I have been planning to use a DSP crossover with high quality DACs feeding one amplifier per driver in a 3 way active speaker. My first thoughts were whether to use a MiniSHARC or a NanoDIGI - anyone know the advantages of the SHARC?
Anyway, whilst looking for information that compare these two DSPs, I found myself stumbling over several articles that compare analog crossovers with their DSP counterparts. My target (like most people) is to get the most ACCURATE sound reproduction which would imply that the DSP solution is better. Several people point towards the downsides with a DSP solution (even with good DACs) but I just can't see why - is it just 'fear of the new' or are there aspects that genuinely affect the sound quality. For example, does 'perfect reproduction' sound too harsh or something like that?!!
Anyway, whilst looking for information that compare these two DSPs, I found myself stumbling over several articles that compare analog crossovers with their DSP counterparts. My target (like most people) is to get the most ACCURATE sound reproduction which would imply that the DSP solution is better. Several people point towards the downsides with a DSP solution (even with good DACs) but I just can't see why - is it just 'fear of the new' or are there aspects that genuinely affect the sound quality. For example, does 'perfect reproduction' sound too harsh or something like that?!!
There are a lot of audio"philes" who basically believe in magic and if your system doesn't use what ever their magic feather is, then it's inferior. They will tell you loudly that lots of essentially insignificant or barely detectable factors will provide dramatic and magical improvements, while ignoring actual issues.
Those are going to be the majority of the people telling you a dsp based crossover is bad.
However, like with anything, a poorly done solution wt a DSP will be inferior to a well executed passive crossover. A DSP will not magically make you an expert speaker builder either.
For a learning experience or if you just want to experiment, a DSP based solution is just infinitely superior to a passive crossover.
1) You can change things in real time
2) No soldering or desoldering of elements to make changes
3) Obviously no stockpiles of different inductors, capacitors and resistors in order to be able to experiment.
My impression from reading the website is that the MiniSHARC may be more intended for OEM solutions. Where as the nanoDIGI is more of a general purpose solution. They seem to have broadly similar capabilities from what I can tell.
The MiniSHARC has only digital inputs and outputs, unlike the nanoDIGI. It appears to be able to do a 4 in/8 out vs 2 in/8 out on the nanoDIGI.
Those are going to be the majority of the people telling you a dsp based crossover is bad.
However, like with anything, a poorly done solution wt a DSP will be inferior to a well executed passive crossover. A DSP will not magically make you an expert speaker builder either.
For a learning experience or if you just want to experiment, a DSP based solution is just infinitely superior to a passive crossover.
1) You can change things in real time
2) No soldering or desoldering of elements to make changes
3) Obviously no stockpiles of different inductors, capacitors and resistors in order to be able to experiment.
My impression from reading the website is that the MiniSHARC may be more intended for OEM solutions. Where as the nanoDIGI is more of a general purpose solution. They seem to have broadly similar capabilities from what I can tell.
The MiniSHARC has only digital inputs and outputs, unlike the nanoDIGI. It appears to be able to do a 4 in/8 out vs 2 in/8 out on the nanoDIGI.
With DSP, you will always pay a price of adding extra active analog circuitry to support both ends of the conversion process as well as any artifacts of the AD/DA conversions themselves. But it is true you can more easily tweak your parameters than by doing a purely analog xover 'right' that lacks these sonic problems.
One problem with near brick-wall digital rolloffs are interdriver timing nonlinearities at other than ideal listening positions that may add some pre and post resonant character around the xover region.
One problem with near brick-wall digital rolloffs are interdriver timing nonlinearities at other than ideal listening positions that may add some pre and post resonant character around the xover region.
Last edited:
Thanks. Maybe I should be clear in that the choice is between an active analogue crossover and a DSP (taken straight from a digital source like a CD with optical link) piped through a set of really good DACs with dejittered clock etc. You would think that the DSP solution would be perfect but I read some reports where the sound quality is claimed to be impaired - in some cases even without filters activated. Maybe it is just one of those things you need to experience yourself. It's just that a full blown DSP solution is quite expensive!!
This is true, but the question is are those issues clearly audible. With current technology, you basically need test equipment to detect the phenomena you are describing.
Essentially, in either case analog or digital, if you are hearing artifacts you've done something wrong.
Another digital route?
I can't be enthused enough about my recent change from analog crossover to DSP-based. I replaced an old op-amp based crossover with a PC running MPD, Ecasound, Sox, etc.
My goal was as close to zero cost as I could manage, and had this list of parts in hand:
old AMD-based PC, 2 gigs ram.
old Turtle Beach Montego ddl sound card
old 200 gig hard disk
All the software was free. Ubuntu linux, ecasound, MPD, Sox, all open source/free software.
I use the same disk files on another system, use either a laptop or tablet to run the headless system, and the sound is enough better my wife noticed it immediately. I'm sure it has to do with better control over the crossover than the best available resistors and caps (5% tolerance isn't all that great)...
I do like the ability to change and test with immediate feedback. Change the config file and restart MPD. Takes seconds.
Oh, you want a better DAC? Sure. Plug in pretty much anything you like, via USB, or other. Ubuntu has drivers for most of what's out there.
How much PC do you need? Well this old single-core system typically runs at 5% use. You could run this on pretty much anything. Yeah, solid state disk would be nice, because it's quiet. Fanless operation would be nice, too. Some day.
I can't be enthused enough about my recent change from analog crossover to DSP-based. I replaced an old op-amp based crossover with a PC running MPD, Ecasound, Sox, etc.
My goal was as close to zero cost as I could manage, and had this list of parts in hand:
old AMD-based PC, 2 gigs ram.
old Turtle Beach Montego ddl sound card
old 200 gig hard disk
All the software was free. Ubuntu linux, ecasound, MPD, Sox, all open source/free software.
I use the same disk files on another system, use either a laptop or tablet to run the headless system, and the sound is enough better my wife noticed it immediately. I'm sure it has to do with better control over the crossover than the best available resistors and caps (5% tolerance isn't all that great)...
I do like the ability to change and test with immediate feedback. Change the config file and restart MPD. Takes seconds.
Oh, you want a better DAC? Sure. Plug in pretty much anything you like, via USB, or other. Ubuntu has drivers for most of what's out there.
How much PC do you need? Well this old single-core system typically runs at 5% use. You could run this on pretty much anything. Yeah, solid state disk would be nice, because it's quiet. Fanless operation would be nice, too. Some day.
The nanoDIGI is an IIR based DSP which basically can do in the digital domain what we can do in the analogue domain with filters and such.
The MiniSHARC, however, can also do FIR which also lets us correct phase and make the speaker linear phase.
Technically DSP is superior to analogue circuitry as long as you use enough precision. The MiniSHARC is good enough unless you want to filter way below 20 hz where you lots and lots of bits but for use in the audio range it is enough. Even experts like Bruno Putzeys and Linkwitz have gone over to DSP. Bruno has also jumped on the FIR train too with his Grimm Audio LS1.
There are also "hidden" advantages of DSP for us mortals which is that prototyping and such is criminally easy. You don't have to bother as much with box designs but can rather just build it and as long something isn't wrong or bad you can fix it in the end and get the response you want.
Addition: If you use DSP crossover then don't put a DAC in front of it, use the DACs after the DSP crossover.
The MiniSHARC, however, can also do FIR which also lets us correct phase and make the speaker linear phase.
Technically DSP is superior to analogue circuitry as long as you use enough precision. The MiniSHARC is good enough unless you want to filter way below 20 hz where you lots and lots of bits but for use in the audio range it is enough. Even experts like Bruno Putzeys and Linkwitz have gone over to DSP. Bruno has also jumped on the FIR train too with his Grimm Audio LS1.
There are also "hidden" advantages of DSP for us mortals which is that prototyping and such is criminally easy. You don't have to bother as much with box designs but can rather just build it and as long something isn't wrong or bad you can fix it in the end and get the response you want.
Addition: If you use DSP crossover then don't put a DAC in front of it, use the DACs after the DSP crossover.
My advice:
I am using DBX driverack PA and it definitely degrades the sound quality compared to a good analogue (passive or active crossover). I am sure, I compared. It is like comparing the sound of a good Wav file to the Mp3. Sound gets harsher, simplified, soundstage flattens. Just a "dead" sound.
Now a good DSP crossover, like Lake LM26 or XTA is much better than DBX. I compared them. Is it as good sound as a good analogue, like pass labs? Don't know, did not compare.
Keep in mind that those good DSP crossovers are expensive $3-$4k. Not sure if it is worth it but I AM sure that the DBX or Behringer are only good enough for prototyping projects, unless it is a very modified Behringer. Those may be good but I have not tried it again because of the $1800 price tag.
I am using DBX driverack PA and it definitely degrades the sound quality compared to a good analogue (passive or active crossover). I am sure, I compared. It is like comparing the sound of a good Wav file to the Mp3. Sound gets harsher, simplified, soundstage flattens. Just a "dead" sound.
Now a good DSP crossover, like Lake LM26 or XTA is much better than DBX. I compared them. Is it as good sound as a good analogue, like pass labs? Don't know, did not compare.
Keep in mind that those good DSP crossovers are expensive $3-$4k. Not sure if it is worth it but I AM sure that the DBX or Behringer are only good enough for prototyping projects, unless it is a very modified Behringer. Those may be good but I have not tried it again because of the $1800 price tag.
I'm a bit of a DSP doubting Thomas.
Going active is definitely worth the effort. Ive just started my first active implementation, in analogue.
In my opinion, cheap DSP is just cheap. My doubts with DSP are that i believe i would need a Good device, read expensive. This is based on absolutely no evidence, circumstantial or otherwise.
All i will say is that my opamp crossover was cheap, built in a day and sounds great.
De-soldering a couple of components to tweak is hardly bothersome. If it was so much effort, why DIY at all....
And 0.1% resistors and 1% caps are hardly difficult to find.
Going active is definitely worth the effort. Ive just started my first active implementation, in analogue.
In my opinion, cheap DSP is just cheap. My doubts with DSP are that i believe i would need a Good device, read expensive. This is based on absolutely no evidence, circumstantial or otherwise.
All i will say is that my opamp crossover was cheap, built in a day and sounds great.
De-soldering a couple of components to tweak is hardly bothersome. If it was so much effort, why DIY at all....
And 0.1% resistors and 1% caps are hardly difficult to find.
Last edited:
The digital of DSP is quite cheap, and as long as you have enough bits for precision you get equal performance. If you know how to do analogue crossovers then sure the only real advantage over analogue is that you can go FIR and correct the phase.
The main gain of DSP is that it is so much easier and faster to set up, especially if you don't want textbook electrical slopes but rather want say 24 dB acoustical slopes instead.
The dbx DriveRack PA is doing A/D conversion, then processing, then D/A conversion. Its convertors are notoriously cheap. The dbx DriveRack 260 does much the same processing, but with better sonic results. Using either of these systems (or equivalents), you can still implement very good crossovers, and get way more flexibility than building in the analog domain.
But if you can avoid the initial A/D conversion (using digital sources), half the sonic problem goes away. And if you also choose your own DACs for the downstream D/A conversion, even better.
I'm using NanoDigi (TOSLINK input, S/PDIF output) with some fairly cheap DACs and it sounds excellent.
But if you can avoid the initial A/D conversion (using digital sources), half the sonic problem goes away. And if you also choose your own DACs for the downstream D/A conversion, even better.
I'm using NanoDigi (TOSLINK input, S/PDIF output) with some fairly cheap DACs and it sounds excellent.
Some very good points have been made here in favour of DSP crossovers. I love rthe concept of active crossovers; more so DSP-based, although my current loudspeaker is passive. I was until recently using a Najda DSP crossover, and will do so again for my next speaker. The concept for this DSP crossover was born here on diyAudio forums. Do have a look:
WAF-Audio
http://www.diyaudio.com/forums/digital-line-level/215379-dsp-xover-project-part-2-a.html
WAF-Audio
http://www.diyaudio.com/forums/digital-line-level/215379-dsp-xover-project-part-2-a.html
Ground Sound from denmark have an excellent digital filter with advance futures, have used my self with non problems . a excellent product if you are an advance builder.
One thing that got me thinking is the difference that phase linearisation (is that a word??) has on the flatness of the frequency response. I would therefore assume that different listening positions (and even more so, speaker locations) have a similar consequence. The advantage with a dsp setup is that this can be quickly adjusted. Try doing that with an analogue crossover!
This combination has always puzzled me. The MiniDSP 4x10 has good DACs already built in (6 output and 2 input channels) and has SPDIF digital input plus volume control and four selectable presets that can use either the digital or analog inputs. It's almost like a preamp. I find this to be an excellent combination for my needs.
This seems to me to be missing the point. Assuming plentiful precision, the bottleneck will always be the DACs (assuming a digital source with sufficient bits). DSP precision is getting more affordable so isn't an issue whereas DACs haven't received much development since ESS.
Phase linearization doesn't affect the frequency response that much except that it removes pre-ringing in the crossover regions. But unless you have steep filters that isn't that much of an issue. The main advantage is that it removes group delay so the transient time response of the speaker is far far better.
Last edited:
As an owner of an analogue already, I'd buy the miniDsp... then if you don't like it, find someone wanting to trade. I am sure there are plenty of us out there... On my analogue crossover, it's impossible to tell what the actual setting is, the notches aren't 100% accurate. I hate knowing that. If you can live with it or have measurement software, you may see things differently. The miniDSP does everything you need it to do for next to nothing, and its abilities will likely far outweigh the drawbacks. That said, I have never heard one myself, so cannot attest to whether or not I could live with that either, but am open to trying it.
I listen to 99.9% vinyl, but would risk the DAC if it meant making my speaker system much more coherent and balanced for minimal cost.
I listen to 99.9% vinyl, but would risk the DAC if it meant making my speaker system much more coherent and balanced for minimal cost.
Have you found a good fix for the on-off "POP"? That seems to be the one big flaw with these units.