Hi,
This sounds very interesting.
I still remember the early analogue "processor" PA systems and the beginnings of "steered arrays" before I got out of Pro-Audio. Almost all implementations I see nowadays for larger scale PA's are simply infeasible without DSP.
As usual, the Pro-Audio side is in the Vanguard, HiFi will catch up.
I am already beginning to see "steered directivity" systems for home AV applications.
Ciao T
This sounds very interesting.
I still remember the early analogue "processor" PA systems and the beginnings of "steered arrays" before I got out of Pro-Audio. Almost all implementations I see nowadays for larger scale PA's are simply infeasible without DSP.
As usual, the Pro-Audio side is in the Vanguard, HiFi will catch up.
I am already beginning to see "steered directivity" systems for home AV applications.
Ciao T
They are in fact available for quite some time now:
Digital Sound Projector - Yamaha - United States
Regards
Charles
Hi,
IF (big if) I understood the way these work they are not classic steered arrays... But I may be wrong.
Ciao T
IF (big if) I understood the way these work they are not classic steered arrays... But I may be wrong.
Ciao T
They indeed are steered arrays but not classic ones. They are indeed about 4 arrays in one !!!
1. Front center direct radiating
2. Front left and right: for each one a steered beam that is meant to bounce off the sidewalls.
3. Back: steered beam that is meant to bounce of the back wall with an intermediate hop via the ceiling.
It has some dozens of tiny wideband drivers each with its own tiny class-d amp.
Regards
Charles
1. Front center direct radiating
2. Front left and right: for each one a steered beam that is meant to bounce off the sidewalls.
3. Back: steered beam that is meant to bounce of the back wall with an intermediate hop via the ceiling.
It has some dozens of tiny wideband drivers each with its own tiny class-d amp.
Regards
Charles
Hi,
My understanding was that the drivers as such where not actually fed audio as such, but just (digital) impulses.
Maybe this was about a prototype, instead the real ones went back to traditional steered arrays?
Ciao T
My understanding was that the drivers as such where not actually fed audio as such, but just (digital) impulses.
Maybe this was about a prototype, instead the real ones went back to traditional steered arrays?
Ciao T
Yes, there are a number of DSP solutions. The ADAU devices from Analog are popular and a development kit and SigmaStudio software are available. Some of these setups are the heart of a few commercial solutions (like the miniDSP) as well.
SigmaStudio? Graphical Development Tool | SigmaStudio Overview | ADAU1701 | Audio Signal Processors | Audio/Video Products | Analog Devices
Cheers,
Dave.
SigmaStudio? Graphical Development Tool | SigmaStudio Overview | ADAU1701 | Audio Signal Processors | Audio/Video Products | Analog Devices
Cheers,
Dave.
It was probably a tiny bit of exageration when I wrote that the only (low level) analog circuits of (any remaining) interest were microphone pre-amps . . . you are correct in noting that the interface to the A/D converter and from the D/A remain. But if ever there was a function that suggests a single-chip "solution" that's it, and if the chip designers haven't done it already I'm sure they soon will. Something like the INA134 and DRV135, perhaps, optimized for direct connection to the converters (if those won't work as-is).
A discussion of crossovers, then, would be about . . . crossovers (or more accurately, signal splitters). A discussion of how we implemented them back in the dark ages with analog circuits would make a welcome chapter (that could be skipped by anyone interested merely in modern crossover design and implementation), but all the important aspects of the entire "crossover" discussion could be done without a single reference to physical implementation. The following chapter on how we "make it real" would take only a couple paragraphs to show why it's no longer done "analog" except in the most trivial cases (small two-way active monitors, for example).
It seems the thread has drifted to another book about another related subject.
As one who is in the process of of converting a near-field 2-way to active crossover, I'm actually interested in the OP subject. I don't see that DSP makes any economical sense in simple or analog speaker systems unless it is part of a larger digital system such as PC based, Home Theatre or as earlier discussed; commercial installations.
I accept that others, professionals particularly, may now be accustomed to all-digital processing in audio, at least in their workplace or at the computer desk. This does not mean that we all use these systems or have any need for ongoing adjustability or complex functions in crossovers. So how about a little more constructive thinking on the OP subject?
As one who is in the process of of converting a near-field 2-way to active crossover, I'm actually interested in the OP subject. I don't see that DSP makes any economical sense in simple or analog speaker systems unless it is part of a larger digital system such as PC based, Home Theatre or as earlier discussed; commercial installations.
I accept that others, professionals particularly, may now be accustomed to all-digital processing in audio, at least in their workplace or at the computer desk. This does not mean that we all use these systems or have any need for ongoing adjustability or complex functions in crossovers. So how about a little more constructive thinking on the OP subject?
You mean this?
I thought that was (mostly) what folks were talking about.
Thank you for reminding me Pano, point taken also from earlier posts. The inference from many posts is that these guys want a different book about a different subject, i.e DSP. I think that is clear. Seeing this was knocked out in the OP and following, I kinda thought I should stay in bounds but others have their views and all was fine, I hoped.
However, when dewardh suggested only one chapter in recognition of active filters, you plainly see that we are on divergent paths and not talking about missing content at all IMHO.
However, when dewardh suggested only one chapter in recognition of active filters, you plainly see that we are on divergent paths and not talking about missing content at all IMHO.
As potential purchasers of the book, I think it is fair that people also be concerned about the scope of the work to be covered. My guess is that the author means to share his experience with the reader. But some are questioning whether this would comprise a sufficiently significant body of work, given the current state of technology. This might make it "a different book", but I think it is still constructive. Free market research, I'd say.
But what do you need to know about that . . . that isn't already available for free on the web? Certainly not 500 pages worth . . .
For sure it doesn't make sense to put a $100 DSP board in a $200 "price point" speaker. And if you're going to get the design right the first time, or amortize the design cost over a long production run (or do it for free, for fun), easy adjustability (during the design process) may not matter either. But everything you need to simulate that simple 2-way is available free (PCD), can be designed with circuits already described free (linkwitzlab, westhost, etc.), and implemented on circuit boards available from a variety of hobby vendors, who generally provide free yet another set of calculators to determine circuit values. You'll probably still end up with $100 in parts and many tens of hours of time "designing", shopping, building, testing, etc., but hey, it's a hobby, right? Or you could just build a design from one of those already mentioned vendors.
Or you can spend the $100 "up front" for a miniDSP board, program it in an hour or two (first time through, including "learning curve"), and hear what the finished speaker will sound like. And then change it in a matter of minutes if you want to. With biquads and delay for adjusting acoustic center and all that good stuff that's hard[er] to implement analog if you want it. All without buying a book to tell you how to do it the hard way (which, again, you can learn for free elsewhere). And if you don't want to leave the DSP in the speaker you can still implement the final feature set analog using the free stuff mentioned above (plus parts).
For a simple 2-way, analog or digital, you don't *need* a book . . . for a "complicated" design you don't want to do it analog (and if you do anyway then linkwitz has explained, and demonstrated (see ORION), how. What seems to be the common sentiment about Doug's proposed book is that a "review of the field" is good "as it used to be" not so much. But if it's what it takes to get you "active" then by all means buy a copy. Just don't forget to charge it as a cost to your 2-way monitors.
Hi,
That is debatable.
Again, a DSP based solution can do a lot of things you cannot do with analogue systems, or if only with immense effort.
I can see someone implementing a really cheap 2-Way crossover (if need be it can be wrapped directly around the power amplifiers, no need for anything extra) in order to cut costs in a commercial product.
However, given how time-consuming it is to optimise an analogue crossover to actually match the drivers, produce the correct lobing (or better, no lobing at all), which is especially critical in nearfield systems and so on I would feel uncomfortable to invest soi much time on changing components etc.
And are you telling me that the ultimate source for your two nearfields is an analogue microphone feed or an LP-Player, solely?
Otherwise you do have a digital source, that is converted to analog somewhere.
I worked quite extensively on active speakers in the 80's (it was THE thing at the time), including allpass time delay and even such niceties as current feed on midranges and tweeters and current-feed plus motional feedback on the woofers.
The amount of time taken to optimise merely the analogue filtering part was dramatic (okay, it was a 3-way system) and textbook filters did not work, without adding even more stuff to equalise.
By comparison I look at this:
UE_OB
And then I ask myself, why work so hard as I did in the 80's if I don't have to?
But maybe you are not looking for really high quality results, or even a speaker with a flat resulting response and you just want a basic textbook filter.
In that case there is SO MUCH information on filters (including for example an on-line designer by AD) that designing the whole thing should not take more than 5 minutes.
Ciao T
That is debatable.
Again, a DSP based solution can do a lot of things you cannot do with analogue systems, or if only with immense effort.
I can see someone implementing a really cheap 2-Way crossover (if need be it can be wrapped directly around the power amplifiers, no need for anything extra) in order to cut costs in a commercial product.
However, given how time-consuming it is to optimise an analogue crossover to actually match the drivers, produce the correct lobing (or better, no lobing at all), which is especially critical in nearfield systems and so on I would feel uncomfortable to invest soi much time on changing components etc.
And are you telling me that the ultimate source for your two nearfields is an analogue microphone feed or an LP-Player, solely?
Otherwise you do have a digital source, that is converted to analog somewhere.
I worked quite extensively on active speakers in the 80's (it was THE thing at the time), including allpass time delay and even such niceties as current feed on midranges and tweeters and current-feed plus motional feedback on the woofers.
The amount of time taken to optimise merely the analogue filtering part was dramatic (okay, it was a 3-way system) and textbook filters did not work, without adding even more stuff to equalise.
By comparison I look at this:
UE_OB
And then I ask myself, why work so hard as I did in the 80's if I don't have to?
But maybe you are not looking for really high quality results, or even a speaker with a flat resulting response and you just want a basic textbook filter.
In that case there is SO MUCH information on filters (including for example an on-line designer by AD) that designing the whole thing should not take more than 5 minutes.
Ciao T
I'm not going to argue with the point that DSP's / digital are the way of the future, and I agree that there should be some space given to them in the book.
However I suspect there are many different reasons why a person may decide to make an analogue crossover rather than a digital one. Some of us get just as much enjoyment out of designing and building something as we do from listening to it (sometimes more if we didn't do a great job on the design side
) I can see why people would want to do this rather than just get a plug and play module and program it with a gui interface. If your goal is to get great sound the easiest possible way then sure the DSP seems a no brainer, but that is not everyone's bent.
Some might like the challenge of doing it the old way rather than taking the easy way out. Maybe the results won't be as good, but maybe they will be better!
Some may be technophobes and have no interest in those new fangled computer thingies.
Some may strongly feel that digital is still just an approximation of what analogue can reproduce and not have any digital in their systems at all.
Some may just want to "not be main stream".
On top of all that, I can see danger in the power of the DSP. I can see design considerations that would be absolutely necessary when done the old way being brushed aside because "we can fix it with the DSP". Is this always (ever?) going to result in a better sound? I can see people becoming lazy and massaging the signal in order to get it to come out right in the end, rather than designing to minimise the need for correction in the first place... which will be better?
Of course one can apply all of the old design rules and then need minimal massaging of the digital signal as well, I'm not trying to imply that can't happen as well.
On the point of all of the information being available on the net, yes maybe it is, but it can take a LONG time to find it! Many people actually like to have a reference in a nice handy portable form. I couldn't get by these days without google, but I also love to have a well written reference book too.
Just some food for thought.
Tony.
PS. Yes I am attempting to design and build an analogue crossover, and yes some (but not all) of the points above apply to me
PPS. If this book had have been available when I first started to think about my active crossover I probably would have bought it. Though then I'd likely never have gone down the path I have.
However I suspect there are many different reasons why a person may decide to make an analogue crossover rather than a digital one. Some of us get just as much enjoyment out of designing and building something as we do from listening to it (sometimes more if we didn't do a great job on the design side
) I can see why people would want to do this rather than just get a plug and play module and program it with a gui interface. If your goal is to get great sound the easiest possible way then sure the DSP seems a no brainer, but that is not everyone's bent. Some might like the challenge of doing it the old way rather than taking the easy way out. Maybe the results won't be as good, but maybe they will be better!
Some may be technophobes and have no interest in those new fangled computer thingies.
Some may strongly feel that digital is still just an approximation of what analogue can reproduce and not have any digital in their systems at all.
Some may just want to "not be main stream".
On top of all that, I can see danger in the power of the DSP. I can see design considerations that would be absolutely necessary when done the old way being brushed aside because "we can fix it with the DSP". Is this always (ever?) going to result in a better sound? I can see people becoming lazy and massaging the signal in order to get it to come out right in the end, rather than designing to minimise the need for correction in the first place... which will be better?
Of course one can apply all of the old design rules and then need minimal massaging of the digital signal as well, I'm not trying to imply that can't happen as well.
On the point of all of the information being available on the net, yes maybe it is, but it can take a LONG time to find it! Many people actually like to have a reference in a nice handy portable form. I couldn't get by these days without google, but I also love to have a well written reference book too.
Just some food for thought.
Tony.
PS. Yes I am attempting to design and build an analogue crossover, and yes some (but not all) of the points above apply to me
PPS. If this book had have been available when I first started to think about my active crossover I probably would have bought it. Though then I'd likely never have gone down the path I have.
Last edited:
Certainly, Shaun,dewardh, ThorstenL & Tony, I understand your viewpoints. My prototype is growing on a spare chunk of 35 x 60 mm stripboard and uses scrapbox bits. No dollars have been directly spent, which is how I like crossovers, active or passive. I probably overspend on parts for amps and other, bigger toys to get better price breaks. These flow into smaller works like this and still give me parts matching benefits. 'Nothing new there.
As wintermute has comprehensively illustrated, the sensible, modern and prebuilt way is not always best or satisfying to everyone. On another note, how do you see PLLXOs... another face of the same outmoded pursuit, or are they too recent and popular a concept to criticise?
My hope however, regardless of current interests, is to see some other considered views about what's in and what's not in regard simply to active crossovers. Perhaps this is just too hard an ask.
In courtesy however, if I didn't need or welcome a book by its title description or preferred some other subject, I would not post there. Simple.
If the whole offering was, in your view, such a universal waste of effort that you felt compelled to make a remark of disapproval, I would suggest making it with the least pretence of constructive criticism, wind and waffle possible. There is no point labouring such comments unless it is particularly necessary to justify your stance.
As wintermute has comprehensively illustrated, the sensible, modern and prebuilt way is not always best or satisfying to everyone. On another note, how do you see PLLXOs... another face of the same outmoded pursuit, or are they too recent and popular a concept to criticise?
My hope however, regardless of current interests, is to see some other considered views about what's in and what's not in regard simply to active crossovers. Perhaps this is just too hard an ask.
In courtesy however, if I didn't need or welcome a book by its title description or preferred some other subject, I would not post there. Simple.
If the whole offering was, in your view, such a universal waste of effort that you felt compelled to make a remark of disapproval, I would suggest making it with the least pretence of constructive criticism, wind and waffle possible. There is no point labouring such comments unless it is particularly necessary to justify your stance.
I *certainly* would have (and said so). Then came the DCX2496, and with it a new, better, and much easier way of thinking about crossover design (still a bit expensive for implementation in a home system). Now we've got DSP boards for $100, which is below the cost of 3-way, and approaching the cost of even simple 2-way, analog implementations. See the trend?
Doug asked, and got some answers. Some people still want to build (and optimize) old analog designs. Other's are going (or have gone) digital, part of a trend to "active" crossovers that is, by and large, *driven by the ease of design and good performance of digital crossovers*. The more the book addresses that the more useful it will be for years to come.
I also use the DCX2496, but will still definitely buy the book because, as Doug already stated many times, much of the content is applicable to both DSP and analogue.
I really think this shared part is actually the most interesting topic, filter theory and how to choose between them. So for me the book doesn’t need to be a cookbook with implementation details, but a book that teaches me to know what it is I exactly need. ..
Once I know what I need, actually building it is most of the times not that hard, when using google and this forum…
So its clear, Doug, if you want to write a book that makes everybody happy here, I think you will need 50.000 pages (without raising the cost of the book, of coarse J )
I really think this shared part is actually the most interesting topic, filter theory and how to choose between them. So for me the book doesn’t need to be a cookbook with implementation details, but a book that teaches me to know what it is I exactly need. ..
Once I know what I need, actually building it is most of the times not that hard, when using google and this forum…
So its clear, Doug, if you want to write a book that makes everybody happy here, I think you will need 50.000 pages (without raising the cost of the book, of coarse J )
Hi,
Yes, I used to stuff like this. Long ago.
PLLXO's are not a recent concept. They are nearly as old as multiway audio systems (say late 1920's).
Note, my point was not that no-one uses passive crossovers (line-level or speaker-level) or analogue active ones, but the fact that almost all pertinent information for these has been long available and nowadays is trivial to find.
A Book on audio circuits I purchased in the early 80's in East Germany had several chapters dealing with active crossovers, equalisation and time-delay.
So what DS proposes is to write a book on the state of the art appx. a generation ago. Even his approach to synthesising the filters needed seems to hark back to this time.
So I feel such a book would be of mostly historical interest. Sure, there are still people writing books on steam locomotives and people buying them. But not many will pretend it had anything to do with current technology.
As for:
Well, DS asked for opinions as to what was missing in his proposed TOC.
He got answers.
He then defended his proposed TOC on the grounds that DSP lack the necessary performance. His view in that context was clearly wrong. He insisted that 2/3 rds of his book are applicable to DSP implementations, however all they can help with is to emulate analogue circuitry with DSP, which is fairly stupid in this day...
Also, to be clear, I do not think that the book DS proposes as such is a waste of time. It would be nice to have good book on the subject. However it seems extremely unbalanced in favour of technology that is rapidly becoming obsolete in those areas where it not already is.
As it is I suspect John K's few pages on his website on using a digital XO running on a PC to produce an active speaker that offers flat frequency domain response AND completely linear phase has more relevance than the whole Book DS proposes.
As DS is a commercial Author who charges for his work I think he SHOULD know that. What he does about it is his call, but it will influence strongly the way his book sells I should think.
As for your crossover build from scrap, have fun. I suspect you will be able to spend a lot of time soldering and adjusting the responses, when instead you could have spend halve an hour pressing buttons and then listening to music.
I guess the work in hardware is by far more therapeutic.
Ciao T
Yes, I used to stuff like this. Long ago.
PLLXO's are not a recent concept. They are nearly as old as multiway audio systems (say late 1920's).
Note, my point was not that no-one uses passive crossovers (line-level or speaker-level) or analogue active ones, but the fact that almost all pertinent information for these has been long available and nowadays is trivial to find.
A Book on audio circuits I purchased in the early 80's in East Germany had several chapters dealing with active crossovers, equalisation and time-delay.
So what DS proposes is to write a book on the state of the art appx. a generation ago. Even his approach to synthesising the filters needed seems to hark back to this time.
So I feel such a book would be of mostly historical interest. Sure, there are still people writing books on steam locomotives and people buying them. But not many will pretend it had anything to do with current technology.
As for:
Well, DS asked for opinions as to what was missing in his proposed TOC.
He got answers.
He then defended his proposed TOC on the grounds that DSP lack the necessary performance. His view in that context was clearly wrong. He insisted that 2/3 rds of his book are applicable to DSP implementations, however all they can help with is to emulate analogue circuitry with DSP, which is fairly stupid in this day...
Also, to be clear, I do not think that the book DS proposes as such is a waste of time. It would be nice to have good book on the subject. However it seems extremely unbalanced in favour of technology that is rapidly becoming obsolete in those areas where it not already is.
As it is I suspect John K's few pages on his website on using a digital XO running on a PC to produce an active speaker that offers flat frequency domain response AND completely linear phase has more relevance than the whole Book DS proposes.
As DS is a commercial Author who charges for his work I think he SHOULD know that. What he does about it is his call, but it will influence strongly the way his book sells I should think.
As for your crossover build from scrap, have fun. I suspect you will be able to spend a lot of time soldering and adjusting the responses, when instead you could have spend halve an hour pressing buttons and then listening to music.
I guess the work in hardware is by far more therapeutic.
Ciao T
By now it should be clear that a least the hobbyists are interested in DSP implementations of crossover filters. I would think the pros are, too.Quite likely. But is that the only audience that Douglas is trying to reach? Probably not.
He asked what was missing. There has been a resounding answer of "More DSP, please!" Seems pretty simple.
No?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.