAD Calculator. That is absolutely free from effects of cal/mic files and offers neutral zone for comparisons.
            Sure - could you please PM your e-mail.

            One thing I forgot to mention was that I had my Mic calibrated by CrossSpectrum Labs - and they provided me the calibration file in *.FRD format.

            Both VituixCAD and ARTA require the calibration file in *.mic format (i think VituixCAD also accepts *.cal or *.txt extension). So I opened the *.FRD file in Microsoft Notepad - and did "Save as" a *.MIC file.

            Do you think that can cause an issue?

            In regards to how I'm comparing simulation to measurement, I do the following:

            1) In Vituix - I go to File --> Export --> Frequency Response
            2) Then in ARTA - I measure the speaker (with mic calibration file enabled) from the same distance when I took my initial measurement for crossover design (with the mic at tweeter-level pointed directly at the tweeter since I specified the tweeter as my primary axis by inputting 0-0-0 for the tweeter's X-Y-Z coordinates).
            3) I then gate the ARTA measurement prior to first reflection
            4) I then go the "Analysis"-->"Single Gated Smooted Frequency Response" --> "File"-->"Export"-->"ASCII file"- and save the FR response
            5) I then open both measurements (i.e., the actual measured response via ARTA, and the simulated exported response via VituixCAD) and open the files in Xsim
            - I use Xsim only because I've gotten used to viewing files in it - no other particular reason.

            Then I see if there is any discrepancy between the two measurements. In my case, there is a discrepancy and it is nearly 100% explained by the mic calibration file - because if I perform Step 2 above WITHOUT the mic calibration file loaded into ARTA, then the measured response matches simulated response nearly 100%.




            Comment


            • e-mail is public, found in Help->About of VituixCAD 2, or in the end of index page https://kimmosaunisto.net/

              FRD should be totally okay but I'd like to see it.

              Comment


              • Originally posted by kimmosto View Post
                Not needed anymore. Clicking of Far 1 button (and possible fine tuning of Reference time in ms) in Convert IR to FR tool makes the same trick ie sets Reference time to peak of tweeter's IR.
                That is awesome! Just be clear - is it okay to click the "Far1" button for every subsequent set of fair field measurements (e.g., mid, woofer, etc.) to be processed?

                I noticed in the Preparation document, it only says to click "Far 1 to set default values for 1st far field export (tweeter's)". If after processing the Tweeter's far field response, I next open up the mid-woofer's far field response, is it okay to click the "Far 1" button again to indicate that the loaded measurement represents field measurement?

                Comment


                • Originally posted by dkalsi View Post
                  it okay to click the "Far 1" button again to indicate that the loaded measurement represents field measurement?
                  No it's not because re-clicking of Far 1 button for mid and woofer would remove timing/acoustical center differences between drivers. Whole idea of dual channel measurement method would be lost.

                  Comment


                  • The following was answered by e-mail but I will copy here to inform other user who might have the same puzzle:

                    I think and hope that possible problem has been found. Convert IR to FR tool requires Scale type=Pascal to enable frequency response compensation with mic/cal file. If Scale type=Volt, program assumes that PIR file is not acoustical measurement, and does not use frequency response compensation even Calibration is checked and valid cal/mic file is selected into text box. I will probably change this feature soon because this kind of secret disabling could be confusing for users, and calibration may be needed also with Volt scale.

                    At the moment you need to check Microphone Used On in ARTA's Audio Devices Setup window to specify PIR file as a mic measurement. That should be checked before measurements. This is proper setting for acoustical measurements, though ARTA allows mic measurements also as a "voltage measurement" (because program does not know what device has been connected to input). I have already added screenshot of Audio Devices Setup window to Measurement preparations -document.



                    Calibration files were okay as I expected. You can load also original FRD file to Calibration text box by entering *.frd to Open file dialog.

                    Comment


                    • FYI, the latest built (2.0.9.5, 2018-12-31 14:35 GMT+2) allows frequency response compensation with mic/cal file also for Volt and unitless (wav) signals. User is responsible to uncheck calibration for IR files which does not need FR compensation. Calibration is forced off only for CLIO .mls files with Ohm scale.
                      In addition, *.frd filter added to Open dialog while selecting calibration file.

                      Comment


                      • Originally posted by kimmosto View Post

                        I think and hope that possible problem has been found..
                        BINGO!!!!!! Worked perfectly!!!! Thanks Kimmo - this resolved the issue.

                        Thanks and Happy New Year!

                        Comment


                        • Excellent! Happy New Year for all!

                          Comment


                          • Originally posted by kimmosto View Post
                            Excellent! Happy New Year for all!
                            Happy New Year Kimmo. I have one comment on something which confused me and one very very minor request.

                            I recently had an issue simulating series wired woofers in the diffraction simulator. It seems simulating 2 woofers automatically increases their output by 6db, which is probably what most people want, but caused me some confusion in a recent project. I simulated this project by making a LF tail which was specified as series wired, but the HF portion was 6db too high. I fixed this but it caused some headscratching.

                            Second issue- when I'm designing with Vituixcad I use the optimizer constantly. In previous versions there was a hotkey to bring it up, but I have to use the drop down menu every time. Do you think you could assign a hotkey to it as before?

                            Third, thanks again for this software, it amazes me how much it's been improving.

                            Comment


                            • Originally posted by civit View Post
                              It seems simulating 2 woofers automatically increases their output by 6db...
                              Actually number of drivers does not affect to SPL in diffraction simulation. I suppose the "problem" is that baffle effect response is normalized to full space i.e. very low frequencies are at 0 dB line and high frequencies at +6 dB, regardless of driver count. This is quite common in diffraction simulators, and feature can be found also for example in LspCAD 5 and Edge.

                              Merger tool compensates normalization to full space by decreasing SPL 6 dB if baffle effect response is selected. Merger has also automatic and manual scaling, so automatic -6 dB with baffle effect response would not be mandatory. I'm aware that this could be confusing in some cases if scaling is not self evident, and especially if diffraction response is used somewhere else than Merger tool.

                              Originally posted by civit View Post
                              Second issue- when I'm designing with Vituixcad I use the optimizer constantly. In previous versions there was a hotkey to bring it up, but I have to use the drop down menu every time. Do you think you could assign a hotkey to it as before?
                              Ctrl+T was changed for Tune block..., but I will try to find something else.
                              PS. You can also minimize optimizer to task bar without closing it.

                              Comment


                              • Originally posted by kimmosto View Post

                                Ctrl+T was changed for Tune block..., but I will try to find something else.
                                PS. You can also minimize optimizer to task bar without closing it.
                                Thanks for your comments. I will try to be more aware of the different options in the merger.

                                Oh by the way - I absolutely love the listening window graph.




                                Comment

                                Working...
                                X