In the DCX2496, each of the 9 parametrics in each output may be individually configured as low-shelf or hi-shelf with 6dB/oct or 12dB/oct slopes, or as bell. It really gets much closer to true crossover design than more expensive systems. Remember that it's under $300. Note that older firmware versions only allowed 6 parametrics per output and were probably suffering from other limitations that the latest version does not have.
As you mention, the more expensive units from very reputable brands are actually useless because they are focused on textbook responses with little ability to alter them. These units are for sure designed by people not understanding speakers. Very little people in PA world understands them anyway
As you mention, the more expensive units from very reputable brands are actually useless because they are focused on textbook responses with little ability to alter them. These units are for sure designed by people not understanding speakers. Very little people in PA world understands them anyway
I use the CS2310
I don't know the what type of bookshelf speakers you are trying to convert.
As many already pointed out some bookshelf speakers have reactive components to compensate for baffle diffraction and for impedance flattening etc.
Many cheap speakers use very simple straight forward 1 - 4th order passive filters.
In my case for my DIY speakers, baffle diffraction was not an issue.
In the evolution of going from passive to active - I ran separate leads to the drivers and kept the passive crossover external breadboarded near the amps. This facilitated changing crossover topologies and components for tweaking and evaluation. And I tried a wide variety.
After purchasing the aforementioned CS2310, and going totally active, there is no way I would go back to passive.
For a variety of reasons ( some already mentioned ).
I don't know the what type of bookshelf speakers you are trying to convert.
As many already pointed out some bookshelf speakers have reactive components to compensate for baffle diffraction and for impedance flattening etc.
Many cheap speakers use very simple straight forward 1 - 4th order passive filters.
In my case for my DIY speakers, baffle diffraction was not an issue.
In the evolution of going from passive to active - I ran separate leads to the drivers and kept the passive crossover external breadboarded near the amps. This facilitated changing crossover topologies and components for tweaking and evaluation. And I tried a wide variety.
After purchasing the aforementioned CS2310, and going totally active, there is no way I would go back to passive.
For a variety of reasons ( some already mentioned ).
For the DSP knowledgeable DIYer I propose buying a small PC box used from ebay and slapping a multi channel sound card in it. You'll get by less than $150 for the PC plus however much you want to spend on the sound card. With even a weak CPU (by today's standards) you can run any arbitrary crossover/EQ filters.
As you mention, the more expensive units from very reputable brands are actually useless because they are focused on textbook responses with little ability to alter them. These units are for sure designed by people not understanding speakers. Very little people in PA world understands them anyway
That actually is not what I said, what I said was that the very high end digital loudspeaker management systems are more useful, not less. The Dolby and DBX units are developed by people who understand the physics of speakers just fine, and I think that the majority of PA guys using them do as well, or at least, they should if they are running 10-20,000 dollar speaker management systems. Those systems are so complicated to setup I don't think a lay person would even know where to begin.
The DCX is a very inexpensive speaker management system with a great many capabilities. I don't happen to like it in home audio use without a lot of special considerations, and I abhor the 50 or so jrc jellybean opamps used in its circuit. I really wish it used a lower noise and higher slew unit than it does. I also think it has too much gain, and even with the gains properly adjusted so that the unit functions close to unity gain, I've found that it clips rather easily at the analogue input. It's definitely night and day running digital into it, but you then have to use a 6 channel analogue volume control after it. Using a digital domain volume control causes the unit to function with less bits, thus causing a loss of resolution and increased distortion. I am definitely no digital expert, so maybe the gentlemen from Ground Sound can speaker more appropriately on this, but I've seen the graphs of what happens when you use either its built in digital volume control and gain controls, or control the volume digital before the unit, and its not pretty. I swear either a magazine or person posted some graphs of this through a DSP2496 unit, but I've actually been involved in testing the DCX unit and when fed normal signal levels for normal listening, it gave an average bit resolution of less than 10. Comparing the waveforms encoded on the source to the output was night and day. And mind you, at least as far as I understand it, this isn't a problem with just the DCX, its not really a design flaw, its a problem with all units of this style. I think less gain would go a very long way in fixing some of the problems, but it wouldn't change the fact that the unit throws away bits and thus modifies the signal in a noticeable way. We aren't talking a change from 16 bits to 12 bits, which you might argue at these lower levels isn't audible, we are talking 6-10 bit loss in resolution.
One solution that I haven't seen in many units, which would fix a lot of these problems, would be a digitally controlled analogue volume control just like is used in pretty much all home theater receivers and pre/pro's. In fact, I didn't realize this until recently, but most all of these processors and receivers do the gain adjustment in the digital domain, not the analogue domain. This means if you make large reductions in level to a particular channel, you end up throwing away a lot of the bit resolution. This can cause lower dynamic range in that particular channel as well as lower its resolution. Also, if you do just the opposite and raise the level quite a bit, you stand a chance of overloading that channel during dynamic peaks (This has actually been a problem ever since bass management went digital). If you look at the particular systems speced by THX or TMH systems, they specify analogue domain bass management, not digital. When I talked with Mr. Holman, he told that the reason I mentioned above was why. He told me that, until the digital chips caught up with the need, he would not suggest doing any sort of gain or volume adjustments in the digital domain.
The PRO units have to be bullet proof in terms of levels to avoid overload and live up to the “standard” of very high input and output margins, which leads to low or very low resolution in the converters.
You’ll also have to do the right thing in the digital domain to avoiding “throwing” away bit resolution – a combination of software design, DSP and user control is often the issue. If you optimize the loudspeaker system and amplifiers you’ll have a system which performs much better than the PRO units in their normal configurations. Our complete systems are optimized all the way and a customer can concentrate on the loudspeaker setup.
The implementation can be a tough job; especially the room will affect the listening quality.
It’s often seen that people tend to lower the overall bass output to avoid booming bass, because they do not cure the problem of room modes. BUT if you take care of the room modes – you can have a nice "juicy" tight bass giving a perfect tonal balance. Getting the right bass level will actually cure many aggressive tendencies in a loudspeaker system without doing anything to the upper frequency range.
Some years back I demonstrated our active system to a Danish Audio Society/Club/group on a pair of B&W 2-way shelf loudspeakers. After the initial setup I made a new setup in which I boosted the bass around 50Hz and made a sharp cut below 40Hz. I updated the setup without telling them what I have done – then I asked them what they thought I did. They thought I made some EQ at the tweeter!
I know the DCX2496 and it can be optimized in terms of level and I have done this with a few units, but the whole box should be redesigned - also the main board.
You’ll also have to do the right thing in the digital domain to avoiding “throwing” away bit resolution – a combination of software design, DSP and user control is often the issue. If you optimize the loudspeaker system and amplifiers you’ll have a system which performs much better than the PRO units in their normal configurations. Our complete systems are optimized all the way and a customer can concentrate on the loudspeaker setup.
The implementation can be a tough job; especially the room will affect the listening quality.
It’s often seen that people tend to lower the overall bass output to avoid booming bass, because they do not cure the problem of room modes. BUT if you take care of the room modes – you can have a nice "juicy" tight bass giving a perfect tonal balance. Getting the right bass level will actually cure many aggressive tendencies in a loudspeaker system without doing anything to the upper frequency range.
Some years back I demonstrated our active system to a Danish Audio Society/Club/group on a pair of B&W 2-way shelf loudspeakers. After the initial setup I made a new setup in which I boosted the bass around 50Hz and made a sharp cut below 40Hz. I updated the setup without telling them what I have done – then I asked them what they thought I did. They thought I made some EQ at the tweeter!
I know the DCX2496 and it can be optimized in terms of level and I have done this with a few units, but the whole box should be redesigned - also the main board.
Robert GS said:The implementation can be a tough job; especially the room will affect the listening quality.
It’s often seen that people tend to lower the overall bass output to avoid booming bass, because they do not cure the problem of room modes. BUT if you take care of the room modes – you can have a nice "juicy" tight bass giving a perfect tonal balance. Getting the right bass level will actually cure many aggressive tendencies in a loudspeaker system without doing anything to the upper frequency range.
Some years back I demonstrated our active system to a Danish Audio Society/Club/group on a pair of B&W 2-way shelf loudspeakers. After the initial setup I made a new setup in which I boosted the bass around 50Hz and made a sharp cut below 40Hz. I updated the setup without telling them what I have done – then I asked them what they thought I did. They thought I made some EQ at the tweeter!
The problem is in human hearing as usual. Bass and midbass have a strong effect on the way we perceive higher frequencies. A loud tweeter without midrange or bass always sounds harsh, distorted, terrible... Adding the lower frequencies changes this dramatically, although nothing has actually changed about the tweeter. Dips between 100Hz and 500Hz (due to room acoustics) are perceived as harsh mid-highs. Our hearing is full of distortion mechanisms, only our sight is relatively good.
BTW: I use to pad the outputs of the DCX2496 to get improved resolution. It's as easy as not setting the gain knobs to the maximum on the amplifiers
I have not worked with other DSP crossovers but I have read many manuals from reputable brands in search of a more versatile unit, including BSS, dbx, SSL, Xilica, Yamaha... No one allows independent adjustment of the biquads (or the taps for FIR). Most don't even have phase shifters and don't allow more than 6 parametric EQs per output. The newest Xilica units include phase shifters as a type of parametric EQ (so you can use more than one per output) which is a nice feature, though.
That article is an opinion piece as far as I'm concerned. Not only is much of the arguments full of wholes, but more importantly, no evidence is presented to support their contention. I've talked with many knowledgeable speaker designers as well, and I can tell you that none of them agree with the premises presented in that article. Even companies that do offer active crossovers feel its one answer to a series of trade-offs that must be made in speaker design, not the only one. This isn't to say that I feel active crossovers is a bad thing or that it can't be implemented properly. That is not the case at all. I just don't agree that its universally better, and I don't think the average person can properly implement it. I see a lot of people go with active crossovers because they don't know how to properly design crossovers, and they think that buying an active crossover will be an easier way to build their own speakers. The reality is, properly tuning an active crossover is every bit as difficult as a passive crossover (The fundamental theories are the same), and these people unable to properly implement a passive crossover will not be able to implement an active crossover either.
You are too pessimistic. Making optimum active crossovers requires substantially less work because the impedance plot of every speaker driver no longer matters and because phase shifters are there to help you with acoustical phase matching.
Not to mention that dealing with op-amps and small resistors and capacitors (or DSP parameters) is easier and cheaper than dealing with big components.
For example, in a 2 way system good results may be already obtained with just a textbook LR24 and two adjustable phase shifters (each one mimics the phase shift of the opposite driver resulting in decent acoustical phase matching over a wide frequency range). This can't be done in passive.
Not to mention that dealing with op-amps and small resistors and capacitors (or DSP parameters) is easier and cheaper than dealing with big components.
For example, in a 2 way system good results may be already obtained with just a textbook LR24 and two adjustable phase shifters (each one mimics the phase shift of the opposite driver resulting in decent acoustical phase matching over a wide frequency range). This can't be done in passive.
and Eva I think you are too optimistic. I showed graphs and cam present more to support my view. I've done the foot work with evidence to support my contention. Your view is your view, and thats fine, we have a right to our opinions, but don't act like I must clearly be wrong because of what you believe.
Dear pjpoes,
we aren't here to convince you you can believe anything you want. But I'll stay with with the physics and the experts, who understand it like Klein+Hummel and Musicelectronic Geithain
Just one of a few things to think about: Yes, you need more amps, but a 60 W + 30 W amplifier in a 2-way active system have the same dynamic range as a single 175 W amp on a passive crossover for that speaker (see e.g. V. Dickason).
As a well educated person you are, you know Galilei's famous words: "And yet it moves!"
we aren't here to convince you
you can believe anything you want. But I'll stay with with the physics and the experts, who understand it like Klein+Hummel and Musicelectronic Geithain Just one of a few things to think about: Yes, you need more amps, but a 60 W + 30 W amplifier in a 2-way active system have the same dynamic range as a single 175 W amp on a passive crossover for that speaker (see e.g. V. Dickason).
As a well educated person you are, you know Galilei's famous words: "And yet it moves!"
I think that it’s not black and white. To change from a passive to active you need to really understand the passive network and what it’s doing. You would have to measure the voltage drive and be able to duplicate it with the active crossover. That may be much more difficult than it sounds depending on the specific crossover and driver set.
I am in the middle of “cloning” a 4 way pair of speakers and decided to bi-amp them between the bass and lower midrange driver. You would think you could just build the passive network and. leave out the sections for the woofer and transition to the lower midrange. Not so. By taking out one half of the bandpass network on the lower midrange driver you have to redo both the upper bandpass and driver zobel to keep the original voltage drive.
To simulate the voltage drives in this system with an active crossover you would need a rather sophisticated loudspeaker control system. A generic off the shelf active crossover would never work for this particular system.
In other systems it way work fine but you would have to look at the original passive design and determine up front if the active crossover was up to the task of not. You may end up with something that works and sounds good to you but you can’t just assume an active will work well in all circumstances.
Rob
I am in the middle of “cloning” a 4 way pair of speakers and decided to bi-amp them between the bass and lower midrange driver. You would think you could just build the passive network and. leave out the sections for the woofer and transition to the lower midrange. Not so. By taking out one half of the bandpass network on the lower midrange driver you have to redo both the upper bandpass and driver zobel to keep the original voltage drive.
To simulate the voltage drives in this system with an active crossover you would need a rather sophisticated loudspeaker control system. A generic off the shelf active crossover would never work for this particular system.
In other systems it way work fine but you would have to look at the original passive design and determine up front if the active crossover was up to the task of not. You may end up with something that works and sounds good to you but you can’t just assume an active will work well in all circumstances.
Rob
winny said:
Have you tried Lab.Gruppen DSP24? As you could alter just about everything, it was very very complicated and flopped totally. I would expect you to like it though...
I have have a pair and they could sound really really good IF you were lucky. It was very very hard to figure out what happened in the digital domain (clumsy software design) - I gave up and sold them after a while.
They were also pretty noisy - bad box planning. At that time the converters didn't have the quality of today. DSP24 should have been 3-way stereo and a display/control interface on the front for live edit.
Fosti said:Dear pjpoes,
we aren't here to convince you
Just one of a few things to think about: Yes, you need more amps, but a 60 W + 30 W amplifier in a 2-way active system have the same dynamic range as a single 175 W amp on a passive crossover for that speaker (see e.g. V. Dickason).
As a well educated person you are, you know Galilei's famous words: "And yet it moves!"
Dear Fosti,
I've probably taken a stronger and harder stance on this issue than even I believe, but it was to make a certain point, which I still believe is true. First let me make it perfectly clear to you, in an ideal and perfect world of electronics, I do agree and believe that an active system is better. I.e. Theoretically its a superior system, and when you begin to throw the realities of life and "physics" into the equation, under many circumstances it remains a better system. That has not been my point here.
My point was that an active crossover is no more flexible than a passive crossover and that they are just as difficult to properly setup as a passive crossover.
There are also certain flawed arguments used to support active crossovers over passive ones, which simply tell me that the these people do not understand the physics we are dealing with. One of those arguments is that there is no inductance, capacitance, or resistance variability to effect the crossover, so the slopes are much better. While fundamentally correct, the active crossover will not be effected by the speakers own inductance, capacitance, and resistance, it also doesn't matter. If you are correctly designing a passive crossover, you will have measured and controlled for all of that. This means that the speakers own effect will be known and well accounted for in a properly optimized passive crossover, and its effect will be negligible. The counter argument to this is that the crossover is based on a fairly weak signal (Which actually isn't usually true), and that the inductance, resistance, and capacitance of the speaker will change depending on where the speaker cone is. This is true, but its effect is still negligable for most of its range, and when it does become a problem, it would be a problem for an active crossover too.
The biggest flawed argument used, related to the above argument, is that the amplifier sees an easier load from the speakers. While its true that the amplifier has a straighter path, and thus is not interacting with the crossover, the speakers own voice coil is still an inductor, and that load on the amplifier is actually generally worse, not better. With a properly optimized crossover you can add conjugate filters to properly deal things like rising inductance, mechanical ringing, or rising impedance. In fact, the only way to deal with those are with passive parts, and they do have an effect on the frequency response and ease with which an amplifier can drive a speaker. The only area where active crossovers do win out here is that an active crossover will not have the same effect on the electrical phase of the system as a passive crossover would, and so as is the case with some highend speakers which present a very demanding load of low impedance and strong negative phase, this will be partly mitigated. However a lot of this has to do with the ease with which higher order crossover slopes can be achieved using complex transfer functions which achieve near transient perfect results. However, this isn't true of the large majority of active crossovers, and many people have argued that with real musical signals, their is no benefit to a transient perfect crossover anyway.
Then you have the dynamic range discussion. Again, its a flawed argument that doesn't hold up to scientific investigation. Mind you, I'm not saying its completely wrong or untrue, its just exaggerated and misrepresented. Vance Dickinson didn't create two completely equal speakers, one having a passive crossover at its best and one having an active crossover at its best, and show 3+ db's greater efficiency. His comment is mostly subjective opinion about the subjective effect of setting up a speaker this way. While this is some objective evidence to support the view, there are also numerous counter arguments to suggest the reason for the dynamic range increase are due to other factors having nothing directly to do with the different crossover types. Another problem with this theory is that it seems to imply that the major limiting factor in a speakers peak output is amplifier power, and that some how, as if by magic, driving the drivers by separate crossovers increases this output beyond the physical limits of the system. Again though, I'm not saying there are no benefits to multi-amping, just that the improvements are a bit exaggerated and mis-represented.
My final comment on this here is that you can not take active crossovers at their best and passive crossovers at their worst and say, "See here, active crossovers are clearly superior, look at all the bad things that passive crossovers do." You can turn that right around and take passive crossovers at their best and show how superior they are to active crossovers at their worst. However, that isn't a fair comparison, its not representative of the technology.
To perform passive design one needs some knowledge of filter design couple with working experience with filter design(comes with practice) with speakers in the real world. Hence this needs time and money with purchases of different values of caps and inductors. This process can be time consuming and fructrating at times. To get them(ie filter) right/optimal is by no means easy.
However using active digital x-over like DCX2496, one really do NOT need any good knowlegde of electrical filtering concepts to get a speaker working in a reasonable "listenable" conditions without performing transfer functions analysis. What is required is a microphone and s/w to adjust the FR to a reasonably flat and adust the filter contour to a state where the speakers means your satisfaction, in this case mine, BTW it is easy I don't design and build for commercial purposes or even for other listeners. That had been what I was doing all this while and I found it quite satisfactory.
The next step for me is actually design a pasive filter using circuit anlaysis via s/w and microphone measurement and then compare it against what I have done above. I wonder whether is would be clearly be superior to my technique via DCX.
I think(suspect, no proof) a superior sound system would be to design an "optimal passive filter" and drive them separately through 2/3 amps depending whether it is a 2 or 3 way speaker., ie without any active filter in between.
cheers.
However using active digital x-over like DCX2496, one really do NOT need any good knowlegde of electrical filtering concepts to get a speaker working in a reasonable "listenable" conditions without performing transfer functions analysis. What is required is a microphone and s/w to adjust the FR to a reasonably flat and adust the filter contour to a state where the speakers means your satisfaction, in this case mine, BTW it is easy I don't design and build for commercial purposes or even for other listeners. That had been what I was doing all this while and I found it quite satisfactory.
The next step for me is actually design a pasive filter using circuit anlaysis via s/w and microphone measurement and then compare it against what I have done above. I wonder whether is would be clearly be superior to my technique via DCX.
I think(suspect, no proof) a superior sound system would be to design an "optimal passive filter" and drive them separately through 2/3 amps depending whether it is a 2 or 3 way speaker., ie without any active filter in between.
cheers.
Interesting... soundengine355
Asked a fairly straightforward question, and a lengthy debate begins.
Though out the discussion is the common agreement of compromise.
1st big compromise - a round speaker in a rectilinear box.
Does the existing passive crossover attempt to solve baffle diffraction?
Speakers are reactive - Does it incorporate Zobels to compensate for rising impedance?
If the existing bookshelf offered some of these elements and indicated that it was designed with some level of sophistication beyond 2 inductors, 2 capacitors and perhaps a resistor to pad a tweeter - Then I would be much more reluctant to modify it.
AudioXpress ran a 2 part article of Passive Crossover Linear Phase Speakers ( 6/08, 7/08 ) focusing on a transient perfect system. Some of the approaches are very sophisticated - all with some compromises ( or side effects if you will ). There are numerous linear phase CO designs that can ONLY be implemented in an active format. It recommended Active CO and digital technology if linear phase was the goal.
Sept 2008 audioXpress has an article on tweeter setback - with the goal of reducing delay distortion.
The problem was with no tweeter setback, there is too much delay dispersion, and with complete setback, a poor off axis frequency response.
Passive components introduce delay. And while the author admits that he could have used a passive approach, he chose to use a DCX2496 for ease of experimentation.
One of the problems with caps and inductors is they come in standard values with 3% tolerance for quality inductors, and quality caps @ 5%. This means ( unless your winding your own coils ) a compromise on component values or careful parts matching. As the the parts count increases so does the possibility of cumulative error.
A small voice coil in the average inefficient bookshelf will heat rather quickly. It does not take much power to cause the resistance to change by 8% or more. Since a passive crossover is designed with "cold" parameters in mind that causes problems. Active CO avoids this because it is not subject to the heat problems caused by handling amperes.
As a strategy - I would 1st outboard the crossovers. I have found this alone is an improvement over inside the cab.
2nd. Check the quality of the passive components and replace with better grade and implement capacitor bypassing.
This would still allow you to experiment with active CO and revert to an optimized Passive if it didn't sound better.
Not to repeat things that Eva said - my conversion using the CX2310 - by removing the reactive components from the amp load - I gained a large boost in clarity.
Asked a fairly straightforward question, and a lengthy debate begins.
Though out the discussion is the common agreement of compromise.
1st big compromise - a round speaker in a rectilinear box.
Does the existing passive crossover attempt to solve baffle diffraction?
Speakers are reactive - Does it incorporate Zobels to compensate for rising impedance?
If the existing bookshelf offered some of these elements and indicated that it was designed with some level of sophistication beyond 2 inductors, 2 capacitors and perhaps a resistor to pad a tweeter - Then I would be much more reluctant to modify it.
AudioXpress ran a 2 part article of Passive Crossover Linear Phase Speakers ( 6/08, 7/08 ) focusing on a transient perfect system. Some of the approaches are very sophisticated - all with some compromises ( or side effects if you will ). There are numerous linear phase CO designs that can ONLY be implemented in an active format. It recommended Active CO and digital technology if linear phase was the goal.
Sept 2008 audioXpress has an article on tweeter setback - with the goal of reducing delay distortion.
The problem was with no tweeter setback, there is too much delay dispersion, and with complete setback, a poor off axis frequency response.
Passive components introduce delay. And while the author admits that he could have used a passive approach, he chose to use a DCX2496 for ease of experimentation.
One of the problems with caps and inductors is they come in standard values with 3% tolerance for quality inductors, and quality caps @ 5%. This means ( unless your winding your own coils ) a compromise on component values or careful parts matching. As the the parts count increases so does the possibility of cumulative error.
A small voice coil in the average inefficient bookshelf will heat rather quickly. It does not take much power to cause the resistance to change by 8% or more. Since a passive crossover is designed with "cold" parameters in mind that causes problems. Active CO avoids this because it is not subject to the heat problems caused by handling amperes.
As a strategy - I would 1st outboard the crossovers. I have found this alone is an improvement over inside the cab.
2nd. Check the quality of the passive components and replace with better grade and implement capacitor bypassing.
This would still allow you to experiment with active CO and revert to an optimized Passive if it didn't sound better.
Not to repeat things that Eva said - my conversion using the CX2310 - by removing the reactive components from the amp load - I gained a large boost in clarity.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.