Announcement

Collapse

Columbus Day Sale

Extended One More Day! Columbus Day Sale at Parts Express! 10/15/19 Click the Parts Express Logo at the top left of this page.
See more
See less

VituixCAD v2 released

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

  • dlr
    replied
    Originally posted by kimmosto View Post
    As far as I know, OmniMic v2 is single channel gear which is not able to measure sound flying time so timing/phase differences between different measurements and drivers are not included in exported response data. That would make life quite hard if target is to simulate multi-way speaker with VituixCAD (or any other software) which calculates off-axis responses in both planes, power and DI responses by polar measurements and simulated crossover. Recommended measurement gear is "normal" analog with two input and output channels and external (or internal) time reference with loop back in the second channel.
    I have to disagree. Measured phase is most definitely not a requirement. Actually, quite the opposite. If measured phase is used exclusively, then determining the absolute acoustic center of each driver would be required to model off-axis responses accurately. But it's not possible to accurately determine absolute acoustic center, which is why relative acoustic center has been used going back at least to the 80's. I first used CALSOD in the mid-90s, initially using the LMS measurement system which cannot measure phase. It had a built-in tailing function for use in generating minimum-phase of the resulting SPL response. However, that alone is insufficient. What is required is the three-measurement scheme to determine relative acoustic offset, as detailed in my SpeakerBuilder article here.

    What this provides is a set of minimum-phase SPL responses that allow for easy determination of acoustic offset. With that offset and those specific (not to be further modified) files any software which can model off-axis response will show accurate off-axis response based on the crossover. The one caveat, a big one of course, is that variations due to diffraction differences in the off-axis are not taken into account, but the direct on-axis response which does have that if taken from measurements in situ of the box will include diffraction at the test point, which is often the listening position or close to it.

    I have and will continue to point out the usefulness of this process for minimum-phase files. Again, it is not necessary to measure phase for basic crossover work. Since absolute acoustic center is, at best, an estimate, using relative acoustic offset provides precise offset values, as long as those SPL files are not further modified. The OmniMic is fully adequate for the task.

    My software, WinPCD, has built in functions to automatically calculate vertical/horizontal responses and for the entire front hemisphere in 5 degree increments and as long as minimum-phase files are used. Measured phase is immaterial and, in fact, not usable for this.

    dlr

    Leave a comment:


  • kimmosto
    replied
    ^Yes, there are few places. Time offset is needed also with dual channel measurements e.g. if all measurements have extra delay and excessive wrapping at high frequencies makes phase responses in graphs fuzzy and reduces accuracy of phase response calculation. Or you have stepped baffle or deep horn, and drivers need to be located physically at correct distance from common reference point on front baffle.

    First option is to adjust physical location of driver instance in the crossover. Positive Z coordinate in mm moves driver further from listener, and negative to closer. X,Y,Z mm simulates physical location of the driver compared to common reference point on the baffle. Common reference point 0,0,0 is typically perpendicular point of listening axis on baffle surface.

    Second option is to adjust Delay [us] of driver's frequency responses. Scale in dB, delay in us and polarity inversion adjustments in Drivers tab are meant for normalizing frequency response data of that driver model. Note that Delay adjustment is not exactly same as location adjustment with X,Y,Z coordinates because delay offset effects to all off-axis directions, but Z-coordinate more to 0 deg and 180 deg directions, and less to 90 deg.
    Last edited by kimmosto; 11-01-2018, 03:37 PM.

    Leave a comment:


  • jhollander
    replied
    I don't think I was clear. If I use single channel measurements or manipulate the measured FRD files for near filed splicing, the measured phase is not accurate. If I extract minimum phase for the FRD files I lose the measured phase. I do know the difference in the phase between the drivers, so is there a location within your program where I can add a delay to the driver, so I can restore (make accurate) the phase alignment?

    Leave a comment:


  • kimmosto
    replied
    ^VituixCAD shows phase response of all individual drivers in speaker simulation. Optimizer function does not have automatic adjustment for phase response so you need to make all adjustments to crossover or mechanics manually. Quite complete set of active IIR and linear phase (FIR) and passive components and blocks is available for controlling phase with filter.
    Is this what you meant?

    Leave a comment:


  • jhollander
    replied
    So, in vituixcad is there a way to align the phase of the drivers?

    Leave a comment:


  • kimmosto
    replied
    Originally posted by DDF View Post
    Absolute phase isnt required, only relative between drivers.
    Sure, and therefore I used " " characters around absolute word. In practice, measured timing is relative to reference channel whatever signal it has, and relative timing for simulation is created with timing offset (pre-delay in ARTA) while calculating phase response for exported frequency responses.

    Originally posted by DDF View Post
    As long as the group delay from mic to acquisition file is constant, two ch isnt necessary.
    This is usually just useless theory, because measurement programs decide (without asking us) how they behave if reference channel is not available or not in use. Most common behavior is that timing is normalized by the program so that each measurement could have different delay between actual acquisition signal and zero time of impulse response. Typical application is that IR peak is automatically moved to 0 ms or to constant number of samples (e.g. ARTA, 300 samples). Nothing in that is absolute or relative so timing of measurements could variate quite randomly preventing summing of different drivers while simulating off-axis.

    So in practice semi-dual or dual channel connection and dual channel (timing reference) mode is usually mandatory, and therefore recommended in this thread.

    Leave a comment:


  • DDF
    replied
    Originally posted by kimmosto View Post
    ^It was just information for you and all other Umik-1, OmniMic v2 etc. users. Those products are better for something else than measuring individual drivers for multi-way design. Capability to measure "absolute" phase/timing directly (without multiple measurements, summing and comparing) is quite basic requirement in order to simulate off-axis with different radiator types. That requires timing reference signal with loop back to second input channel of sound card which records signal from microphone. That is not possible with single channel usb mic which does not generate measurement signal.
    Absolute phase isnt required, only relative between drivers. As long as the group delay from mic to acquisition file is constant, two ch isnt necessary. The real issue with single ch method is the antialiasing errors and errors in source whiteness creep into the measurements and final result but these have no impact on how drivers sum in cad xover packages as these response errors are superimposed on all files equally Sent from my SM-P600 using Tapatalk

    Leave a comment:


  • civit
    replied
    Originally posted by kimmosto View Post
    ^It was just information for you and all other Umik-1, OmniMic v2 etc. users. Those products are better for something else than measuring individual drivers for multi-way design. Capability to measure "absolute" phase/timing directly (without multiple measurements, summing and comparing) is quite basic requirement in order to simulate off-axis with different radiator types. That requires timing reference signal with loop back to second input channel of sound card which records signal from microphone. That is not possible with single channel usb mic which does not generate measurement signal.
    I've learned this the hard way. I use ARTA with the partial-loopback now and it works well.

    Leave a comment:


  • kimmosto
    replied
    ^It was just information for you and all other Umik-1, OmniMic v2 etc. users. Those products are better for something else than measuring individual drivers for multi-way design. Capability to measure "absolute" phase/timing directly (without multiple measurements, summing and comparing) is quite basic requirement in order to simulate off-axis with different radiator types. That requires timing reference signal with loop back to second input channel of sound card which records signal from microphone. That is not possible with single channel usb mic which does not generate measurement signal.
    Last edited by kimmosto; 10-29-2018, 06:00 AM.

    Leave a comment:


  • Thump
    replied
    Originally posted by kimmosto View Post
    As far as I know, OmniMic v2 is single channel gear which is not able to measure sound flying time so timing/phase differences between different measurements and drivers are not included in exported response data. That would make life quite hard if target is to simulate multi-way speaker with VituixCAD (or any other software) which calculates off-axis responses in both planes, power and DI responses by polar measurements and simulated crossover. Recommended measurement gear is "normal" analog with two input and output channels and external (or internal) time reference with loop back in the second channel.
    Not certain if this was to me, although I wound up nearly hi-jacking with my last post (sorry about that) and will move this conversation out of here to another thread - I do NOT want to derail your thread at all. I want it focused on your excellent program, as does everyone else that wanders in here I'm sure!

    In case it was replying to me, to finish up and move out - mine wasn't about OmniMic or using a "mic" at all. This is just about the DATS V2 initialization, and making sure I'm not about to toast a ribbon I can't replace with a DATS T/S, ZMA and FRD sweep.

    Leave a comment:


  • kimmosto
    replied
    As far as I know, OmniMic v2 is single channel gear which is not able to measure sound flying time so timing/phase differences between different measurements and drivers are not included in exported response data. That would make life quite hard if target is to simulate multi-way speaker with VituixCAD (or any other software) which calculates off-axis responses in both planes, power and DI responses by polar measurements and simulated crossover. Recommended measurement gear is "normal" analog with two input and output channels and external (or internal) time reference with loop back in the second channel.

    Leave a comment:


  • Thump
    replied
    Originally posted by mobius View Post
    Thump, it sounds like it's just about time for you to learn to take measurements. In the meantime, would CSS's spec sheet help you out?
    Right I've had that PDF a long time. I located an FRD and ZMA on my NAS but it's been so long I do NOT remember if I wound up personally making that from a trace program, but I believe that's what I did based on the contents of the files. It doesn't have any typical manufacturer notated entries, it's just raw line list data points. So I likely used the graphs from that PDF to do it.

    What I want instead is to use DATS v2 to pull as much information as possible. However research over the last couple days has me partially confused. From what I see, tweeters / ribbons don't typically have T/S parameters used as much as mids and woofers get used, or that seems to be the implication. The ZMA is what's mostly used.

    I can't even confirm if I need to be worried about hooking up the RT2 to DATS V2 without damaging it. I know I had purchased a cap specifically to protect tweeter from having too low a freq sent, thereby damaging it during a sweep.

    However, It's been over 4 years when I was gathering up so much of the stuff I thought I would need (such as a cap for that) I'm not sure where that wound up. It's likely buried in one of the umpteen boxes of stuff I have and it's just a matter of actually finding the thing, if I can. It's a bit of a needle in a haystack at this point for something like that vs the total amount of stuff I've got for the project overall

    -----------------
    Q: So does DATS V2 provide a configurable frequency range I can specify -before- doing a sweep so I can safely extract and interact with the RT2 from DATS without damaging it?
    • This is a very big concern because I can't replace the CSS-RT2 with a new set of CSS-RT2 and I love these things - I don't want to use any other ribbon. These are bad *** high quality wonderful little closed back bricks of true ribbon goodness. I'm partial to them at this point having listened to them for so long, not to mention all the work done to make them precisely fit including flush mount, the latter is true for all drivers in my case.
    -----------------

    Ultimately being able to do fully fledged, extensive measurements is where I want to reach, and I'm literally excited about finally being able to deep dive into measurement aspects. My original goal, which is still the goal, is using the 2nd order H/BP/L custom XO PE designed for me as the base for LCR. I made all of them identical, construction and component wise. I want to do a full, complete, comprehensive set of measurements. Then, use my DSP to actively compensate for "reality" to smooth out the resulting FR into a finely tuned beautiful audio experience for the room.

    I purchased all the tools in preparation for it early on; OmniMic + DATS V2 been sitting in a closet waiting for the day. I simply hadn't (and still quite haven't) reached that stage. I've wanted to, for such a long time, and I'm closer than I've ever been. The reason I'm now starting to poll the forums and search related to using measurement software is because I'm getting close and want to try getting a small head start on some initial knowledge I'll need to begin rudimentary measurements. In this case, just T/S + ZMA and if possible, ideally, FRD just from the ribbon outside the cabinet in free air. This last part is at minimum for personal experience in doing it and to see what it looks like.

    Right now I'm finishing the edges of the MDF baffles with filler + hardener. I'm being overly meticulous because once this is done the paint is high dollar quality oil primer and coat. I want to make certain the surface is truly finished and be as professionally done as I can manage. The results so far regarding edge prep is really nice and winding up looking like I actually put edge banding on it, when in reality it's just repeated thin and super sanded coats of wood filler + hardener for ultra smooth substrate.

    Based on previous experiments, with the above being done to a scrap piece for test, applying final sanding up to 5000 grit and then hitting it with light coats using HVLP the results are super clean, ultra smooth satin from inside the huge pressurized paint booth I made for this project.

    Leave a comment:


  • mobius
    replied
    Thump, it sounds like it's just about time for you to learn to take measurements. In the meantime, would CSS's spec sheet help you out?
    Attached Files

    Leave a comment:


  • kimmosto
    replied
    ^Thanks for your positive feedback!

    Leave a comment:


  • Thump
    replied
    kimmosto Just wanted to say thanks for this outstanding work. Finally look a break from working on a baffle and decided to take a dive into VCAD v2.

    This is just damn impressive. Really, really nice. I'm an infant when it comes to XO. However I'm using a reference design (made for me by PE for my Esoteric LCR) as my template, just to see if I can figure out how to re-create what they did (using X-Over 3 Pro) in your program. Not there yet but it's actually starting to "get there".

    Unfortunately I appear to be one of the few in the world that scooped a set of the incredible CSS-RT2 true ribbon tweeter (3x) before it disappeared. Can't find any FRD/ZMA. If I can find the thread or people that picked up the IP for that Creative Sound Solutions they may have those files somewhere. I also picked up DATS and also have the complete OmniMic kit too, but I haven't even attempted to dive into those things yet.

    So the only reference files I have presently are the original Esoteric mid + woofer series.

    Anyway, Vituixcad v2 seems incredible. Keep up the good work! I'll donate soon as I'm able to. This is as worthy as anything I've ever seen, and I've scoured the planet looking at what's out there. For noobs you've made this thing extremely accessible, flexible and usable because it does have very intuitive aspects that aren't present in several others I've looked at.

    Job well done man. Very.

    Leave a comment:

Working...
X