Interesting, but outside 100-10khz is where practiced ears cannot calibrate sound pressure levels so require help most. First-order XO regions can go ouside that range. Cab LF tuning and HF directivity measurements?
I have a dumb question. If the measurement mic is omnidirectional and calibrated, how does the method take into account the ears are not omnidirectional, but much most sensitive front-left/right, not straight front nor straight to one side. This phenomenon is of course critical to stereo sound perception. Thanks.
The crossover will still be correct even if the microphone varies all over the spectrum. The errors will be the same for all the measurements on all the drivers, so all that happens is your global voicing will be the opposite of the microphone response but the relative crossover contributions won't be affected.
Polar plots (directivity information) will also still be intact.
All these matters form a complex part of speaker design in general. The microphone shouldn't get involved in this process, and it doesn't. The fact that it is omnidirectional is irrelevant if sound isn't coming from anywhere but the speaker itself during a measurement, which it shouldn't for crossover type measurements.
OK. I tune my diy by ear using a tone generator/comparator, making sure the driver points axially at my ear but from an angle front-left/right so perception is most sensitive (and mimicks actual listening). Others do the same?
I experimented with crossfeed headphone stereo perception and got the musicians/stage to be in front of me, only by twisting the headphone pads "open-wing" as if front-left/right (then EQ down trebble/bass and compensate for any non-seal loss).
I experimented with crossfeed headphone stereo perception and got the musicians/stage to be in front of me, only by twisting the headphone pads "open-wing" as if front-left/right (then EQ down trebble/bass and compensate for any non-seal loss).
I initiated a thread over in amplifier -> headphones systems https://www.diyaudio.com/community/threads/open-wing-headphone-crossfeed-stereo-sound.391630/
More than certainly. With Rephase i have some really old cloth surround +/-1.5dB and on phase speakers from 100-20k. Take a look into Elliot sound State Variable crossover 😉 .
U can get even flatter response, yet i like those 3dB for tweaking the sound. U ll be amazed at how much 1dB (low q - big bell) changes everything when the system is sort of flat than what one is usually used to.
My 5 cents are, build a proto box, perfection it. When u got tired, make GOOD measurements from it. Bewhare of signal incidence if the mic is random incidence. Dont be scared of putting the mic really close to the floor (u get one less reflection at least in practical terms).
But mainly, do MANY measurments changing mic, source, even room if possible (just try to keep mic distance, and angle (no so much if using random incidence mic). So u end up "nullifying" the things u did wrong, or the constraints that comes with the real world. There was a node from a standing wave at the mic or speaker position, no prob! when u vector average (with rew for example or Smaart) all the measurments, all those errors will influence less the final measurment. Same with any problem any measurment may have. Once u got bored, load that into rephase and let the fun part begin!
Best of lucks! try not to overthink too much, just make errors! big ones! and learn from them, and have fun ! expensive stuff aint really needed (says someone who never had, and prob never will, have expensive gear, at least i diy it myself).
Nothing is funnier (actually its sad) than someone trying to get top dollar stuff, or 0.001% distortion something, when there is a speaker at the end of the chain (call it the distorter-oven), have 40dB noise floor at listening position, where also reverberant field is signifficant, and there are some awfull polyurethane absorbers somewhere because accoustics is improved by the looks, like everything else in this virtual world!
In a nutshell, its better to fail gloriously than to think about every variable involved. Luckily, we have friendly UI dsp nowdays that make everything awfully simple. A plus u get from this sort of system is that u can listen really close to the speakers, making reflections less dominant, the acoustical axis is way more forgiving, and we are really deaf below 200hz when it comes to source localization, so u can place the sub at the corner of the room to excite as much as maximun pressure frequencies (at those freqs the response of the room is modal eitherway) possible, while also effectively using all the power at 1/4 of the sphere (lows are omni, or dipole) . Also, i dont think phase is of concern sub 150-200Hz...
U can get even flatter response, yet i like those 3dB for tweaking the sound. U ll be amazed at how much 1dB (low q - big bell) changes everything when the system is sort of flat than what one is usually used to.
My 5 cents are, build a proto box, perfection it. When u got tired, make GOOD measurements from it. Bewhare of signal incidence if the mic is random incidence. Dont be scared of putting the mic really close to the floor (u get one less reflection at least in practical terms).
But mainly, do MANY measurments changing mic, source, even room if possible (just try to keep mic distance, and angle (no so much if using random incidence mic). So u end up "nullifying" the things u did wrong, or the constraints that comes with the real world. There was a node from a standing wave at the mic or speaker position, no prob! when u vector average (with rew for example or Smaart) all the measurments, all those errors will influence less the final measurment. Same with any problem any measurment may have. Once u got bored, load that into rephase and let the fun part begin!
Best of lucks! try not to overthink too much, just make errors! big ones! and learn from them, and have fun ! expensive stuff aint really needed (says someone who never had, and prob never will, have expensive gear, at least i diy it myself).
Nothing is funnier (actually its sad) than someone trying to get top dollar stuff, or 0.001% distortion something, when there is a speaker at the end of the chain (call it the distorter-oven), have 40dB noise floor at listening position, where also reverberant field is signifficant, and there are some awfull polyurethane absorbers somewhere because accoustics is improved by the looks, like everything else in this virtual world!
In a nutshell, its better to fail gloriously than to think about every variable involved. Luckily, we have friendly UI dsp nowdays that make everything awfully simple. A plus u get from this sort of system is that u can listen really close to the speakers, making reflections less dominant, the acoustical axis is way more forgiving, and we are really deaf below 200hz when it comes to source localization, so u can place the sub at the corner of the room to excite as much as maximun pressure frequencies (at those freqs the response of the room is modal eitherway) possible, while also effectively using all the power at 1/4 of the sphere (lows are omni, or dipole) . Also, i dont think phase is of concern sub 150-200Hz...
No, they sell the calibrated mic directly. Their cal files are for both 0deg and 90deg. $110 plus shipping.
https://cross-spectrum.com/measurement/calibrated_umik.html
https://cross-spectrum.com/measurement/calibrated_umik.html
Hrm,
Is it common practice to build a speaker and instead of just building a crossover, just use a DSP to make it an active crossover temporarily basically until you are happy with the real world results and then build a crossover?
Very best,
Is it common practice to build a speaker and instead of just building a crossover, just use a DSP to make it an active crossover temporarily basically until you are happy with the real world results and then build a crossover?
Very best,
It’s a way to do it so you can listen to different topologies and settings. I used to use DSP crossovers but prefer passive ones now as it allows transportability between different amps and I don’t need 4 Chan of the same amp.
What I do is have a complete set of all common crossover component values and Wago connectors. I wire up the temp crossover in 15 min and listen.
It’s actually faster than setting and tweaking DSP filters if you can believe that. DSP may be easier to tweak but I find that a lot more measurement of the adjustments are needed.
What I do is have a complete set of all common crossover component values and Wago connectors. I wire up the temp crossover in 15 min and listen.
It’s actually faster than setting and tweaking DSP filters if you can believe that. DSP may be easier to tweak but I find that a lot more measurement of the adjustments are needed.
Sorry i was a little sleep deprived yday, and i have that tendency to come here when in such state...
The DSP i use (actually, as most stuff, got that from here), is at software level only. U make ur measurment. Load that into Rephase, export an impulse response, load that into eqAPO for convolution with OS output. Downsides: it makes little thumps when i switch from non sound to sound (i put another song in tidal for example), but maybe its a configuration issue, i dont recall that happening a year ago. Also, I'm scared of passive crossovers and dont know how to properly make them.
I forgot to mention, i use just one stereo amp, and the sub i had lying arround is active. The Home theatre system crossover the signal awfully, so i will make the State Variable crossover, but i forgot to mention, both low outputs will feed a simple op amp adder (maybe its called summer). So i ll end up with L R and L+R for sub. Maybe its far from HI FI but i ll cost me a couple of op amps and not much more.
https://sound-au.com/project148.htm
It would be interesting to make comparisson tests of mono sub signal vs estereo, different sub position configurations, and different phase allingments. For example if i recall correctly, its pretty documented that we love massive ammounts of distortion in the low end. Yes, I'm the overthinker ! 🤐
Best regards!
The DSP i use (actually, as most stuff, got that from here), is at software level only. U make ur measurment. Load that into Rephase, export an impulse response, load that into eqAPO for convolution with OS output. Downsides: it makes little thumps when i switch from non sound to sound (i put another song in tidal for example), but maybe its a configuration issue, i dont recall that happening a year ago. Also, I'm scared of passive crossovers and dont know how to properly make them.
I forgot to mention, i use just one stereo amp, and the sub i had lying arround is active. The Home theatre system crossover the signal awfully, so i will make the State Variable crossover, but i forgot to mention, both low outputs will feed a simple op amp adder (maybe its called summer). So i ll end up with L R and L+R for sub. Maybe its far from HI FI but i ll cost me a couple of op amps and not much more.
https://sound-au.com/project148.htm
It would be interesting to make comparisson tests of mono sub signal vs estereo, different sub position configurations, and different phase allingments. For example if i recall correctly, its pretty documented that we love massive ammounts of distortion in the low end. Yes, I'm the overthinker ! 🤐
Best regards!
Hrm,
Well, before I jump on more equipment, I think I need to try and learn on what I have so I can get used to the process. The idea of selecting some drivers, building an enclosure with them, measuring them individually (frequency response and how to do that; impedance and how to do that), making FRD/ZMA data out of the measurements, and how to then approach simple crossover design. Ultimately I would really like to build a similar speaker like the above with an 8 inch quality driver and a 3~4 inch full range driver with a very simple crossover as a full range speaker.
I have: MiniDSP Umik-1 and the Dayton DATS
I'm thinking of picking up some inexpensive drivers ($10 full range or tweeter and $20 bass woofer) to tinker with and torture and learn from. I have plenty of scrap birch to make smaller enclosures for some mini-tower or just large bookshelf class enclosures. I could build it and put cheap drivers in and then just go through the process or learning to do the above. Then with some help and guidance maybe move on to a larger tower build with quality drivers.
Example cheap drivers to tinker with and learn from:
https://www.parts-express.com/GRS-1TD1-8-1-Dome-Tweeter-8-Ohm-292-462
https://www.parts-express.com/GRS-6PR-8-6-1-2-Poly-Cone-Rubber-Surround-Woofer-292-426
$23 per set; $46 per pair
(The above would be a bookshelf tweeter + woofer to make a 2 way crossover and learn to do that)
https://www.parts-express.com/RESTOCKED-Dayton-Audio-RS100-8-4-Reference-Full-Range-Driver
https://www.parts-express.com/GRS-8SW-4-8-Poly-Cone-Subwoofer-4-Ohm-292-480
$45 per set; $90 per pair
(The above would be a mini-tower full range + bass woofer to also make a simple 2 way crossover and learn on this)
Does this sound reasonable? Or should I not bother with those and just get quality drivers (like the Dayton Reference 8" RS225 & 4" RS100)?
https://www.parts-express.com/Dayton-Audio-RS225-4-8-Reference-Woofer-4-Ohm-295-376
https://www.parts-express.com/Dayton-Audio-RS100-4-4-Reference-Full-Range-Driver-4-Ohm-295-378
$125 per set; $250 per pair
(The above would be a tower full range and learn to make a simple crossover)
Very best,
Well, before I jump on more equipment, I think I need to try and learn on what I have so I can get used to the process. The idea of selecting some drivers, building an enclosure with them, measuring them individually (frequency response and how to do that; impedance and how to do that), making FRD/ZMA data out of the measurements, and how to then approach simple crossover design. Ultimately I would really like to build a similar speaker like the above with an 8 inch quality driver and a 3~4 inch full range driver with a very simple crossover as a full range speaker.
I have: MiniDSP Umik-1 and the Dayton DATS
I'm thinking of picking up some inexpensive drivers ($10 full range or tweeter and $20 bass woofer) to tinker with and torture and learn from. I have plenty of scrap birch to make smaller enclosures for some mini-tower or just large bookshelf class enclosures. I could build it and put cheap drivers in and then just go through the process or learning to do the above. Then with some help and guidance maybe move on to a larger tower build with quality drivers.
Example cheap drivers to tinker with and learn from:
https://www.parts-express.com/GRS-1TD1-8-1-Dome-Tweeter-8-Ohm-292-462
https://www.parts-express.com/GRS-6PR-8-6-1-2-Poly-Cone-Rubber-Surround-Woofer-292-426
$23 per set; $46 per pair
(The above would be a bookshelf tweeter + woofer to make a 2 way crossover and learn to do that)
https://www.parts-express.com/RESTOCKED-Dayton-Audio-RS100-8-4-Reference-Full-Range-Driver
https://www.parts-express.com/GRS-8SW-4-8-Poly-Cone-Subwoofer-4-Ohm-292-480
$45 per set; $90 per pair
(The above would be a mini-tower full range + bass woofer to also make a simple 2 way crossover and learn on this)
Does this sound reasonable? Or should I not bother with those and just get quality drivers (like the Dayton Reference 8" RS225 & 4" RS100)?
https://www.parts-express.com/Dayton-Audio-RS225-4-8-Reference-Woofer-4-Ohm-295-376
https://www.parts-express.com/Dayton-Audio-RS100-4-4-Reference-Full-Range-Driver-4-Ohm-295-378
$125 per set; $250 per pair
(The above would be a tower full range and learn to make a simple crossover)
Very best,
Last edited:
And then I'm also still enamored with the idea of just using several smaller drivers to do similar to be able to make a tall skinny cabinet using a bunch of 4" drivers.
https://www.parts-express.com/RESTOCKED-Dayton-Audio-RS100-8-4-Reference-Full-Range-Driver
https://www.parts-express.com/Dayto...eated-Paper-Cone-Midbass-Woofer-8-Ohm-295-416
$85 per set; $170 per pair
(The above would be 4 of those 4" woofers in parallel & series (final 8ohm) and the full range driver with a 2 way crossover)
I've modeled the little 4" array and it looks like it could work ok. Obviously not as well as larger ones, but the idea of a 6~7" face cabinet that is tall (40~48") and not terribly deep (7~8 inches) would make for a nice skinny cabinet that would fit into most spaces. Relying on wall boundaries and room modes to lift that bottom end enough to sound full range-ish.
Ugh... then I think of that and I'm back to just putting a tweeter & mid-woofer on the front of a skinny tower, and putting an 8" sub driver on the side of it. Ends up being better and similar size (skinny and tall, not crazy deep, will fit in with most furniture and rooms without requiring a big rearrangement). This would require a more complex 3 way crossover, far beyond me. So I'd have to really get help to learn this.
Very best,
https://www.parts-express.com/RESTOCKED-Dayton-Audio-RS100-8-4-Reference-Full-Range-Driver
https://www.parts-express.com/Dayto...eated-Paper-Cone-Midbass-Woofer-8-Ohm-295-416
$85 per set; $170 per pair
(The above would be 4 of those 4" woofers in parallel & series (final 8ohm) and the full range driver with a 2 way crossover)
I've modeled the little 4" array and it looks like it could work ok. Obviously not as well as larger ones, but the idea of a 6~7" face cabinet that is tall (40~48") and not terribly deep (7~8 inches) would make for a nice skinny cabinet that would fit into most spaces. Relying on wall boundaries and room modes to lift that bottom end enough to sound full range-ish.
Ugh... then I think of that and I'm back to just putting a tweeter & mid-woofer on the front of a skinny tower, and putting an 8" sub driver on the side of it. Ends up being better and similar size (skinny and tall, not crazy deep, will fit in with most furniture and rooms without requiring a big rearrangement). This would require a more complex 3 way crossover, far beyond me. So I'd have to really get help to learn this.
Very best,
Last edited:
The drivers you pick are fine. Budget priced doesn’t mean budget sound. Look for ones that are wide bandwidth and have less breakup.
Great candidates are TC9FD for a mid.
If you want to add a tweeter - do it above 5.5kHz and retain advantages of full range driver.
A good dome tweeter like RST28F is great and easy to integrate since well behaved.
The DC28F is also very good.
AMT tweeters are also good generally.
For woofer the 6.5in poly cones are nice. The paper cone Dayton Classic series are also. Good.
Faital Pro 5FE100 and 6FE10 are also excellent woofer to start.
Great candidates are TC9FD for a mid.
If you want to add a tweeter - do it above 5.5kHz and retain advantages of full range driver.
A good dome tweeter like RST28F is great and easy to integrate since well behaved.
The DC28F is also very good.
AMT tweeters are also good generally.
For woofer the 6.5in poly cones are nice. The paper cone Dayton Classic series are also. Good.
Faital Pro 5FE100 and 6FE10 are also excellent woofer to start.
Thanks,
I think my main hold up right now is figuring what direction to go as a learning model. I could get some inexpensive but ok drivers and just work on them a while and learn to measure them from an enclosure and build the metrics to then design crossovers. That's my main goal is to learn to do all that. Part of me wants to just get the cheapest stuff and fiddle with it until I feel good about bothering with more expensive things to make an honest go of it. Then again, it's not like the drivers go to waste, so I wonder about just getting some better drivers and doing the same thing since I can always reuse them in a better build after I learn a lot more. I have "too many" things I want to try of course. I love full range towers for simplicity that will work on anything with no need for sub integration or electronics to integrate in other speakers, just basic good old stereo but full range down to 30hz in a room. But trying to keep it realistic with size dimensions that will work for most rooms.
I should probably get drivers I ultimately want to actually use, so nothing is left sitting around after a project. I don't mind making new enclosures and transplanting drivers from one project to another to learn from. Ultimately I have to learn to make appropriate crossovers. That's my biggest hurdle at the moment I think moving forward on any of my projects. I'm used to subs with no crossover need. Full range drivers were a step to move into speakers without a crossover, but then I figured, that's no going to teach me anything, I need to learn crossovers.
Just not sure if I want to start with a dedicated tweeter + mid-woofer (like 1" dome + 6.5" woofer) and use that to learn to measure and make 2 way crossovers. Or if I want to just get a 4" full range driver and a larger bass driver like 8" and do the same thing with that. Ultimately I like the idea of a full range + bass driver, it just creates a very wide faced speaker compared to doing a tweeter + smaller woofer on a smaller face and then putting a larger bass driver on the side of the cabinet (and learn a 3 way design or whatever it would take to integrate that properly).
And of course, if the Dayton DATS and MiniDSP Umik-1 are ok for right now to learn on for this purpose.
Very best,
UMIK-1 and DATS are what I use. If you want to have smaller frontal profile, look at slot loaded open baffle. My XSD speaker uses 8 opposed 6.5in woofers (low cost poly one) to make some amazing bass.
More here.
https://www.diyaudio.com/community/threads/the-xsd-speaker.385717/
More here.
https://www.diyaudio.com/community/threads/the-xsd-speaker.385717/