View Full Version : Active Crossover as plugin or otherwise
By popular (or at least Curt962) demand I am opening this as a new thread, from its humble origins as a ramble at the end of a Pissing Contest.
I wonder there may be some scope to introduce a crossover correction as an SBS plugin, along the lines of Inguz DRC. It seems to me that this would be an elegant solution cutting out the need for expensive and possibly sound degrading additional boxes.
Advocates of active crossover system argue that it reduces load on amps and means that cheaper speakers and amps can be used to produce the same level of performance or higher.
However I am not sure how (even say with 2 way speakers) the corrected signal could be turned into 4 analogue line level signals. Is it possible to create 2 separate synched 2 channel streams which could be sent via 2 squeezeboxes via analog out or perhaps coax to 2 dacs. Or could there be scope for a double decker squeezebox with 4 analog / 2 simultaneous digital outputs.
Does anyone have any interest in the development of such a product?
Well this is nice isn't it? Crisps anyone? I'm sure the others will be along in a minute...
Well I have a pair of home-built 2-way transmission line speakers, in which I wired the drive units directly to a 4-way speaker plug in the back of the cabinet. Currently I have conventional crossovers in the speaker bases, but connecting each drive unit to a separate source remains a possibility. A double-decker squeezebox could be an interesting DIY project. One squeezebox per speaker would mean that left and right channel on each squeezebox would be completely different - ie left=tweeter, right=bass driver.
Just as a matter of interest, how tight is the synchronisation between 2 synched squeezeboxes - is it good enough for each speaker to be on a separate squeezebox, or for the bass drivers and tweeters to be on separate squeezeboxes?
Hey Adam,
I hope you didn't think I was "shooing" you to a new thread,or something.
I just thought that the topic was interesting, and should have it's OWN thread rather than being buried in the midst of another.
Deaf Cat
2011-03-28, 14:27
Sorry, just to check what you are talking about, or more like do I understand what you are talking about ;-)
Left signal from amp goes to left speaker, passive crossover in speaker blocks off low freq's to tweeter with capacitor.. and blocks high freq's with a coil thing.. to the woofer.
Ditto with right side.
Active crossover filters out the unwanted high and low bits before the signal gets to the amp(s), so one stereo amp for the tweeters and one stereo amp for the woofers, no wasted amp power...
If I understand this crossover thing correctly, I'll reading any further comments with interest - Very Good Thinking, I like the idea.
No passive crossovers and No active crossover, just software....?
Cool!
:-D
Sorry, just to check what you are talking about, or more like do I understand what you are talking about ;-)
Left signal from amp goes to left speaker, passive crossover in speaker blocks off low freq's to tweeter with capacitor.. and blocks high freq's with a coil thing.. to the woofer.
Ditto with right side.
Active crossover filters out the unwanted high and low bits before the signal gets to the amp(s), so one stereo amp for the tweeters and one stereo amp for the woofers, no wasted amp power...
If I understand this crossover thing correctly, I'll reading any further comments with interest - Very Good Thinking, I like the idea.
No passive crossovers and No active crossover, just software....?
Cool!
:-D
Yes that was the idea. As Phil pointed out elsewhere you can do this with a tact preamp/dac but that costs thousands whereas this would hopefully need only an additional squeezebox and amp. The problem is that I'm not sure that the 2 player synch would work. I was hoping that one of the clever developer folks around here could be persuaded to have a look into it.
If it is done in the SBS we are talking about at least a TCP stream per pair of speaker element, where each stream carries the left and right channel for, eg. one stream for woofer, one for midrange, etc.
These has to be synchronized so they time gap is not too large and weīre speaking of micro seconds here.
Although technical feasible to do in software given that the netwotk is stable enough, I donīt see that the current SBS architecture supporting that.
Without some major software project...
A theoretical idea (assuming a 2-way system)
* TOSlink output of Touch/Transporter split into a "Y".
* TWO (2) DEQ2496s. One for high pass, and one for low pass.
* TOSlink or AES/EBU output of DEQ2496 to HP/LP DACs
* DACs to amps.
Assuming you can even split a TOSlink connection into a "Y"...
In the digital domain, the DEQ is no less transparent IMO than Inguz, and therein you can create your XO curves. Delay features in the DEQ could be tweaked for time alignment of drivers. You could use the DEQ's internal DACs if needed.
This would be MY off the shelf attempt to keep the XOs in a purely digital domain, and then sending the split signal off to nice quality DACs and amps.
The problem is that you're still buying multiples of everything. It will get prohibitively expensive quickly.
For a DIYer with money to spend it might be worthwhile. I still believe a commercial product with all the engineering already done is the most effective solution.
Without some major software project...
A theoretical idea (assuming a 2-way system)
* TOSlink output of Touch/Transporter split into a "Y".
* TWO (2) DEQ2496s. One for high pass, and one for low pass.
* TOSlink or AES/EBU output of DEQ2496 to HP/LP DACs
* DACs to amps.
Assuming you can even split a TOSlink connection into a "Y"...
In the digital domain, the DEQ is no less transparent IMO than Inguz, and therein you can create your XO curves. Delay features in the DEQ could be tweaked for time alignment of drivers. You could use the DEQ's internal DACs if needed.
This would be MY off the shelf attempt to keep the XOs in a purely digital domain, and then sending the split signal off to nice quality DACs and amps.
The problem is that you're still buying multiples of everything. It will get prohibitively expensive quickly.
For a DIYer with money to spend it might be worthwhile. I still believe a commercial product with all the engineering already done is the most effective solution.
I'm not usually this blunt, but someone has to say it - this is a CRAZY idea. People like Phil have been gently and politely trying to point out the futility of this scheme for a while, but it seems not to have got through.
Crossovers that operate from a full range stereo signal are well understood (regardless of whether they operate in the analogue or digital domain) and need not cost a lot of money. The cost of the crossover in an active system (which requires one power amp per drive unit) is not that significant.
Contemplating doing all this as some sort of SBS plugin and hoping to use multiple sync'd Squeezeboxes to deliver the split signals is frankly bonkers.
Phil Leigh
2011-03-29, 00:31
I thought the aim was to do the XO server-side? - if not it's easy:
You need:
1 SB player
2 DEQ's
2 identical stereo amps
pair of 2-way speakers with xo removed
4 speaker cables
1 toslink lead from SB to DEQ 1 (configured as Low pass, connected to bass drivers via amp 1)
1 s/pdif lead from SB to DEQ 2 (configured as high pass, connected to treble drivers via amp 2)
No sync issues at all.
You can use the internal DEQ dacs, or external dacs
and of course you can still run Inguz for DRC server side!!!!
Robin Bowes
2011-03-29, 00:32
On 29/03/11 08:26, cliveb wrote:
>
> I'm not usually this blunt, but someone has to say it - this is a CRAZY
> idea. People like Phil have been gently and politely trying to point
> out the futility of this scheme for a while, but it seems not to have
> got through.
>
> Crossovers that operate from a full range stereo signal are well
> understood (regardless of whether they operate in the analogue or
> digital domain) and need not cost a lot of money. The cost of the
> crossover in an active system (which requires one power amp per drive
> unit) is not that significant.
>
> Contemplating doing all this as some sort of SBS plugin and hoping to
> use multiple sync'd Squeezeboxes to deliver the split signals is
> frankly bonkers.
C'mon Clive, this is the *Audiophile* forum! ;)
R.
--
"Feed that ego and you starve the soul" - Colonel J.D. Wilkes
http://www.theshackshakers.com/
Contemplating doing all this as some sort of SBS plugin and hoping to use multiple sync'd Squeezeboxes to deliver the split signals is frankly bonkers.
I was trying to say the same but in other words. :)
I thought the aim was to do the XO server-side? - if not it's easy:
You need:
1 SB player
2 DEQ's
2 identical stereo amps
pair of 2-way speakers with xo removed
4 speaker cables
1 toslink lead from SB to DEQ 1 (configured as Low pass, connected to bass drivers via amp 1)
1 s/pdif lead from SB to DEQ 2 (configured as high pass, connected to treble drivers via amp 2)
No sync issues at all.
You can use the internal DEQ dacs, or external dacs
and of course you can still run Inguz for DRC server side!!!!
You're right: I was talking about doing it server side. And yes it is an unashamedly theoretical posting.
I can see that it can be done by 2 DEQs. In fact I think it may be achieved by one DCX 2496 (at a cost of about Ģ200) as suggested elsewhere.
I'm not usually this blunt, but someone has to say it - this is a CRAZY idea. People like Phil have been gently and politely trying to point out the futility of this scheme for a while, but it seems not to have got through.
Crossovers that operate from a full range stereo signal are well understood (regardless of whether they operate in the analogue or digital domain) and need not cost a lot of money. The cost of the crossover in an active system (which requires one power amp per drive unit) is not that significant.
Contemplating doing all this as some sort of SBS plugin and hoping to use multiple sync'd Squeezeboxes to deliver the split signals is frankly bonkers.
I will freely accept that the idea I was suggesting may not be practical given the existing architecture, but I can't see any sense in which it is bonkers. Anyway none of this was tied particularly to a particular implementation- it was based upon an idea of what would in principle be the best way of doing things.
The train of thought I was on was about the way that we are still tied to using boxes to do stuff that can be done server side. It offends my sense of elegance to be using 2 separate overlapping systems (Room EQ and electronic crossover) where one would do. It also strikes me that apart from the theoretical inelegance of this solution, there is an unnecessary connection involved and the possibility that the 2 systems may interact unhelpfully. Hmmm the Tact idea sounds increasingly attractive.
As I imagine it the room eq working on 4 channel would optimise the 4 outputs (or 6 for 3 way) in one go taking into account driver and room characteristics. Whilst it may be cheap to buy Linn's active crossovers for its own speakers, I dare say the speakers are not cheap. The imaginary solution I am suggesting does not require you to buy into a particular dac or amp or speaker. It is also a serious point that by and large active speakers are relatively rare in the domestic market and many people would not wish to be tied to the existing choice available.
Now if it doesn't work then yah boo sucks to me: but I really do think that the point is worth making that the power of a server based system to solve hitherto fiddly audio problems is not yet being tapped properly. Perhaps it never will be because basically not enough people care to make it worthwhile developing- but as regards cost effective audiphilia -wouldn't it be a wonderful thing if someone could buy just a double squeezebox [maybe a double dac (equivalent of MF M1 or an rDAC); maybe not] 2 cheapish power amps and 2 crossover less speaker boxes [no idea but maybe 1500-2000 the lot]. You would probably get better results than 99% of audiophiles currently get with systems costing 10k plus.
How is any of this bonkers on a forum where people spend ages agonising about whether they should be using asynch usb from a touch (NB a hack for which it was not designed) rather than s/PDIF? Or changing the PSU for a linear one costing more than the Squeezebox?or taking the screen of their machine, or discussing the relative merits of different capacitors? or putting rocks on things? The interesting point for thoughtful observers of the audiophile species is that people happily try to bodge what they have at enormous expense rather than rethinking from scratch
Believe me- I know that that the x-over can be done downstream but only at some expense of flexibility and/or potential signal degradation. This is one of the reasons why traditional 2 channel audiophiles have been suspicious of active systems. The idea that this could be done costlessly (in money and signal degradation)on the host server seems worth mentioning.
Now you can laugh but one of the reasons I started thinking about this was by noticing that inguz contains an Ambisonic UHJ decoder, apparently free and gratis. Ambisonics is a cautionary tale since it seems to have been a conceptually elegant solution to the problem of recording and delivering surround sound; but a commercial failure [Apologies Phil; as I am sure you know all this] As far as i can tell no one has impugned either the maths or the results in the real world. Be that as it may, it is striking that what previously required a separate and probably expensive ambisonic box [6000 plus from meridian maybe less elsewhere] can now be obtained for nothing.
Equally as soon as one starts thinking about room eq it becomes obvious that there is no hard and fast dividing line between room eq and speaker eq.
Anyway that is enough of a ramble. The bottom line is that if the platform will not support it then it can't be done: but it occurs to me that if logitech ever make a multichannel squeezebox it might just work.
Back to bonkers mansions for a cup of tea.
Phil Leigh
2011-03-29, 04:29
My (now 7 years old) Linn front speakers (+ amps+XO's) were circa Ģ6.5k all-in. You can spend MUCH MUCH more than this!
Yes you can use a DCX2496 instead of 2 DEQ's... (BTW my opinion of Behringer gear is that it is cheap and OK when it works... but I've had a fair bit of it that just died...)
The Linn XO's are relatively cheap because all the normal expensive bits (case, PSU) are not required, they are provided by the host amplifier(s). They are however, high-quality... and in an active system there is nowhere to hide when it comes to "the weakest link in the chain" :-)
Adamdea
I think your 'blue sky' approach to this is interesting. If it stimulates some discussion of a future SB product, or even a DIY hack of existing products, then it can only be a good thing.
The main criticism so far seems to be that the synchronisation of two squeezeboxes would not be tight enough for this to work, so I have a couple of questions along that line.
Which scenario would minimise the impact of slight synch errors - would it be better to have one squeezebox driving both bass drivers and one driving both tweeters, or would it be better to have one squeezebox for the left, and one for the right?
What stops two SBs from synching better? Is it a hardware issue that might be solvable if, say, two SB units shared some common (timing) hardware? Or is it something that could be improved in software if there was sufficient will?
Phil Leigh
2011-03-29, 05:04
Adamdea
I think your 'blue sky' approach to this is interesting. If it stimulates some discussion of a future SB product, or even a DIY hack of existing products, then it can only be a good thing.
The main criticism so far seems to be that the synchronisation of two squeezeboxes would not be tight enough for this to work, so I have a couple of questions along that line.
Which scenario would minimise the impact of slight synch errors - would it be better to have one squeezebox driving both bass drivers and one driving both tweeters, or would it be better to have one squeezebox for the left, and one for the right?
What stops two SBs from synching better? Is it a hardware issue that might be solvable if, say, two SB units shared some common (timing) hardware? Or is it something that could be improved in software if there was sufficient will?
Interesting question... using one SB for treble and another for bass, sync errors would manifest as changes in the relative driver time alignment, but with stereo image solid... whereas using one for left and another for right would create a wandering image with fixed driver alignment...
Much depends on the frequency and size of any sync errors...
I'd be intrigued to hear some real-life results!
Interesting question... using one SB for treble and another for bass, sync errors would manifest as changes in the relative driver time alignment, but with stereo image solid... whereas using one for left and another for right would create a wandering image with fixed driver alignment...
Exactly. And as far as I can see it would ONLY be a timing issue. The server could theoretically send whatever it likes to each channel of each SB, so left bass and left treble could certainly be sent to one SB. After that, appropriate amping could send the SB outputs to the correct driver units. So amps with appropriate characteristics could be used for each driver type. Such an approach would be scalable for multi-drive speakers too.
Exactly. And as far as I can see it would ONLY be a timing issue. The server could theoretically send whatever it likes to each channel of each SB, so left bass and left treble could certainly be sent to one SB. After that, appropriate amping could send the SB outputs to the correct driver units. So amps with appropriate characteristics could be used for each driver type. Such an approach would be scalable for multi-drive speakers too.
Yes timing is the issue but it is not trivial.
Assuming we only have stereo with two channels it makes more sense to have both channels for each type of drivers to one SB.
Phil Leigh
2011-03-29, 05:39
Exactly. And as far as I can see it would ONLY be a timing issue. The server could theoretically send whatever it likes to each channel of each SB, so left bass and left treble could certainly be sent to one SB. After that, appropriate amping could send the SB outputs to the correct driver units. So amps with appropriate characteristics could be used for each driver type. Such an approach would be scalable for multi-drive speakers too.
When you say "amps with appropriate characteristics could be used for each driver type" - the crucial thing here is that the amps have identical phase angle shifts and slew rates... in fact, they should be sonically identical amps, with only absolute current delivery maybe being different. So yes, you can get away with smaller transformers and reservoir caps for HF, but pretty much everything else needs to be the same. In general it's best IME to use identical amps (unless the amps have been expressly designed for - and into - the speakers...ATC, Meridian et al)
Also, there are choices regarding vertical vs horizontal vs diagonal amping... over the years different manufacturers have followed different paths on this, although once you start putting the amps INSIDE the speakers then vertical is the only option...
When you say "amps with appropriate characteristics could be used for each driver type" - the crucial thing here is that the amps have identical phase angle shifts and slew rates... in fact, they should be sonically identical amps, with only absolute current delivery maybe being different.
OK, I hadn't appreciated that. I had assumed that it might be possible to simply use a lower powered amp for the treble. I imagine that phase shifts between two SBs might anyway swamp any differences between two amps, but I'm only guessing here. Perhaps there could be a calibration process that could deal with both effects together. The beauty of what Adamdea is proposing is that the crossover characteristics and driver responses could be tweaked in the server software, and although it would inevitably be a very complex process it might have the potential to deal with sonic differences between amps.
Is the timing issue due to the way TC/IP streams are transmitted, or is it about getting the start command sent to the 2 boxes?
I can see that any difference has to be less than the sort of timing issues you are seeking to correct for in the first place.
Is the timing issue due to the way TC/IP streams are transmitted, or is it about getting the start command sent to the 2 boxes?
I can see that any difference has to be less than the sort of timing issues you are seeking to correct for in the first place.
It is related to keeping them in sync with micro second resolution over time. TCP/IP is part of the issue since it basically is not time oriented but order oriented. When you only have one stream, playing from a buffer solves that problem but with two or more streams to multiple devices are involved, you need to keep all these devices sync. If you can keep them in sync over time, you have also solved the start.
earwaxer9
2011-03-29, 09:25
This subject makes my cap/inductor "rolling" seem like kids stuff! I dont think I could stand the stress of it!
Is the timing issue due to the way TC/IP streams are transmitted, or is it about getting the start command sent to the 2 boxes?
I can see that any difference has to be less than the sort of timing issues you are seeking to correct for in the first place.
Even if the audio is adequately buffered, and the two streams start perfectly in sync, I guess the problem is the small differences in the rates of the clocks used to play out the data. I have only a rudimentary understanding of the syncing mechanism in SBS, but I understand that individual SBs send information back to the server that enables it to figure out the timing of the whole playback chain. But perhaps a hardware hack that forced two SBs to share a common playback clock would do the trick.
Even if the audio is adequately buffered, and the two streams start perfectly in sync, I guess the problem is the small differences in the rates of the clocks used to play out the data. I have only a rudimentary understanding of the syncing mechanism in SBS, but I understand that individual SBs send information back to the server that enables it to figure out the timing of the whole playback chain. But perhaps a hardware hack that forced two SBs to share a common playback clock would do the trick.
Yes. Which sort of amounts to the doubledecker squeezebox I suppose.
tcutting
2011-03-29, 10:51
Does anyone know what the target accuracy of the synchronization of two SBs is? Unfortunately, even if it's a few milliseconds, that probably kills this idea. (Quick calculation for what it's worth: at 2500hz, 1 cycle is only 0.4msec. A higher-frequency XO point, this number gets smaller).
The digital/active crossover seems much more feasible... the DSP gets full-range L & R information via digital interface with no timing errors. Could create digital XO's with arbitrary roll-offs, phase/amplitude corrections. Then have separate DACs feeding amp/speakers. Drawback would be no off-the-shelf products for performing amplitude/phase corrections. Advantage would be you should be able to equalize each XO output for actual amp/speaker combination, as well as delay and amplitude match each set to each other.
Does anyone know what the target accuracy of the synchronization of two SBs is? Unfortunately, even if it's a few milliseconds, that probably kills this idea.
You're probably right, if we're talking about two independent SBs. But what's the possibility that one of the mods used to slave an SB to a DAC clock could be adapted in order to slave it to another SB? Wouldn't that ensure perfect synchronisation?
tcutting
2011-03-29, 11:10
Has anyone looked at this company?
http://www.DEQX.com
They have products similar to the aforementioned Behringer, but you have fully digital I/O available (it looks like the Behringer ALWAYS needs or converts to analog...) They also provide the software for doing setup/correction.
Their products seem a bit pricier ($2000-$4000), and not sure what you actually get for the money. It looks like they have options with digital outputs, so you can run into your own favorite DACs (although you need at least 2 stereo DACs for 2-way speakers, more typical would probably be 3 stereo DACs for 2-ways plus subs). Of course you still need separate amps for each speaker.
Does anyone know what the target accuracy of the synchronization of two SBs is? Unfortunately, even if it's a few milliseconds, that probably kills this idea. (Quick calculation for what it's worth: at 2500hz, 1 cycle is only 0.4msec. A higher-frequency XO point, this number gets smaller).
It is not frequenccy related but speed related.
Sound travel at around 300m/s, you do the math...
As if the design idea wasnīt flawed enough by timing issues, you would also need to have 2 or more pre-amps/volume controls or use a multi channel pre-amp or similar to contrpol the volume. That is unless you want to do volume attentuation in the digital domain.
Given that the whole exercise is to improve sound, you donīt want to do it in the digital domain, meaning you would need to have 3 capable pre-amps or similar or a multi channel pre amp of some kind. Most, if not all, reasnobaly priced multi-channel pre-amps used ADC conversion of analouge inputs and for obvious reasons we donīt want that, which leaves us with a very few high-end multi-channel pre-amps that have true analog multi-channel paths. All of sudden, it gets very expensive already before solving the timing issues.
It would be like building an AWD driven car that instead of having one engine that drives all wheel through the transmission and a normal steering to control direction, had an engine for each of it wheels and using the speed of the wheels to control not only speed but also direction. The engines would be using carburators and old fashioned ignition and be controlled with mechanical wires individually. That would be possible but would neither be efficient, easy to control, accurate nor cheaper.
I forgot to mention one thing, weīre obviously talking about a car used for roads or tracks with lots of corners.
As if the design idea wasnīt flawed enough by timing issues, you would also need to have 2 or more pre-amps/volume controls or use a multi channel pre-amp or similar to contrpol the volume. That is unless you want to do volume attentuation in the digital domain.
Given that the whole exercise is to improve sound, you donīt want to do it in the digital domain, meaning you would need to have 3 capable pre-amps or similar or a multi channel pre amp of some kind. Most, if not all, reasnobaly priced multi-channel pre-amps used ADC conversion of analouge inputs and for obvious reasons we donīt want that, which leaves us with a very few high-end multi-channel pre-amps that have true analog multi-channel paths. All of sudden, it gets very expensive already before solving the timing issues.
It would be like building an AWD driven car that instead of having one engine that drives all wheel through the transmission and a normal steering to control direction, had an engine for each of it wheels and using the speed of the wheels to control not only speed but also direction. The engines would be using carburators and old fashioned ignition and be controlled with mecha
nical wires individually. That would be possible but would neither be efficient, easy to control, accurate nor cheaper.
I forgot to mention one thing, weīre obviously talking about a car used for roads or tracks with lots of corners. actually digital domain volume control is fine by me. Lets face it you wouldnt want to do this if you dont like dsp. Otherwise 2 passive preamps or just 2 integrateds may do . Not convinced this bit is major problem.
actually digital domain volume control is fine by me. Lets face it you wouldnt want to do this if you dont like dsp. Otherwise 2 passive preamps or just 2 integrateds may do . Not convinced this bit is major problem.
Manipulating the signal for room correction in SBS or in DSP does not necessarily mean you lose information but digital volume control will.
2 (or more if you have more than 2 way speakers) passive preamps, integrated, whatever analouge attenuation with good enough quality would do but they still would have to be matched level wise for it to make sense from a sound quality point of view and you still have to deal with the timing issue.
Given that you want to use the existing SBS and some version of SB players, does not make a lot of sense from software engineering point of view from neither cost nor sound quality perspective and especially not when trying to accomplish better sound at lower cost.
Phil Leigh
2011-03-29, 23:47
ALL EQ/DRC involves some element of gain reduction - it's necessary to avoid clipping when boosting and happens whether you like it or not when cutting.
Level attenuation in the digital domain is not a problem if done properly (ie in 24 or 32 bit) - and not excessively. After all it has already happened on all the CD's you bought and it didn't ruin their sound...
actually digital domain volume control is fine by me.
Manipulating the signal for room correction in SBS or in DSP does not necessarily mean you lose information but digital volume control will.
I'm perfectly happy with digital volume control too. I've heard all the arguments about losing signal while the noise level stays constant, therefore SNR suffers, but if the noise level is inaudible to start with, what does it matter if I'm losing data bits below that level?
I'm also happy that this setup will require another power amp. So from my point of view, it still comes back to JUST a timing issue. I'm not saying it's not a significant issue, but it seems it is the only one that has to be addressed (plus the plugin itself, obviously) to get this working.
It's not the only way to do this, and nobody is arguing that it's the best way. I think Adamdea's intention with this thread was just to have a discussion about whether it's possible. I suppose it's analogous to some of the other plugins out there - the iPlayer plugin might not be the BEST way to listen to BBC radio (if you've got a tuner in your rack), and the CD player plugin might not be the best way to listen to a CD (if you've got a hardware CD player in your rack), but they're all useful tools.
If this plugin existed, I would enjoy the possibility of fooling around with different crossover profiles, just to see what they do to the system.
Also, wouldn't you have the same volume control problem with any digital crossover- eg if we went touch -DCX2496 to 2 poweramps how would you be controlling volume?
and even with analogue crossovers, are they necessarily perfectly pair-matched?
Still the synching problem with the 2 SBs remains. You may notice from the Pissing Contest thread that this issue has apparently been ventilated before. if only someone really clever could think of an answer.
ALL EQ/DRC involves some element of gain reduction - it's necessary to avoid clipping when boosting and happens whether you like it or not when cutting.
Level attenuation in the digital domain is not a problem if done properly (ie in 24 or 32 bit) - and not excessively. After all it has already happened on all the CD's you bought and it didn't ruin their sound...
It doesnīt have to happen but it can happen. Besides in this case it is for the good cause and like volume attenuation it could possibly done at the 24 or 32 bit resolution.
As for digital level attenuation, it can be done so there are no hearable loss. However in the context where we use SB and SBS for xover and still keep the main architecture intact, the level attenuation is done in the SB, having, so in 24 bit. That gives a whole 12 dB to work with untl we start to have impact on 16 bit recordings. Since weīre discussing speakers without passive xover, we can also assume the speakers are easy to drive. letīs say 95db/w (simplified since one have more than one "speaker" in reality).
Even with moderate amplifiers, letīs say 64 W per amp channel, the listening volume would still be very very loud when you have used up all your window of 12 dB without resolution losses, everything at lower volume lever than that 101 dB will quickly start using bits lower than the bits over 16.
So in conlcusion digital volume control when using 24 bit will in reality affect the resolution so it can be heared unless one add extra attenutaion, have very weak amps or speakers that are hard to drive.
Also, wouldn't you have the same volume control problem with any digital crossover- eg if we went touch -DCX2496 to 2 poweramps how would you be controlling volume?
and even with analogue crossovers, are they necessarily perfectly pair-matched?
Yes and you could have similar issues with analogue xovers themselves, or power amps and speaker drivers for that matter but at least as long as there are just one volume control there are less sources of errors.
Yes and you could have similar issues with analogue xovers themselves, or power amps and speaker drivers for that matter but at least as long as there are just one volume control there are less sources of errors.
Surely the answer is that as with a transporter you trim the analog output level to roughly the right amount and then use digital attenuation mjust for the fine tuning.
I do this at home using a passive preamp and very rarely change its volume setting.
For the active system you could do this using single resistors on the line outs which can be pretty accurately matched, but 2 passive preamps or a single 4 way unit would be fine.
Perhaps it would be helpful if you could identify which active crossover system you think avoids this issue. For my part i think that if you are worried about 4 resistors not matching you ought to be far more worried about all those components in 2 separate 2 way crossovers.
Surely the answer is that as with a transporter you trim the analog output level to roughly the right amount and then use digital attenuation mjust for the fine tuning.
I do this at home using a passive preamp and very rarely change its volume setting.
For the active system you could do this using single resistors on the line outs which can be pretty accurately matched, but 2 passive preamps or a single 4 way unit would be fine.
Perhaps it would be helpful if you could identify which active crossover system you think avoids this issue. For my part i think that if you are worried about 4 resistors not matching you ought to be far more worried about all those components in 2 separate 2 way crossovers.
Are you seriously debating that having 4 volume controllers instead of one is not harder to get well-aligned? Besides, you donīt have 2 sources feeeding your setup at once do you?
BTW, having tow transporters is very far from your "cheap" solution isnīt it?
IMO, Linn and Meridian got it right and there are quite a few active speakers from other manufacturers like Dynaudio that works well too.
Phil Leigh
2011-03-30, 04:39
Just to be clear on this, if DSP DRC requires a 3dB boost (at any frequencies), it will reduce the overall level by 3dB first to ensure that there is enough headroom for the boost without inducing clipping.
If DRC determines a 3dB cut is required then it will achieve this cut by reducing the level at the relevant frequencies - that's what a "filter" does - frequency-dependant level reduction.
So, digital attenuation is a given with DRC.
In practical applications, DRC can require up to 10dB of boost/cut depending on how the speakers and room are sounding.
Digital XO's will also often require cut for one or more drivers, to compensate for the unequal frequency response of the drivers - if your passive XO has a resistor or l-pad in it (usually on the tweeter) this needs replicating with digital attenuation.
Of course a digital (or active) XO can do something that a passive can never do and that is boost the signal to a driver...but it can only do this if there is headroom and that usually means an across the board cut in level first...
Just to be clear on this, if DSP DRC requires a 3dB boost (at any frequencies), it will reduce the overall level by 3dB first to ensure that there is enough headroom for the boost without inducing clipping.
If DRC determines a 3dB cut is required then it will achieve this cut by reducing the level at the relevant frequencies - that's what a "filter" does - frequency-dependant level reduction.
So, digital attenuation is a given with DRC.
Just like you can increase the dynamic headroom by having the volume attenuation at 24 bit or 32 bin in DAC with digital volume control, this could be done in SW for DSP DRC meaning as long as your soruce material is lower than your headroom for the signed integer or whatever you are using. So if we have 16 bit music stream using a signed integer in a 32-bit system, there is enough headroom. If this was impossible, so would digtial attenuation without losses.
Are you seriously debating that having 4 volume controllers instead of one is not harder to get well-aligned? Besides, you donīt have 2 sources feeeding your setup at once do you?
BTW, having tow transporters is very far from your "cheap" solution isnīt it?
IMO, Linn and Meridian got it right and there are quite a few active speakers from other manufacturers like Dynaudio that works well too.
No, don't worry I was not debating that seriously or otherwise.
Perhaps I haven't made my point clearly. I am not arguing about the problem of synching to players just the suggestion that volume control considerations make the server side crossover suggestion unfeasible.
You are right to point out that somewhere somehow levels have to be made to match, and that this is a little trickier with active crossovers (especially non-proprietary ones). But how serious a problem is this?
It seems to me that there is a channel matching probelm in any stereo system. The pre-existing stereo volume control is in fact two volume controls L and R albeit that they may be synched physically.
Obviously (assuming digital vol control won't do all attentuation) you now have to match 4 levels not 2. MY point is that
1) getting the signal into the range where it can be harmlessly controlled by digital attenuation is simple and can probably be done accurately because you are just matching 4 resistors (on left high/ left low/ right high/ right low)
2) in any event any additional complexity (and pair matching problem) is nothing like as complex as the problem of getting passive crossovers to work and match.
3) How is this any different from what will have to occur in (or after)any digital crossover?
4) In Linn's case they have as far as I can see crossovers matched exactly to a particular amp and pair of speakers which is not the same (see 6).
5) In meridian's case I think they are using digital attenuation. I am not sure whether their solution works unless your system is all meridan.
6) I think the volume control problem is an inherent problem of any system where you try to get generic amp+ generic (x-over less) speaker + generic crossover and is not peculiar to the server side solution.
No, don't worry I was not debating that seriously or otherwise.
Perhaps I haven't made my point clearly. I am not arguing about the problem of synching to players just the suggestion that volume control considerations make the server side crossover suggestion unfeasible.
You are right to point out that somewhere somehow levels have to be made to match, and that this is a little trickier with active crossovers (especially non-proprietary ones). But how serious a problem is this?
It seems to me that there is a channel matching probelm in any stereo system. The pre-existing stereo volume control is in fact two volume controls L and R albeit that they may be synched physically.
Obviously (assuming digital vol control won't do all attentuation) you now have to match 4 levels not 2. MY point is that
1) getting the signal into the range where it can be harmlessly controlled by digital attenuation is simple and can probably be done accurately because you are just matching 4 resistors (on left high/ left low/ right high/ right low)
2) in any event any additional complexity (and pair matching problem) is nothing like as complex as the problem of getting passive crossovers to work and match.
3) How is this any different from what will have to occur in (or after)any digital crossover?
4) In Linn's case they have as far as I can see crossovers matched exactly to a particular amp and pair of speakers which is not the same (see 6).
5) In meridian's case I think they are using digital attenuation. I am not sure whether their solution works unless your system is all meridan.
6) I think the volume control problem is an inherent problem of any system where you try to get generic amp+ generic (x-over less) speaker + generic crossover and is not peculiar to the server side solution.
Fair enough and yes the volume control can be an issue with other designs too.
Forgot to mention B&O as an example of good active speakers.
Agreed. However, sigh, it does seem that the stream synching problem is going to be difficult to surmount.
Are there any very clever people out there with an idea.
Phil Leigh
2011-03-30, 10:28
No, don't worry I was not debating that seriously or otherwise.
Perhaps I haven't made my point clearly. I am not arguing about the problem of synching to players just the suggestion that volume control considerations make the server side crossover suggestion unfeasible.
You are right to point out that somewhere somehow levels have to be made to match, and that this is a little trickier with active crossovers (especially non-proprietary ones). But how serious a problem is this?
It seems to me that there is a channel matching probelm in any stereo system. The pre-existing stereo volume control is in fact two volume controls L and R albeit that they may be synched physically.
Obviously (assuming digital vol control won't do all attentuation) you now have to match 4 levels not 2. MY point is that
1) getting the signal into the range where it can be harmlessly controlled by digital attenuation is simple and can probably be done accurately because you are just matching 4 resistors (on left high/ left low/ right high/ right low)
2) in any event any additional complexity (and pair matching problem) is nothing like as complex as the problem of getting passive crossovers to work and match.
3) How is this any different from what will have to occur in (or after)any digital crossover?
4) In Linn's case they have as far as I can see crossovers matched exactly to a particular amp and pair of speakers which is not the same (see 6).
5) In meridian's case I think they are using digital attenuation. I am not sure whether their solution works unless your system is all meridan.
6) I think the volume control problem is an inherent problem of any system where you try to get generic amp+ generic (x-over less) speaker + generic crossover and is not peculiar to the server side solution.
Volume control of an active system is NORMALLY never a problem... it is done in the pre-amp that drives the XO (or in the source if you don't have a pre-amp)... in your proposed server-side system this would have to be done digitally with synced SB volume controls (This aspect is entirely possible today)
Deaf Cat
2011-03-30, 13:45
Bit behind on the tech talk side of things, but, thinking to avoid the sync problem of two SB's could the software not seperate the left low, right low, left high, and right hi, put them on a DD signal and then out of the SB digi out, in to a nice AV processor, using the front channels for the left and right low's and the rear channels for the left and right high's forwarded into power amps/stereo speakers?
One volume control on the av processor.
Just a thought,
I have an av processor acting as a stereo pre amp dac at the mo, channels spare ;-)
PS I have no idea as to size of signals, what would be needed, and what the SB can handle...
earwaxer9
2011-04-01, 07:56
2) in any event any additional complexity (and pair matching problem) is nothing like as complex as the problem of getting passive crossovers to work and match.
Not sure what you are referring to --- Good caps are usually within a couple of percent of each other. In a 1st order crossover there is just one of them per channel. Very easy.
I've been "bi-amping" for decades with a "band-quality" Furman tunable active crossover. With control over crossover frequency and amp volume out, it really is the ultimate "loudness button."
If you froogle.com for "active stereo crossover" you'll find good quality stuff (like Pyle) for less than US$100. Far cheaper than buying a second receiver.
If you have balanced ins/outs, your in business because they are supported. If not the only real inconvenience is you will have to buy RCA to 1/4" phone plug cables (3 sets) since "band" equipment eschews RCA connectors.
P
dizzysnakepilot
2011-04-01, 22:11
If I'm understanding this thread correctly- proposed is the idea of doing a software crossover and then independently amplifying the results? Sound promising.
I switched from a mono-amped Linn setup to a active bi-amped some years ago. I have two amplifiers which are modded with hardware cards specific to the particular model of speaker I have. Each amp drives one of the types of speaker cones in the speaker. It works great, and is different as well as better than the mono-amped system.
What the software version would be missing, it seems, is the correct and precise crossover profile that is matched exactly to each physical speaker cone in the speaker (as Linn can do since it's all a matched system)
MediaCenter
2011-04-03, 20:27
Sorry to be blunt but this makes no sense. You can buy miniDSP for $125 for a 2x4 (2 way) with PEQ, Delay etc. Going active is affordable now! lets focus our attention to something more important.
Sorry to be blunt but this makes no sense. You can buy miniDSP for $125 for a 2x4 (2 way) with PEQ, Delay etc. Going active is affordable now! lets focus our attention to something more important.
Looks very interesting, but I am not sure I follow how this works- you take the analog outs of your transporter send to the minidsp which converts to digital (sampled at 48kHz), does the magic and outputs analog to your amp? Now you may well be right that this is an excellent solution, but some people might regard it as a teenie bit of a compromise. Doesn't seem as smart as using somethig with digital input and analog out, let alone digtial in digital out (although you would need 4 channels of digital output and a 4 channel dac at least). Or is there a version of the mini dsp which is all digital?
On the other hand it may be the killer point in the argument that room eq and active crossovers are more important than Dacs.
Mini DSP can take the digital SPDiF output and process it. The limitation is the 48 kHz sampling rate. I even don't know if it can handle the 96 kHz signal . The advantage is you can make nearly any kind of filter or DSP with them .
Chris
Ps another interseting processor for crossovers and EQ is the Xilica 4080 ca 1500$
http://www.xilica.com/?c=78&cat=1&id=1
Phil Leigh
2011-04-04, 08:37
Mini DSP can take the digital SPDiF output and process it. The limitation is the 48 kHz sampling rate. I even don't know if it can handle the 96 kHz signal . The advantage is you can make nearly any kind of filter or DSP with them .
Chris
Ps another interseting processor for crossovers and EQ is the Xilica 4080 ca 1500$
http://www.xilica.com/?c=78&cat=1&id=1
If this miniDSP kit could do 24/96 in and feed multiple s/pdif outs configured as a 3-way crossover it would be very interesting for me in the future...
I'm looking at some alternatives in the event that Linn pull support for my amps and XO's...
MediaCenter
2011-04-07, 12:25
Mini DSP can take the digital SPDiF output and process it. The limitation is the 48 kHz sampling rate. I even don't know if it can handle the 96 kHz signal . The advantage is you can make nearly any kind of filter or DSP with them .
Chris
Ps another interseting processor for crossovers and EQ is the Xilica 4080 ca 1500$
http://www.xilica.com/?c=78&cat=1&id=1
Just a matter of time when they start offering 24/96 and higher DACs with crossover and PEQ it can be a great kit and at reasonable price point may bring many to an active setup. Still, cannot deny the convenience of DEQX with analog volume control but at a much higher price point.
Builder Brad
2011-09-19, 11:54
looks like the Mini Dsp can do the higher sample rates, switch inputs and do a 4-4 way active system.
http://www.minidsp.com/products/minidspkits/2-x-in-8-x-out
its a bit biggers board, I just got one to experiment with on a Linkwitz Orion inspired OB system.
be rude not to at that price.
Powered by vBulletin® Version 4.1.12 Copyright © 2013 vBulletin Solutions, Inc. All rights reserved.