Xsim Tutorial Part 2 | New Features Like Active DSP Filters
HTML-код
- Опубликовано: 8 сен 2024
- Bill Waslo has added some features to Xsim that I'm really pumped about, so I wanted to show you what they are and how they function. This should really elevate DIY Speaker projects to a new level. Considering this software is free, audio nuts on any budget have access to top tier software!
Link to software: libinst.com/Xsi...
![Run Away - Tzuyu(TWICE) ツウィ 쯔위 [Music Bank] | KBS WORLD TV 240906](http://i.ytimg.com/vi/KnHNNyAYiu0/mqdefault.jpg)
Thanks Ryan.
A couple of user interface notes: you can use a mouse wheel to change values instead of the little up/down arrows if you want - it's a little easier, like twisting a knob. You can also type numbers right into those value boxes too -- make sure to hit your Enter key after you change something that way.
The FRD/ZMA editing tool is mostly meant for situations where all you have is a manufacturer's data sheet. You can read off and enter as many points as needed to make the shape look right for modeling. This is particularly helpful for ZMA curves, as many driver makers don't give impedance phase data at all! -- this gives you a way to get it via Hilbert Transform. (BTW, the reason the graph goes down to 2Hz and up to some stupid high values is because Hilbert derived phase within the usual 20 to 20kHz unfortunately depends somewhat on magnitude shape and values at frequencies way outside that range as well. A few frequency points placed outside there will let you clean up phase at the extremes.
Lastly, on that baffle off-axis animation, you can also put your cursor on the picture itself, hold down the left mouse key, and drag the box orientation around instead of using the sliders. When I get back to programming (whenever that happens) I hope to add some polar plots so you can see what crossover changes do to off-axis responses.
Great info Bill. I didn't even think about the HBT phase thing and the scale. Makes sense. Also didn't realize you can click and drag the speaker around. Nice! Thanks for the updates to the software. It's really great. I don't feel like I did this video much justice. Very busy and challenging past month. But it's mostly just information for the die hards so hopefully they can watch past the mundane boring I was and get the most out of the software.
Great video, Ryan. The baffle and off axis modeling look very useful. Looking forward to trying it out. Big thanks to Bill Waslo!
Much appreciated - I am getting very stoked about DSP crossover and driver controls - especially with regards to compression drivers.
Thank you. I always wanted to progress in my know he with speaker building and for your time
Very cool. Thanks for the video! Am definitely keen to see an active build!
Glad you are showing xsim. Great tool. Please discuss on your next xsim video about 1st order and how makers like Hue (HUMAN Speakers) can get away with just a cap on the tweeter.
nice, do you trust more the manufacturer data sheets(Vas and other parameters) or the ones measured with the DATS V2 device?
i raul that sort of depends. I actually have mostly great confidence in the manufacturer's TS parameters but I like to measure myself and make a judgement call. TS parameters are quitr fickle. If someone on the internet measures TS and claims the manyfacturer numbers are wrong I am always skeptical. They can very a lot based on drive level and test method.
Aryan Sunil nimkar
Aryan Sunil. Nikar
Aryan Sunil
Seems very interesting, thanks. I use a software crossover/DSP on my PC, but have never before come across Xsim as a design tool. I switch between several, but all use other types and formats as input. So I wonder, what kind of output/format does Xsim use to transfer the active filter parameters? Or does it make convolution files?
Thank you Ryan for your videos. You've turned me into a budding speaker builder. But I wonder how closely the XSim results match your measured results? If there are differences, is there a pattern to any discrepancies? (the frequency response is not what is predicted, rolloff points not the same, etc)
Richard Comerford generally it’s pretty accurate. Once in a while there will be little tiny discrepancies here and there, but not something with a pattern and small enough to ignore.
Thanks for the video Ryan! I posted on the FB DIY project pad page earlier today about how to add in driver placement on the baffle - looks like I just have to download the new version. So will the addition of the frd modification and the baffle placement eliminate the need to modify the basic frd/zma files in response modeler before importing into Xsim?
No, XSim still doesn't model baffle step or baffle edge diffraction (might be later, getting some help with how to go about that for reasonable results). The new does simulate distances between drivers (as well as floor/ceiling/wall reflections) though.
Bill Waslo thanks for the reply Bill!
Hi Ryan, this is a great video i'm a huge fan of your channel! I've been trying to find the acoustic difference of my 2 drivers with no success so far, I watched your video where you took 3 measurements, read jeff bagby's description on how to achieve this, and even found and purchased an online course that discussed the subject, but no matter what I do I can't get the phase to line up (deep null at x-over freq) when designing my crossover. Im at a loss, is there any recommendations or resources you might suggest? Thanks a lot!
Ford Grimm seems like you’ve tapped all the resources. What measurements software are you using?
at least that's how I feel, REW but id rather use the omnimic software i have. i'm just not savvy with it. i have the omnimice v2 system
i think i understand the principal, but in execution something just isn't right.
Ford Grimm I haven’t used omnimic, but know enough about it that it shouldn’t cause an issue.
Forget the Bagby write up for a minute. Something I didnt like about it was that it over complicated things. Put the mic near the tweeter about 0.5 to 1.0 meters away. Ok, 3 easy steps. Measure the tweeter. Measure the woofer. Measure the tweeter and woofer wired in parallel. Nothing should change. Put the first two measurements into Xsim (or what ever software) and overlay that 3rd measurement. Adjust the woofer delay until they overlap. Usually about 1-2”.
Is that what you’re doing?
@@ImpulseAudioSpeakers yes, i'm also deriving phase, it helps when i use "measured" phase
Hello Ryan and thank you so much for these so useful videos. I am new to speakers DIY, I do not have easy access to resources as Parts Express. I am using ARTA to get my measurements (impedance and frequency response-phase). I have been able to open my files using VituixCAD, but I like Xsim more. The problem is I have not been able to open my (“measured”) FRD files on Xsim; I just get a view of them when I use the option derived from FRD phase source, and not so when I pick up as measured. Can you see something I might be doing wrong? Thank you.
scienceofcreativity that’s kind of weird. I’m not much of a software guy and don’t understand bugs. I can only suggest using the Xsim fed file manipulator thing and trying to save a copy of the frd file from Xsim. It’s possibly it’ll accept the file more likely that way. That’s just a wild guess though. Sorry I’m not much help...
@@ImpulseAudioSpeakers Thank you so much for your reply Ryan, I will try that option and see what happens :)
If you add 3/4 bass drivers to a cabinet what actually changes? is is that you get more bass or do you have to all come off the same xover point?
Do you get more detailed refinement. I would like your advice on this please.
Dave Jones sorry Im not sure I understand what it is you're asking. You're wondering what happens if you add 3 or 4 woofers instead of just one like I did? Basically you get more and more output the more woofers you add. You can do it by just adding drivers and loading the correct measurement in the driver, which I covered more in the original tutorial.
Hi Ryan,
I am looking to build a crossover, but i have computer problem. I bought a couple woofers and a pair of tweeter from partsexpress. Can you or anyone willing to help me calculate a crossover at 1.6k crossover? Thank for the advance
Sorry Xi, I’m not able to do that right now. I’m having trouble finding the time to make another video. Maybe someone else will see your comment and offer.
Ryan, I am planning to use a DSP for a bass driver and passive crossovers for a midrange and a tweeter. The DSP has about 2 ms data processing latency. Should I then just add that 2 ms delay to the DSP channel? Or is there some separate data setting in the Xsim software to reflect that latency?
Well, I assume the mid and tweeter will still go through the dsp as the high pass? If that’s the case the latency will affect both equally. If not, you’ll need to add delay anyways to get the phase properly aligned. Hopefully you don’t need 1.5ms and need more like 2.5ms. Then you’ll just need to add 0.5ms. Hope that makes sense.
@@ImpulseAudioSpeakers No, I am actually planning to run via DSP only bass section of a speaker. The reason is that Hi-pass outputs of many budget DSP plate amps like miniDSP PWR-ICE, Dayton Audio SPA250DSP and Hypex Fusion FA-series do not provide suffiicent sound quality. So I would like to make something similar to Troels Gravesen's FUSION-BAD design.
Based on your instructions from this video I just have tried to make such circuit in Xsim with 2 ideal drivers crossed at 200 Hz and got somewhat decent result with +/- 1 dB ripple in a 200-900 Hz range. I have used 2-nd order passive filter for a "midrange" driver and 3 DSP modules: PEQ (+1.7 dB at 105 Hz), LR2 filter with 2 ms delay, and OpAmp with 0 dB gain for a "woofer". Both drivers are connected with the same polarity. Reversed polarity on the midrange gives pretty deep -26 dB "null" at about 200 Hz and a small 2 dB bump at 500 Hz.
@@ImpulseAudioSpeakers The results of xSim simulation can be seen here: forum.vegalab.ru/showthread.php?t=89899&page=3&p=2917074&viewfull=1#post2917074
@@IliyaOsnovikov in that case the phase change from the acoustic slopes will hopefully reach 2ms, and likely will. It’ll probably work out.
Nice updates. To bad that the interface is still not very intuitive
Have you ever used FPGraphTracer to create FRD and ZMA files?
Hexaudio yes but not a big fan of doing it that way. Labour intensive. But can help.
I use this tool. Very quick and easy. apps.automeris.io/wpd/
Spl trace isn't bad either
did we have a discussion before with disagreement of me saying , that active Xovers dont have phase issue , and someone did say no mini DSP shown it, they have , its inherent , all have , im like no , 4th order does not , so im curious if this software can do 4th order and whats the phase like ?
Pavle Pavlovic 4th order has 360 degrees of phase rotation so it ends up in phase. Not sure if we had a conversation about that but I’ve heard people confused about that from time to time.
If you apply a 4th order filter to each driver using this software you will see the phase change. If the drivers start in phase then the XO should go 360 and end up in phase. If they start out of phase like real drivers normally do, then they will end up out of phase and you’ll have to add phase correction (delay) to make it all work out in the end.
Yes we did , you found a phase issue in Mini DSP i told you ether you are doing it wrong or Mini DSP is bad , or whatever was the setup , even linked you an article where it is explained that there is no phase issue in 4th order and above, every filter adds 90 degree phase 4 of the 360 , there is no 360 its 0 again , so you said " your understanding of the article is wrong ", but now you say there is no phase issue just like the article , there you go , active crossovers when done proper dont have phase issue .
Now you say just because Xover is in phase doesn't mean drivers will be" , how that happens ? Crossovers do delay , they do volume as well , but interested to hear how that driver phase issue happens if you care to share .
Pavle Pavlovic I never said a 4th order XO is out of phase in theory. It certainly can. This isn’t really the place to explain such a deep topic but I will leave you with one answer to your question. Drivers can be out of phase based on their acoustic offset and frequency response, which are never even or flat in the real world, so you must fix the phase using the XO.
Are you referring to a comment you made on my active XO video where I blindly plugged in the mini dsp and got bad results?
Yes , m8 , dont let us go on a bad note again , audiophiles always have disagreements if i come across as negative its because Eng is not my main language , its hard for me , tho , im educated in electronics 4 years , and 2 years in Electroacustics ,and i produced music all my life , so a lot of XP to be able to talk about things in that realm . Still l want to know more , hence the question , ( you tested a lot more then i did ) what can introduce Phase issue assuming ideal preamp , amp and signal ? I was thinking you would say Time alignment of drivers . Do you mean that by Acoustic Property's ?. Sure that is why Tweeter has its own enclosure among other things , but i would like to know more about that in short . How is it that different drivers can reproduce phase mismatched output , and why , IF you have some links , videos ,pdf , anything , or you can provide some insight and if you like to talk about these things , i do , obviously . We can use mail as well , dedskin gmail . I work as audio producer for a long long time , 10 years or so ,
Yes , im referring to that "MINI DSP Argument " , but we sorted that , no phase problems if Xover is good , i think that Mini DSP is good only for measurements , as a crossover obviously it sucks , but idont have it i dont know , trusted you there but did my own measurments and found a problem .
Using Smaart 7 as reference Pink Noise VS my own SPectrum Analyzer of choice ,and that is NUGEN Audio Visualizer 2 , played a Pink Tone on my Drivers , and recorded with Studio microphone , and i got totally different results from SMAART and NUGEN , on same microphone , and test tone played by SMAART , Pink Noise , im sure you know what im talking about . How can one be sure of those measurements . When i Corrected my Drivers and get a flat response , it sounded horrid , totally flat response but my ear tells me something is wrong . I need to get Proper Mic for that , but that wont fix the difference in 2 Audio Analyzers , NUGEN AND SMAART . Did you have such issues .Im sure you have your own Analyzers , I used a third Analyzer as well , and it was showing same as NUGEN , SMAART gives a lot different spectrum . Didnt measure Phase since i just got Smaart , i spent lot of time jsut of make it measure something .
Ryan....I don't see the download link for the updated version of Xsim.
My bad... I found it...thanks
mark viglucci sorry Ill add it as soon as Im home. There are a couple options here but I cant remember which is the version I used in this video: www.libinst.com/Xsim/
I think it may be the beta version but its basically the latest and meant for public consumption. Bill wont be working on Xsim for a while so he told me to just go for it. I think thats the one.
Hi Ryan, the most recent beta is at libinst.com/Xsim/BetaTest/ . It's still called a beta because the user interface and some operating modes might still change.
How do you create your FRD and ZMA files? You are probably using SoundEasy. However, do you have any other free alternatives? I could use Room EQ Wizard, but the problem is it doesn't have an option to compensate for baffle diffraction, when doing nearfield measurements for the bass drivers. ARTA does this, but again, it's not free software. Got any free solutions to my problem?
Marius Tanasescu I'm not quite sure why you would do a nearfield measurement and want to know the baffle effects. I would continue using REW but back up the mic to 1m and get the measured baffled driver.
When you design crossovers you don't consider the response under 200 Hz? I guess you could do that for a 2 way, but not for a 3 way.
Marius Tanasescu i measure outside and get data down to below 100hz. Sometimes more than 20ft above the ground. Some 3 ways like one Im working on now could get away with data down to only 200hz. Such as crossing to a small mid with an 8" woofer around 800hz. The lower the data the better though.
Keep in my how much the room does dominate the amechoic response below 200hz. I still always use anechoic data as low as possible, but keep it in mind.
Well, that's why I'm doing near-field measurements. To measure the low frequencies indoors. Then I splice it with the far-field response (at 1m). I don't have the means to measure outside with the speaker 20 ft in the air. I'm curious on how you are doing that. You got a crane of some sort?
Marius Tanasescu do you have microsoft excel? I think theres a spreadsheet by Jeff Bagby where you can put your frd file into it and simulate the baffle effects. It is free. But it will add baffle effects to the entire measurement even above 200hz. So you'll want to change the nearfield before you splice the measurement.
I measure from on top of my shop roof. It is a pain in the rear but I only have to do it a couple time during a speaker design. Gets tricky to get the mic that high and make sure it's on axis. But it give an anechoic response that is basically reflection free way down into the deep bass. Even if there are reflections in the measurement they are so low in amplitude they dont really show up.
how to download frd and zma files
If you go to the link on one of my driver test videos you should be able to download them. The link is in the video description.
Hellow Ryan! Where can I get XSim 3d? Thanks!
This link should do it: libinst.com/Xsim/BetaTest/