Hi Andy2, I'll list my stuff per your request in next post,
but for now may I ask why you want to simulate active dsp speaker response?
One of the big beauties of dsp imo, is you can so easily make EQ, xover, timing, and levels' changes...... and then simply measure the speaker.
There's no cost other than time...no LRC components $ to get wrong, where sims really pay off.
If you use dual channel FFT measurement, you can literally watch your frequency response and phase traces dial in as you make real time dsp adjustments.
So for me ...it's to heck with sims ... real deal measurements are both easier and more accurate...
Maybe I'm just missing what you're after...
Measurement Software: REW + calibrated mic + adjustable mic stand + USB mic preamp (Behringer FCA610) + manual speaker turntable
DSP & Amp: 3e Audio (1701 based) + 3e Audio 4-channel amp
Crossover Design Software: SigmaStudio, specifically the Auto-EQ control.
DSP & Amp: 3e Audio (1701 based) + 3e Audio 4-channel amp
Crossover Design Software: SigmaStudio, specifically the Auto-EQ control.
REW + calibrated mic, one system with passive 1st order filters with NO DSP. One system with minidsp 2x4, another with dbMark XCA 24, + a small system with ZOUDIO AIO4CH in the making.
Measurement Software: Smaartlive and REW
I've also used ARTA, Systune, Holm, and APL-TDA; but keep coming back to those two.
Dual channel like Smaart or ARTA is definitely the way to go imo....the ability to make real time measurements while tuning is a huge advantage over single channel.
Just became aware of an open source dual channel program
https://opensoundmeter.com/#
Haven't tried it yet...might be a great addition to DIY...
Equipment:
Mics, reference is Isemcon 7150, and a handful of ECM8000s for multiple mic measurements.
I've been to measurement workshops where mics were compared/measured ranging from matched M30's $$$ to the ECM8000 and likes.
For DIY speaker tuning, I don't think mics matter much....I don't even bother with mic cal files anymore.
Although I do set correct mic SPL with a calibrator before measurements.
Soundcards: a bunch. Most used is a UMC404, and it probably has the worst specs of any. Another doesn't matter for making good measurements.
Processor: qsys Core110f 12x12, FIR capable, open architecture design...build whatever processor you want.
These things like Kipman's symetrix, are the bomb imo.
Also 1000% agree with his recommendations for professional speaker management processors (and DACs too).
The value, quality, and functionality of proaudio gear (where gear is a tool and not an object of desire) is sooo much greater than home gear, ime...
Previous processing: miniDSP... openDRC -DA8, multiple DRC_DI's, ice plate-amps. Worked very well. So did PC convolution with JRiver.
DACs: again a bunch, from high to low, cost and specs.
Currently using the DAC's in the q-sys processor, or qsys networked amps. DACs make little difference once a decent quality threshold is crossed ime/imo. ymmv
Amps: Pretty much entirely using various qsc proaudio models.
Older home gear, Adcom and Meitner monoblocks hang around for comparisons sometimes.
Convolver software: FirDesigner and rePhase.
Recently tried using REW as FIR generator....it works. But i think the risk of screwing up is pretty high without prior convolution/tuning/measuring experience.
I think the biggest ingredient in active DSP tuning, is by far and away learning how to make good measurements. Where, and how to make them. How to isolate variables. What the measurements can and can't tell you. What you can and can't fix. And how to...
Thanks for listing your setup.
The reason I asked about simulation is. I'll use a 2way as an example.
1. First I suppose you use REW to make measurement of the tweeter and woofer separately.
2. Export the measurement from REW to miniDSP format.
3. Import the files from REW to miniDSP plug-in.
4. At this point, I would like to simulate the overall speaker output with different DSP filtering. Without simulation, I cannot see how the woofer or tweeter contribute to the overall response. For example, I would like to see how the tweeter responses with different FIR filters.
That is I want to tweak the DSP filter without having to measure each time I made any changes to the FIR filters.
When I work in analog domain, once I got the measurement of the tweeter and woofer, I then use XSim to simulate and the result of the overall response is fairly accurate that I don't have to measure the speaker again.
Last edited:
This is a link from miniDSP
Stereo 2 Way Xover
It shows how to design a 2way. Near the bottom, there is a graph of a tweeter and woofer responses and the overall response (near the bottom of the page). Does anyone know which software tool was used?
Stereo 2 Way Xover
It shows how to design a 2way. Near the bottom, there is a graph of a tweeter and woofer responses and the overall response (near the bottom of the page). Does anyone know which software tool was used?
Pipo X9 Win10 "TV box" running JRMC (JRIver Media Center) ->
MiniDSP USBStreamer (USB to ADAT)
Behringer ADAT8200 to
Behringer MX882 ("preamp") to various Behringer amps.
Worth noting: you might use a DVR with HDMI audio in after the PC and you would effectively have up to 8 channels of active. Downside: system can't do normal multi-channel audio as such. This would be much cheaper than even cheap-*** Behringer 🙂
Correction: I am doing active XO with above. Upon reflection, if you can use a PC as source and you have HDMI out, you could use the HTR as pre- and amp, I guess, and multi-channel would work ok? But with my system, JRMC is the software x-over.
For testing, REW and PC.
MiniDSP USBStreamer (USB to ADAT)
Behringer ADAT8200 to
Behringer MX882 ("preamp") to various Behringer amps.
Worth noting: you might use a DVR with HDMI audio in after the PC and you would effectively have up to 8 channels of active. Downside: system can't do normal multi-channel audio as such. This would be much cheaper than even cheap-*** Behringer 🙂
Correction: I am doing active XO with above. Upon reflection, if you can use a PC as source and you have HDMI out, you could use the HTR as pre- and amp, I guess, and multi-channel would work ok? But with my system, JRMC is the software x-over.
For testing, REW and PC.
Last edited:
No, minidsp software can't do sims. Those are measurements after dsp settings applied. Hypex sw can do sort of sims (zero experience) or at least it can import measurements. REW obviously can create roomeq settings by some kind of simulation, but I have never tried that (but no, nothing like xo settings)
A 2-way is very simple to do with measurements and dsp-settings reiteratively, almost in real time. The cost of this is just your time (family?).
If you want to do simulations and create dsp settings based on that, please study everything about VituixCAD and ARTA But bare in mind that you must encounter thousands of details and options...
More info about VCAD and working with it here VituixCAD
Convolution is something that I know nothing about. Can you eat it, or feed to a dog?
Andy2 - try to find someone who you can visit IRL, all this is very difficult to tell here. Hand in hand guidance is invaluable!
Last edited:
This may be a naïve question, but after taking measurement with REW, you can then export the file to miniDSP format then download the file to miniDSP.
But then what will miniDSP do with that file if miniDSP cannot simulate? Can miniDSP at least show the expected, for example, tweeter output once you establish a FIR filter settings? If you're using miniDSP plug-in software, what does it display? Does it display for example a tweeter curve or something?
For example, if I use XSim, I can change the cap value and visually see how the tweeter curve respond in real time. Is there something similar if you're doing active DSP? I'd like to see how the driver respond without having to measure all the time.
I have not tried those options so I am not sure 🙂
Last edited:
If you're referring to the last graph before the "wrapping up" section, that's just the measured response...from REW no doubt. Follow the 'rephase fir tool link' ...
With regard to your reply about simulation....
1. Yes, measure and tune each driver separately.
2. I don't know what you mean by "Export the measurement from REW to miniDSP format" ....
The only thing that would export from REW to miniDSP that I know of, would be EQs that were auto generated in REW. What is a miniDSP format?
3. which is maybe what I think you meant with 3. EQ's going to miniDSP plugin.....???
4. I'd say why bother with a sim at this point, ...just put some xovers in place and measure...it's faster than simming something till you like how the sim looks,.... then measuring that sim, ....and then re-simming, .....remeasuring...on and on...
Just leave your measurement setup in place till you dial in via actual adjustments in the minidsp...
run a REW sweep after every adjustment...it will help with learning as much as tuning..
(and to make real time tuning even easier, get dual channel FFT...either ARTA or try the open source I linked 😉
Yes just like mark100 says. Just do it step by step. Don't try to integrate everything!
Very simple to do, but difficult to explain.
Very simple to do, but difficult to explain.
+1 to your advice.
What is convolution? My girlfriend says I'm convoluted 😛
And sometimes i think she wants to feed me to the dogs!
Maybe I got convolved somehow .....🙄
But seriously, ime using convolution, or rather a FIR file, is so much easier than dialing in IIR to get equivalent tuning results, I'll never go back to IIR other than when any latency can't be tolerated.
haha Juhazi...same time posts...
A couple more question. In the miniDSP plug-in, are there any graphical display that at least show the filter response?
First now that I understand what REW export to miniDSP format only EQ, so now you have three things: filter for tweeter, filter for woofer, and EQ. Does the miniDSP have the hardware to handle, not only each individual channel, but also the EQ that applies to all channels?
Au contraire !!
It's 1000% the opposite !!!!
(other than overall latency, as previously said)
Last edited:
Yep, of course...that's all the miniDSP graphical interface shows ...the filtering being applied.....measurements are not part of the miniDSP curves shown graphically.
Pls pardon, but i think you're missing things...confusing filtering with measurements...i'd suggest to dig in to the tutorials a bit
An earlier post questioned the "simulate vs measurement" aspect. For a number of reasons, I feel there is wisdom to this.
DSP XO work really well - and does what you expect when everything (EVERYTHING) is ideal. Probably the most often missed aspect here is the phase shift of the drivers at and around the XO point. To this end time aligning drivers, or at the very least characterisation of differential phase of drivers at an around the XO point, which is a detailed manual measurement is essential.
While doing this I find it very useful to work through other XO settings as the added time is minimal.
Oh, equipment:
Measurement:
- Behringer ACM8000
- Dayton EMM6, both with cal files
- custom
- Balanced low noise preamp
Software
- Speaker Workshop - primary
- REW
ADCs
- Custom CS5481 - both balanced and unbalanced versions
- Custom CS5361
- Custom AD1939 (integrated CODEC)
- Custom AK5394
DACs
- Custom CS4398 - multiple versions of this
- Custom CS4396
- Custom CS1939 (CODEC)
DSP
- Analog Devices AD1940
- Analog Devices ADAU1445
- PIC32 Micro running DSP code (yep - it works a treat for 2 way)
- All software and programming outside the sigma studio stuff custom.
There are more variants of the above lot than any sane person would want laying around at home and well everywhere...
DSP XO work really well - and does what you expect when everything (EVERYTHING) is ideal. Probably the most often missed aspect here is the phase shift of the drivers at and around the XO point. To this end time aligning drivers, or at the very least characterisation of differential phase of drivers at an around the XO point, which is a detailed manual measurement is essential.
While doing this I find it very useful to work through other XO settings as the added time is minimal.
Oh, equipment:
Measurement:
- Behringer ACM8000
- Dayton EMM6, both with cal files
- custom
- Balanced low noise preamp
Software
- Speaker Workshop - primary
- REW
ADCs
- Custom CS5481 - both balanced and unbalanced versions
- Custom CS5361
- Custom AD1939 (integrated CODEC)
- Custom AK5394
DACs
- Custom CS4398 - multiple versions of this
- Custom CS4396
- Custom CS1939 (CODEC)
DSP
- Analog Devices AD1940
- Analog Devices ADAU1445
- PIC32 Micro running DSP code (yep - it works a treat for 2 way)
- All software and programming outside the sigma studio stuff custom.
There are more variants of the above lot than any sane person would want laying around at home and well everywhere...
What does this mean? Converting to FIR for convenience can remove information in connection with some of the deeper issues. In some cases you have to ask why doesn't it fit?
Minidsp 2x4 using for MTMs (which needs a 24db slope)
I set it up in a few hours and checked results with HolmImpulse/USB1 Mic
Directions - YouTube University.
Have not gone back into the software for over a year now.
I set it up in a few hours and checked results with HolmImpulse/USB1 Mic
Directions - YouTube University.
Have not gone back into the software for over a year now.
Care to elaborate on that? I guess I've never "converted" anything to FIR...
Remove what exactly...
I'd say let go of what you know of passives and learn what active can do for starters. Crossovers are still crossovers, that part remains the same. Just forget about picking a named crossover from a drop down list as a solution and make the combined driver + crossover hit the acoustical target you desire. But that's the same for both passive and active crossovers.
REW is perfectly suited to do a lot of simulations for you, especially combined together with tools like RePhase etc.
But it is different from XSim, new rules, new possibilities...
Read up on well documented active crossovers, be it IIR or FIR. Let go of the passive way of things. Learn and know what it is you wish to accomplish first.
Some reading like Mitch's several tutorials might shed some light on the subject: Audiolense Digital Loudspeaker and Room Correction Software Walkthrough - CA Academy - Audiophile Style
This doesn't mean you need that software package, but it might give some insight into what's possible. He has even written a book about it to help people get their feet wet.
More links: mitchco - Audiophile Style
Last edited: