3 ms·

https://www.diyaudio.com/community/ https://www.diyaudio.com/community/ is the main online forum. There's a fair bit of discussion over at parts-express.com's f

https://www.diyaudio.com/community/ https://www.diyaudio.com/community/ is the main online forum. There's a fair bit of discussion over at parts-express.com's forums as well. There's a variety of books out there. Sound Reproduction by Dr Toole is a very comprehensive reference to the big picture, but doesn't really get into speaker and crossover design specifically. There's other books you can find for that, but I'm unaware of the most recent good ones as I read most of this stuff decades ago. In that era the Vance Dickenson book was a big one. The main reasons DIY can rival hifi are: 1. Good components are available to DIY builders. Yeah there's some proprietary stuff that's cool you can't get. But there's a ton of very top quality stuff you can get. 2. Retail hifi has insanely restrictive costs. They'll save every penny possible. So simply by buying better capacitors, etc, or using a slightly more complex crossover, or building a bigger box than is fashionable you can gain an edge. 3. The digital room correction systems offered on most home AV receivers today do a lot to fix simple flaws in DIY designs, giving more slack.

JohnBooty 4y ago
For those who want the ultimate TL;DR on Dr. Toole's book, it's something like this: Listeners prefer objectively accurate (ie, the input matches the output) audio reproduction, with some added emphasis on the bass frequencies. Because most of the sound we hear indoors is actually reflected, it is crucial that frequency response of the speakers remains constant as we move off axis. This might seem blindingly obvious, especially the first part, but it's groundbreaking work as far as demystifying audio is concerned. For so many decades, the snake oil salesmen dominated the hifi industry with bogus claims. They wanted us to think that building speakers was like crafting the perfect wine. But it's not. It's engineering. =)
- jasonwatkinspdx 4y ago
  Yeah, I should have elaborated more. What I love about Toole's book is it's basically the whole big picture of audio, completely sourced and grounded in real deal empirical research, but research that also worked to correlate objective measurements to subjective preferences. All those internet arguments over the years about what matters and what doesn't, tubes vs transistors, class A vs class AB amps, you name it... Toole's book is the cliff notes of what's actually real. It's such a huge advantage vs when I first started reading this stuff in the 90s and it was a lot harder to find information that wasn't just someone's "golden ear" opinion imagined out of thin air.
- archi42 4y ago
  The alternative (or an engineering trade off!) can be doing the opposite: Only care for on-axis sound, and while at it put most of the energy on the axis to reduce energy of reflections. Drawback: Works best with one or at most two listeners, plus speakers and listener are constrained in their relative positions. Bad for the living room, not an issue in a study. The bundling is a matter of membrane diameter and frequency. Bigger diameter result in stronger beam forming, however lower frequencies have less beam forming. You can see the effect very well in the directivity pattern of this 8" wide band speaker: https://www.visaton.de/sites/default/files/dd_product/B%20200.pdf https://www.visaton.de/sites/default/files/dd_product/B%2020... I've a FAST/WAW with a 4" wideband (250Hz XO). As personal monitor speakers these work great with an amazing stage and "resolution" (the illusion of being able to pinpoint the exact spatial location of a single instrument). //woah: Thanks you two (you&GP) for triggering me on this (not /s)! While grabbing a coffee I spent a few minutes thinking about beam forming in audio applications; and came up with the line array. Not so novel, I know, but I'm happy to have gotten there on my own.
  - jasonwatkinspdx 4y ago
    Late reply so I dunno if you'll notice this, but I think you'd enjoy reading about Keele's Constant Beamwidth Transducer line array design. He passed away some years ago, but archives of his website are around, and his AES papers are still there. The TL;DR is he adapted ideas from sonar beam forming to audio reproduction.
archi42 4y ago
I honestly believe crossover design is overrated these days (for DIY). It was a necessity in the past, but today we have cheap amps and great DSPs. "The age of the active speaker" I'd call it. The smallest ADAU145x can do the crossover (and corrections) for a 2-way speaker. If necessary even as a FIR. Products I have are the freeDSP-aurora (8 channel DSP, 300€) and a Beocreate (4 channel DSP+amp, 170€). Both use the ADAU145x family, but the aurora is more powerful. Of course this only applies to DIY. Though I wonder what's the cheapest commercial speaker one can get with an ADAU145x in a crossover-less design.
- jasonwatkinspdx 4y ago
  Yeah, I didn't wanna just blather at people but I should have mentioned that with minidsp and all its descendants you can treat the crossover design as a software problem, which definitely simplifies. But also what I was getting at with mentioning the common room correction systems is these days, you can do something like take a textbook Linkwitz Riley 4th order crossover, do the little bit that's necessary to match levels, and then just call it good and let the receiver's DSP handle the large scale FR response. That advice might annoy some more knowledgable people, but the blunt truth is it'll work and done with reasonable care you'll have a truly impressive system for the price. I don't know pro gear that much but my understanding is there's a bunch of vendors now that package DSP + ClassD amps in a way the speaker companies can white label as their own. Kinda like half the AV receivers on the market are ultimately mildly tweaked Audyssey designs.

JohnBooty 4y ago

For those who want the ultimate TL;DR on Dr. Toole's book, it's something like this: Listeners prefer objectively accurate (ie, the input matches the output) audio reproduction, with some added emphasis on the bass frequencies. Because most of the sound we hear indoors is actually reflected, it is crucial that frequency response of the speakers remains constant as we move off axis. This might seem blindingly obvious, especially the first part, but it's groundbreaking work as far as demystifying audio is concerned. For so many decades, the snake oil salesmen dominated the hifi industry with bogus claims. They wanted us to think that building speakers was like crafting the perfect wine. But it's not. It's engineering. =)

jasonwatkinspdx 4y ago
Yeah, I should have elaborated more. What I love about Toole's book is it's basically the whole big picture of audio, completely sourced and grounded in real deal empirical research, but research that also worked to correlate objective measurements to subjective preferences. All those internet arguments over the years about what matters and what doesn't, tubes vs transistors, class A vs class AB amps, you name it... Toole's book is the cliff notes of what's actually real. It's such a huge advantage vs when I first started reading this stuff in the 90s and it was a lot harder to find information that wasn't just someone's "golden ear" opinion imagined out of thin air.
archi42 4y ago
The alternative (or an engineering trade off!) can be doing the opposite: Only care for on-axis sound, and while at it put most of the energy on the axis to reduce energy of reflections. Drawback: Works best with one or at most two listeners, plus speakers and listener are constrained in their relative positions. Bad for the living room, not an issue in a study. The bundling is a matter of membrane diameter and frequency. Bigger diameter result in stronger beam forming, however lower frequencies have less beam forming. You can see the effect very well in the directivity pattern of this 8" wide band speaker: https://www.visaton.de/sites/default/files/dd_product/B%20200.pdf https://www.visaton.de/sites/default/files/dd_product/B%2020... I've a FAST/WAW with a 4" wideband (250Hz XO). As personal monitor speakers these work great with an amazing stage and "resolution" (the illusion of being able to pinpoint the exact spatial location of a single instrument). //woah: Thanks you two (you&GP) for triggering me on this (not /s)! While grabbing a coffee I spent a few minutes thinking about beam forming in audio applications; and came up with the line array. Not so novel, I know, but I'm happy to have gotten there on my own.
- jasonwatkinspdx 4y ago
  Late reply so I dunno if you'll notice this, but I think you'd enjoy reading about Keele's Constant Beamwidth Transducer line array design. He passed away some years ago, but archives of his website are around, and his AES papers are still there. The TL;DR is he adapted ideas from sonar beam forming to audio reproduction.

archi42 4y ago

I honestly believe crossover design is overrated these days (for DIY). It was a necessity in the past, but today we have cheap amps and great DSPs. "The age of the active speaker" I'd call it. The smallest ADAU145x can do the crossover (and corrections) for a 2-way speaker. If necessary even as a FIR. Products I have are the freeDSP-aurora (8 channel DSP, 300€) and a Beocreate (4 channel DSP+amp, 170€). Both use the ADAU145x family, but the aurora is more powerful. Of course this only applies to DIY. Though I wonder what's the cheapest commercial speaker one can get with an ADAU145x in a crossover-less design.

jasonwatkinspdx 4y ago
Yeah, I didn't wanna just blather at people but I should have mentioned that with minidsp and all its descendants you can treat the crossover design as a software problem, which definitely simplifies. But also what I was getting at with mentioning the common room correction systems is these days, you can do something like take a textbook Linkwitz Riley 4th order crossover, do the little bit that's necessary to match levels, and then just call it good and let the receiver's DSP handle the large scale FR response. That advice might annoy some more knowledgable people, but the blunt truth is it'll work and done with reasonable care you'll have a truly impressive system for the price. I don't know pro gear that much but my understanding is there's a bunch of vendors now that package DSP + ClassD amps in a way the speaker companies can white label as their own. Kinda like half the AV receivers on the market are ultimately mildly tweaked Audyssey designs.