For those who are using active DSP, could you post what software you are using for measurement? And what equipment such as amp, DAC ...? And if you're using a Convolution software?
For example:
Measurement Software: REW, miniDSP plug-in, ...?
Equipment: miniDSP plate amp, DAC (brand?), ... ?
Convolver software: Equalizer APO, ....?
Am I missing anything?
For example:
Measurement Software: REW, miniDSP plug-in, ...?
Equipment: miniDSP plate amp, DAC (brand?), ... ?
Convolver software: Equalizer APO, ....?
Am I missing anything?
DSP Crossover project
Minidsp Nanodigi dsp and xmos USB boards with spdif out
4 x Khadas Tone Board ES9038QM dac boards
Rotel RMB 1066 6 channel power amp
REW and Minidsp software plus Boxsim 2.0 and Vituix cad for modelling.
Minidsp Nanodigi dsp and xmos USB boards with spdif out
4 x Khadas Tone Board ES9038QM dac boards
Rotel RMB 1066 6 channel power amp
REW and Minidsp software plus Boxsim 2.0 and Vituix cad for modelling.
MiniDSP 8x8
REW
Parasound Zamp x 2
Behringer inuke 1000
Not very interesting I'm afraid!
Are you looking to switch or doing research?
REW
Parasound Zamp x 2
Behringer inuke 1000
Not very interesting I'm afraid!
Are you looking to switch or doing research?
Thank you.
For now I am doing mostly researching, but hopefully my next project will be some form of active DSP.
I am using Hypex Fusion FA253 for the amp and DSP. The Hypex HFD software is used to design the filter.
I use the Dayton Omnimic for measurement, using the Omnimic software. Since this is a USB microphone, it does not interface directly with the HFD software. But this has not turned out to be a big problem... take a measurement with Omnimic, evaluate the response, make an adjustment to HFD, load the filter into the fusion amp, take another measurement with omnimic...
I may try REW, since my understanding is that REW can export an impulse response which can then be imported into HFD.
I am still building cabinets, so I have not gotten to the stage of evaluating filters and EQ. I hooked things up and did just enough to verify my electronics work and the software does what it needs to do.
I use the Dayton Omnimic for measurement, using the Omnimic software. Since this is a USB microphone, it does not interface directly with the HFD software. But this has not turned out to be a big problem... take a measurement with Omnimic, evaluate the response, make an adjustment to HFD, load the filter into the fusion amp, take another measurement with omnimic...
I may try REW, since my understanding is that REW can export an impulse response which can then be imported into HFD.
I am still building cabinets, so I have not gotten to the stage of evaluating filters and EQ. I hooked things up and did just enough to verify my electronics work and the software does what it needs to do.
Also please describe how you simulate the xover works if you're using active DSP.
For example, if you're using miniDSP, I suppose you can use REW to take the measurement of the tweeter and woofer. You then import the REW measurement to miniDSP?
Then can miniDSP simulate the sum output of the tweeter and woofer?
For example, if you're using miniDSP, I suppose you can use REW to take the measurement of the tweeter and woofer. You then import the REW measurement to miniDSP?
Then can miniDSP simulate the sum output of the tweeter and woofer?
Last edited:
I'm using the convolver in Daphille. That only de-characterizes my FR speakers "rising response". I generated the filter using REW, with help from a Apex220 measurement mic.
I'm using a 4 out miniDSP for the Assist woofers. Only the LR4 low-pass at 70 Hz with an additional "reverse rising response" generated via the miniDSP parametric EQ.
The FRs run full range, the miniDSP taps in via the headphone output of the DAC.
I just got the (4X 40W) Zoudio amp, which may replace the DAC, miniDSP and Jolida amp, depending on how it sounds. We'll see.
I'm using a 4 out miniDSP for the Assist woofers. Only the LR4 low-pass at 70 Hz with an additional "reverse rising response" generated via the miniDSP parametric EQ.
The FRs run full range, the miniDSP taps in via the headphone output of the DAC.
I just got the (4X 40W) Zoudio amp, which may replace the DAC, miniDSP and Jolida amp, depending on how it sounds. We'll see.
Raspberry Pi B+ with Diamond 7.1 usb dac, Raspbian running mpd with my own alsa LADSA plugin. For measurement the Audio Analyzer iPhone app using a homemade Linkwitz/Miller modified microphone, wired to a mini TRRS plug for use in an iPhone 6+.
Win10 with equalizer apo and Asus xonar u7mkii usb 7.1 audio interface. I've also used behringer 204hd and minidsp 2x4hd with and without Dirac 1 and 2. 100db plus horn speakers. The xonar u7mkii is the quietest solution I've come across with the horns as long as I use a laptop and not a desktop PC. Soundwise it's better than the minidsp for sure, tough to call vs. the behringer but the behringer had more noise. The u7mkii is dead quiet ear at the mouth on 110db horns with no padding. Try it you. I think you'll like it and the price is right. I might try a pcie Asus 7.1 card next but I think there's a good change it will again be too noisy so I'm a bit shy to go there at this point. Rew for measurement software with a lab calibrated umik. I've played with for a lot but actually prefer iir. I level match for about a 6db drop from bottom to top and no longer try to eq beyond that. I'm 3 way with 500hz and 10000hz 2nd order butterworths. I have a bit of delay on the supertweeters due to their acoustic center being about 12 inches in front of the mids.
I used to be an amp tweaker - active setup has cured me of that. Even slight eq has way more effect than boutique caps etc. Your mileage may vary!
I used to be an amp tweaker - active setup has cured me of that. Even slight eq has way more effect than boutique caps etc. Your mileage may vary!
Last edited:
My three projects (the last one not in use now)
Measurements with REW and UMIK-1 for all! All use analog signal from AV receiver connected to sources for ease of use for the whole family!
Aino gradient - a collaborative speaker project
- Minidsp 4x10HD, two 2-channel ICEpower modules per speaker in separate boxes
SEAS MR18 3-way
- Hypex FA123 per speaker built-in
Avalanche AS1 modernization
- Minidsp 2x4HD per speaker outside and original built-in amp modules (class AB) 3 per speaker built-in
Measurements with REW and UMIK-1 for all! All use analog signal from AV receiver connected to sources for ease of use for the whole family!
Aino gradient - a collaborative speaker project
- Minidsp 4x10HD, two 2-channel ICEpower modules per speaker in separate boxes
SEAS MR18 3-way
- Hypex FA123 per speaker built-in
Avalanche AS1 modernization
- Minidsp 2x4HD per speaker outside and original built-in amp modules (class AB) 3 per speaker built-in
Last edited:
Measurement Software: REW + Behringer ECM8000 mic
Equipment: PC with internal PCIe Asus 7.1 card, Yamaha P2040 amps
Convolver software: Equalizer APO
Equipment: PC with internal PCIe Asus 7.1 card, Yamaha P2040 amps
Convolver software: Equalizer APO
XTZ room analyzer II pro
Sitronik Lucius 6A
Both all-in-one solution, don't like clutter and add on
Sitronik Lucius 6A
Both all-in-one solution, don't like clutter and add on
Regarding sims and how to:
I don't do any IT simulations for my projects! Just vivid imagination and wild guessing by what I've learned about speaker design, directivity, interferences etc. and mostly "given" speaker morphology and drivers.
When drivers are in proto baffle/box, I take measurements with REW, also off-axis and high spl to check directivity and distortion in low end.
Then I go to DSP softaware and manually set eq parameters to flatten on-axis response in intended passband. After that I set xo points and slopes keeping directivity and distortion in my mind. Then I play tweeter and mid and fine-tune that xo (delay and parameters by measurements, also listening to music!). When that is fixed, I play mid and woofer and set up xo manually. These measurements should be done quasi-anechoically, but indoors in a large room is ok. Bass low end tuning and whole system "tonality" is finalized with the speaker in intedend location in the room, with a pair. I use similar setting for L/R and no room EQ except lowest bass below 60Hz.
This do-measure-correct-measure approach is easy and actually rather fast and definitely the most fun part of the project! A laptop PC runs measurement and DSP software simaultaneously in different windows and it takes just a few mouse moves and clics.
If I were not challenged mathematically and had more time and energy, I might start to learn VituixCAD simulation. Zero motivation for it now.
I don't do any IT simulations for my projects! Just vivid imagination and wild guessing by what I've learned about speaker design, directivity, interferences etc. and mostly "given" speaker morphology and drivers.
When drivers are in proto baffle/box, I take measurements with REW, also off-axis and high spl to check directivity and distortion in low end.
Then I go to DSP softaware and manually set eq parameters to flatten on-axis response in intended passband. After that I set xo points and slopes keeping directivity and distortion in my mind. Then I play tweeter and mid and fine-tune that xo (delay and parameters by measurements, also listening to music!). When that is fixed, I play mid and woofer and set up xo manually. These measurements should be done quasi-anechoically, but indoors in a large room is ok. Bass low end tuning and whole system "tonality" is finalized with the speaker in intedend location in the room, with a pair. I use similar setting for L/R and no room EQ except lowest bass below 60Hz.
This do-measure-correct-measure approach is easy and actually rather fast and definitely the most fun part of the project! A laptop PC runs measurement and DSP software simaultaneously in different windows and it takes just a few mouse moves and clics.
If I were not challenged mathematically and had more time and energy, I might start to learn VituixCAD simulation. Zero motivation for it now.
Last edited:
I use a symetrix symnet 8x8 DSP and DIGIO with REW for measurements. I would recommend using a professional loud speaker management system over HiFi gear. The cost 2nd hand can be lower and the performance, level of integration and level of features are far beyond the gear targeted for HiFi.
My list:
A Workstation PC with JRiver as media player (and build-in convolution engine plus 2 banks of seemingly limitless PEQ's available + several plugins) with an Asus Xonar Essence ST PCI 7.1 sound card feeding mains to a Musical Fidelity M1 DAC using S/PDif optical. Extension card feeds ambient and subwoofer channels.
Measurements made with REW, DSP processing with my own custom template(s) within DRC-FIR with a couple of tweaks done within RePhase.
Measurement mic is a Behringer ECM8000 with calibration file. Phantom power from a IMG Stage Line MPA-102.
M1 DAC feeding a JDS Atom (as pré-amp)* which feeds a Goldmund Telos 400 clone dual mono amplifier. For ambient speakers a Fetzilla amplifier is used. 2x Subs fed by NC500MP Hypex amplifiers.
More info including many measurements: The making of: The Two Towers (a 25 driver Full Range line array)
I do convolution sims within REW and 'live' measurements of end results. Documented in the above thread.
* = temporary until a real pré-amp is build.
A Workstation PC with JRiver as media player (and build-in convolution engine plus 2 banks of seemingly limitless PEQ's available + several plugins) with an Asus Xonar Essence ST PCI 7.1 sound card feeding mains to a Musical Fidelity M1 DAC using S/PDif optical. Extension card feeds ambient and subwoofer channels.
Measurements made with REW, DSP processing with my own custom template(s) within DRC-FIR with a couple of tweaks done within RePhase.
Measurement mic is a Behringer ECM8000 with calibration file. Phantom power from a IMG Stage Line MPA-102.
M1 DAC feeding a JDS Atom (as pré-amp)* which feeds a Goldmund Telos 400 clone dual mono amplifier. For ambient speakers a Fetzilla amplifier is used. 2x Subs fed by NC500MP Hypex amplifiers.
More info including many measurements: The making of: The Two Towers (a 25 driver Full Range line array)
I do convolution sims within REW and 'live' measurements of end results. Documented in the above thread.
* = temporary until a real pré-amp is build.
Last edited:
UMM-6 and REW
Aphex 124a (input level adjustment) Electro-Voice DX34a (DSP) 2 in 4out (pair for 3ways or a Merlin ISP100) < all older tech that just works, quietly.
Rane MA6S six channel amp.
Aphex 124a (input level adjustment) Electro-Voice DX34a (DSP) 2 in 4out (pair for 3ways or a Merlin ISP100) < all older tech that just works, quietly.
Rane MA6S six channel amp.
Measurement: UMIK, REW, HOLM Impulse.
Windows 10 PC => USB => Topping D10 DAC => Toslink => DSP, no AD/DA conversion.
DSP & volume control: miniDSP 4x10HD.
No convolution.
FLAC files are streamed from a Raspberry Pi 4 (running moOde) with a HiFiBerry Digi+ hat to the miniDSP by coaxial SPDIF.
The occasional LP uses the analog input of the miniDSP.
Windows 10 PC => USB => Topping D10 DAC => Toslink => DSP, no AD/DA conversion.
DSP & volume control: miniDSP 4x10HD.
No convolution.
FLAC files are streamed from a Raspberry Pi 4 (running moOde) with a HiFiBerry Digi+ hat to the miniDSP by coaxial SPDIF.
The occasional LP uses the analog input of the miniDSP.
Meas: Line Audio OM-1, Edirol Quad-Capture, REW.
Source: (d) Squeezbox Touch -> toslink -> Ian's re-clocker -> toslink -> toslink active splitter (1->3) ->
Main speaker (LF limited FR) no HP (stuffed ports):
(d) toslink-1 to Left (5 meter) -> DAM DAC in 1 ch diff mode, toslink-2 to Right (5 meter) -> DAM DAC in 1 ch diff mode
(c) All_digital_coefs_1.2 (XL) -> Biquads -> Soekris DAM DAC
LF 45 Hz LP:
(d) toslink-3 (L+R) -> Hypex DLCP
(c) Hypex DSP Filter Designer -> DLCP
So no FIR, only IIR actions. Attachment is from listening position in sofa. Yes, a bit forward but not harsh, just resolved, but probably needs taken down a bit at 1k. Measurements where done using pink noise and REW RTA. I execute a slow change process running over weeks between adjustments. I have actually done one change round to this but have't measured it yet - don't want to know yet 🙂
//
Source: (d) Squeezbox Touch -> toslink -> Ian's re-clocker -> toslink -> toslink active splitter (1->3) ->
Main speaker (LF limited FR) no HP (stuffed ports):
(d) toslink-1 to Left (5 meter) -> DAM DAC in 1 ch diff mode, toslink-2 to Right (5 meter) -> DAM DAC in 1 ch diff mode
(c) All_digital_coefs_1.2 (XL) -> Biquads -> Soekris DAM DAC
LF 45 Hz LP:
(d) toslink-3 (L+R) -> Hypex DLCP
(c) Hypex DSP Filter Designer -> DLCP
So no FIR, only IIR actions. Attachment is from listening position in sofa. Yes, a bit forward but not harsh, just resolved, but probably needs taken down a bit at 1k. Measurements where done using pink noise and REW RTA. I execute a slow change process running over weeks between adjustments. I have actually done one change round to this but have't measured it yet - don't want to know yet 🙂
//
Attachments
Last edited:
A quick question, does miniDSP software have the ability to simulate the speaker response? For analog, for example, one can use XSim. But with active DSP, what software do you use to simulate?
Thanks.
Measurement Software: ARTA + calibrated mic + adjustable mic stand + mic preamp (Focusrite Clarett 2PRE)
Crossover "equipment": Linux based computer + LADSPA host (ecasound, Gstreamer, etc.) + LADSPA plugin I wrote (ACDf)
Convolver software: Just (yesterday!) got a new FIR LADSPA plugin working that uses FFTW. Cool beans
I only build active loudspeakers, and I have a bunch of second hand amps a la Adcom, ATI, and others to draw from.
You probably forgot one of the most important types of software: CROSSOVER DESIGN software! I use my own ACD Tools for the most part. VituixCAD seems to be very capable in this regard but I have yet to give it a real try.
Crossover "equipment": Linux based computer + LADSPA host (ecasound, Gstreamer, etc.) + LADSPA plugin I wrote (ACDf)
Convolver software: Just (yesterday!) got a new FIR LADSPA plugin working that uses FFTW. Cool beans
I only build active loudspeakers, and I have a bunch of second hand amps a la Adcom, ATI, and others to draw from.
You probably forgot one of the most important types of software: CROSSOVER DESIGN software! I use my own ACD Tools for the most part. VituixCAD seems to be very capable in this regard but I have yet to give it a real try.