This is my first post to this forum. I'm very impressed with what I've read concerning the level of expert advice and polite helpfulness on these threads.
A month ago, I started to investigate the possibilities of building speaker enclosures to match the video and component cabinets that I custom build. I've been a building contractor and cabinet maker for 32 years. I'm entertained by the projects I see posted, some are very, very nice. Some show me the difference between the knowledge and skills required to build an amp, and those required to build furniture are sometimes extreme.
I'm a complete novice to audio electronics, and the many hours of reading I've done in the last month, has led me to believe the understanding of all the complex theories of acoustics and electronics will never stick to my tiny and aging brain. The questions don't come fast and furious, they come slowly and as overwhelming as a tidal wave. I feel like a sixth grader trying to learn rocket science. But alas, I am learning.
One question I have is whether or not the holy grail of driver/enclosure experimenters is the digital, PC controlled crossover. Is Digital Signal Processing an artform for the wizards and code writers? Or is it figured out, available in a card that can be purchased? My searches on this site have left me with an equal amount of questions and $2000 answers.
My hope is someone can direct me to an appropriate thread or that something will soon be announced.
My other question is one I hope to understand the answer to.
I have a pair of HeathKit speaker cabinets without any numbers to identify them. They appear to be vintage Altec 859A's. inside is a 15" driver, model #401-149 and a horn 400-3 with its driver, 401-150.
To my ears, they sound very smooth in the base, very efficient, (much louder than my Prophiles) but the highest frequencies just aren't there. I've often thought about adding a tweeter.
Recently, I opened the back panel to see what's inside. and found
that someone has disconnected its original xover (components there, but wires disconnected) and replaced it with one from Radioshack.
In trying to find the proper replacement, my searches at both HeathKit and Altec come up empty. Should I try to rebuild the original crossover? Or can someone help me find a better one?
Any help would be most appreciated.
A month ago, I started to investigate the possibilities of building speaker enclosures to match the video and component cabinets that I custom build. I've been a building contractor and cabinet maker for 32 years. I'm entertained by the projects I see posted, some are very, very nice. Some show me the difference between the knowledge and skills required to build an amp, and those required to build furniture are sometimes extreme.
I'm a complete novice to audio electronics, and the many hours of reading I've done in the last month, has led me to believe the understanding of all the complex theories of acoustics and electronics will never stick to my tiny and aging brain. The questions don't come fast and furious, they come slowly and as overwhelming as a tidal wave. I feel like a sixth grader trying to learn rocket science. But alas, I am learning.
One question I have is whether or not the holy grail of driver/enclosure experimenters is the digital, PC controlled crossover. Is Digital Signal Processing an artform for the wizards and code writers? Or is it figured out, available in a card that can be purchased? My searches on this site have left me with an equal amount of questions and $2000 answers.
My hope is someone can direct me to an appropriate thread or that something will soon be announced.
My other question is one I hope to understand the answer to.
I have a pair of HeathKit speaker cabinets without any numbers to identify them. They appear to be vintage Altec 859A's. inside is a 15" driver, model #401-149 and a horn 400-3 with its driver, 401-150.
To my ears, they sound very smooth in the base, very efficient, (much louder than my Prophiles) but the highest frequencies just aren't there. I've often thought about adding a tweeter.
Recently, I opened the back panel to see what's inside. and found
that someone has disconnected its original xover (components there, but wires disconnected) and replaced it with one from Radioshack.
In trying to find the proper replacement, my searches at both HeathKit and Altec come up empty. Should I try to rebuild the original crossover? Or can someone help me find a better one?
Any help would be most appreciated.
I'll answer the bit on digital crossovers seeing as it's what i do for a living.
Whether the holy grail of audio systems involves a digital xover or not is entirely down to personal opinion. There still exists a significant digi-phobe contingent, even within the pro-audio industry. These days, i'd say there's probably not much to choose between a top quality (active) analogue and digital crossover. However, digital or analogue, active crossovers and separate amps for each driver are better than passive crossovers. I'm sure there'll be some disagreement here, but IMHO there's no contest.
As you've found, the commercially available digital crossovers are not cheap. The only two brands really worth their price are the BSS and XTA units. To be honest not a great deal to choose between them - the BSS FDS-366 is currently the only one which works at 96kHz, 24 bit (the rest are 48k, 24-bit) and some users say the user interface on the XTA's are a bit simpler, but they're both very good manufacturers. They both have external PC control software (via 485/MIDI) although work perfectly well without.
As to the DIY approach on it, well there's a few points. First off, the maths is not easy but not impossible. Once you understand the basics of z-planes, poles on filters and all that stuff it's not too bad, although not the sort of thing you can do with a hangover. The big problem is the electronics. This really needs to be in a separate box with it's own PSU, not in a PC with all that noisy stuff in the same box to screw up your noise floor. Also, latency in a PC is unacceptable in a lot of appilcations. Doing the electronics yourself is not really practical - this is difficult stuff, the design is hard enough and testing/troubleshooting this sort of digital equipment is impossible without decent (expensive) test kit.
So, the short answer to your question is that the best option for a digital xover is a BSS or XTA unit. Personally, I use a BSS Minidrive for testing and working out what xover points sound best, gain settings and that sort of thing, then build the filter up as an active analogue filter on the same board as the amps. There's plenty of problems here too, but it's a much easier (and cheaper) approach for DIY.
Give me a shout if you've got any other questions on this. Hope that helps.
Whether the holy grail of audio systems involves a digital xover or not is entirely down to personal opinion. There still exists a significant digi-phobe contingent, even within the pro-audio industry. These days, i'd say there's probably not much to choose between a top quality (active) analogue and digital crossover. However, digital or analogue, active crossovers and separate amps for each driver are better than passive crossovers. I'm sure there'll be some disagreement here, but IMHO there's no contest.
As you've found, the commercially available digital crossovers are not cheap. The only two brands really worth their price are the BSS and XTA units. To be honest not a great deal to choose between them - the BSS FDS-366 is currently the only one which works at 96kHz, 24 bit (the rest are 48k, 24-bit) and some users say the user interface on the XTA's are a bit simpler, but they're both very good manufacturers. They both have external PC control software (via 485/MIDI) although work perfectly well without.
As to the DIY approach on it, well there's a few points. First off, the maths is not easy but not impossible. Once you understand the basics of z-planes, poles on filters and all that stuff it's not too bad, although not the sort of thing you can do with a hangover. The big problem is the electronics. This really needs to be in a separate box with it's own PSU, not in a PC with all that noisy stuff in the same box to screw up your noise floor. Also, latency in a PC is unacceptable in a lot of appilcations. Doing the electronics yourself is not really practical - this is difficult stuff, the design is hard enough and testing/troubleshooting this sort of digital equipment is impossible without decent (expensive) test kit.
So, the short answer to your question is that the best option for a digital xover is a BSS or XTA unit. Personally, I use a BSS Minidrive for testing and working out what xover points sound best, gain settings and that sort of thing, then build the filter up as an active analogue filter on the same board as the amps. There's plenty of problems here too, but it's a much easier (and cheaper) approach for DIY.
Give me a shout if you've got any other questions on this. Hope that helps.
Another option to the BSS is the Sony SRFP300 which is decent enough sounding (MUCH better than a UC8024 Behringer) and is about 1/3 the price of the BSS here in Oz. If my system was digital only, I'd have one permanently installed, but I listen to analog mainly and I don't want a stinking PCM monstrosity in my system thanks.
Digital x-over
I have been interested in this for several years now. I found a development board made by Analog Devices for their Shark DSP that sells for around $200. It has an SPDIF interface on it. I then went to ebay and bought a few Adcom D/A converters (GDA-600?). Now if I could find a few hundred hours to work on this project, I might get something. This route is pretty complex, as you need to write your own C code to implement the filters etc, and of course you need some other software (Matlab) to help design the filters. The development board does comes with a compiler.
There is no doubt in my mind that nothing beats a properly implemented digital crossover filter. It gives you the ultimate flexibility. The nicest solution I've seen comercially is the Meridian speakers, but they are way out of my price range.
I have been interested in this for several years now. I found a development board made by Analog Devices for their Shark DSP that sells for around $200. It has an SPDIF interface on it. I then went to ebay and bought a few Adcom D/A converters (GDA-600?). Now if I could find a few hundred hours to work on this project, I might get something. This route is pretty complex, as you need to write your own C code to implement the filters etc, and of course you need some other software (Matlab) to help design the filters. The development board does comes with a compiler.
There is no doubt in my mind that nothing beats a properly implemented digital crossover filter. It gives you the ultimate flexibility. The nicest solution I've seen comercially is the Meridian speakers, but they are way out of my price range.
The biggest problem with the digital crossovers, particulary with the FIR filters, is the inherent delay of the algorithm(s). I'm building monitor speakers (for music production), and if you have to monitor yourself via speakers you can't use speakers which delay sound by several hundred milliseconds, a 10 ms delay (or latency) is an absolute maximum tolerable delay, 1 ms is OK, I wouldn't go with anything worse. That means that you can delay the signal max. 44 samples @ 44.1 kHz sample rate (for 1 ms latency). For FIR impulse response this means that the IR length can be max. 88 samples at that sample rate. Obviously there are now sound cards and converters that support >=96 kHz sample rates, so practically an 192 sample IR is OK, which is *just* enough for a simple xover (for 3-way systems) but not much for anything else (for an example you should use about 4k sample IR's if you want to 'sample' the speaker sound and invert the frequency response to get a straight FR, that's just one example of the available DSP applications for digital crossovers). At 96 kHz the latency for ~4k IR would then be about 20 ms which is still OK for mixing duties. So propably it's best to implement two parallel xovers with different size IR's, or use analog xovers (24 dB/oct Linkwitz-Riley is still enough for simple monitoring).
Edit: Regarding the latency don't forget the delay caused by the converters itself (something from 16 to 32 samples with current A/D's), it's better to use digital interfaces (S/PDIF, ADAT) where applicable.
Edit: Regarding the latency don't forget the delay caused by the converters itself (something from 16 to 32 samples with current A/D's), it's better to use digital interfaces (S/PDIF, ADAT) where applicable.
One tip for people considering DSP filters and owning a PC (and obviously having some competence in C++ programming): at www.steinberg.net you can DL demo version of Cubase VST or SX. Then DL the VST SDK and write a nice VST plugin which implements the crossover. There are many free apps that can be used to calculate the IR, though the Matlab solution must be better (for them who have access to it).
Edit:
Links:
SX demo
VST plugin SDK
Edit:
Links:
SX demo
VST plugin SDK
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.