Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[6233638]

Just a theoretical question that's been going around my head the last few days. (started looking at speakers again...)
I only ever plan on having a stereo system, and ideally I think it would be a pair of bookshelf speakers and subwoofers.
 
But it occurred to me that if you were wanting the best possible quality from your system, it might actually be better to drive the tweeters, woofers, and subwoofers completely independently with a multichannel AVR rather than, say, a stereo power amplifier.
 
That way you eliminate whatever crossover design is built into the speakers, and could potentially have better integration between all three units via DSP - especially the subwoofers if they're not located right near the speakers.
 

Would there actually be much of a benefit to this approach, or is it just something that would be theoretically better, but in practice, not really worth the hassle?

[mwillems]

Just a theoretical question that's been going around my head the last few days. (started looking at speakers again...)
I only ever plan on having a stereo system, and ideally I think it would be a pair of bookshelf speakers and subwoofers.
 
But it occurred to me that if you were wanting the best possible quality from your system, it might actually be better to drive the tweeters, woofers, and subwoofers completely independently with a multichannel AVR rather than, say, a stereo power amplifier.
 
That way you eliminate whatever crossover design is built into the speakers, and could potentially have better integration between all three units via DSP - especially the subwoofers if they're not located right near the speakers.
 

Would there actually be much of a benefit to this approach, or is it just something that would be theoretically better, but in practice, not really worth the hassle?

What you're describing is called bi-amping (or tri-amping), and there are serious and real benefits to using an active crossover (whether digitally or in analog like Linkwitz does).  To name just a few:

1) Digital crossovers and filters can be significantly more predictable than traditional passive crossovers (the frequency response of the system is less likely to vary with changes in ambient heat for example)

2) Decreased intermodulation distortion because the amp is reproducing less of the frequency band.

3) Less amplifier power is needed both because each element is amped separately, but also because passive level matching typically requires throwing away a good bit of driver efficiency that can be recouped in digital filters.

4) Much better damping as the amp is coupled directly to the elements (instead of through several components)

5) Delay is very hard to get right in a passive crossover (although it's possible).  Without adequate delay, proper crossover tuning is a "near fit" at best.  With active crossovers delay can be dialed in exactly.

6) Drastically better control of system EQ if your elements have a non-linear response (all elements do to some extent).  With a passive crossover, it's very difficult to EQ anything near the crossover region because both drivers may not be exhibiting non-linearity, but in a passive system you can only EQ both, etc.

If you're willing to put in the time on system tuning, bi-amping a speaker can give truly excellent results. My own system is bi-amped and I would never even think about going back.

Rod Elliot used to have a really excellent article on the subject up, but his site's been down lately, in case it comes back up, the link is sound.westhost.com/bi-amp.htm

[6233638]

That basically sounds like the type of benefits I was imagining from driving the speakers this way.
You say that this is bi-amping/tri-amping - does that mean you should be using separate amplifiers (e.g. three power amplifiers) or simply that a multichannel amplifier is powering each driver independently? (I suppose there are technically six amplifier circuits in there?)
 
Really glad that I decided I'd never want anything more than stereo, and bought that nice Stereo DAC now.
Well, I shouldn't complain - I have use for more than one high quality zone anyway, but I would probably have done things differently if I had known about this in advance.

[mwillems]

That basically sounds like the type of benefits I was imagining from driving the speakers this way.
You say that this is bi-amping/tri-amping - does that mean you should be using separate amplifiers (e.g. three power amplifiers) or simply that a multichannel amplifier is powering each driver independently? (I suppose there are technically six amplifier circuits in there?)

I use discrete power amps, but you could use a multichannel amplifier as long as it supports the channel routing you'd need.  Some receivers don't actually provide user-configurable inputs for all output channels, so you'd have no way to map them, but some receivers do, and multichannel power amps typically do.   One of the reasons that I use discrete power amps is because my high frequency element is about 20dB more efficient than my low frequency elements.  So I prefer to use a high quality low wattage amp for the HF, and a much more powerful amp for the LF. That's one of the advantages of bi-amping: you can mix and match amplifiers if you want to line them up better with your specific needs/speakers. 

But the bottom line is that a separate amp channel is driving each element directly, and all filtering and signal shaping is happening before the amplifier stage.  It doesn't matter if the amp channels are in one box or two boxes.

Really glad that I decided I'd never want anything more than stereo, and bought that nice Stereo DAC now.
Well, I shouldn't complain - I have use for more than one high quality zone anyway, but I would probably have done things differently if I had known about this in advance.

I was in a similar boat, I know how you feel.  You've got a Benchmark DAC right?  As I recall, some of the Benchmark models will slave themselves to the digital output of another DAC, so you'd still need another DAC, but could potentially still take advantage of the nice one you already have.

For what it's worth, I use a Steinberg UR824, and it's excellent.  Mojave uses one as well, if I recall.  It's eight channel, so more than you'd need, but it would give you room to grow if you ever decide to do Tri-amping+Subs  

[6233638]

I use discrete power amps, but you could use a multichannel amplifier as long as it supports the channel routing you'd need.  Some receivers don't actually provide user-configurable inputs for all output channels, so you'd have no way to map them, but some receivers do, and multichannel power amps typically do.   One of the reasons that I use discrete power amps is because my high frequency element is about 20dB more efficient than my low frequency elements.  So I prefer to use a high quality low wattage amp for the HF, and a much more powerful amp for the LF. That's one of the advantages of bi-amping: you can mix and match amplifiers if you want to line them up better with your specific needs/speakers.

I thought it would probably be easier to just have one integrated unit, but that does sound like a good approach. I assumed with Media Center handling playback, it wouldn't matter what each channel is assigned as.

I was in a similar boat, I know how you feel.  You've got a Benchmark DAC right?  As I recall, some of the Benchmark models will slave themselves to the digital output of another DAC, so you'd still need another DAC, but could potentially still take advantage of the nice one you already have.

According to Benchmark, all units are phase-accurate, so you can simply add more of them if you want multichannel.
The Mytek DACs do let you have one DAC act as a master and two other DACs as slaves for a multichannel setup.
But buying another two is completely out of the question - that was a big purchase for me, and why I'm thinking AVR now rather than power amps. (Benchmark announcing a power amplifier is what got me started looking at a new speaker setup though)

For what it's worth, I use a Steinberg UR824, and it's excellent.  Mojave uses one as well, if I recall.  It's eight channel, so more than you'd need, but it would give you room to grow if you ever decide to do Tri-amping+Subs  

Thanks for the recommendation, I'll keep it in mind. I'm likely not going to have anything larger than bookshelf speakers, so I should only need six channels. (but an extra two could probably come in handy)
And really, if you have subwoofers, do you need a speaker with multiple mid-range drivers?
 
 
HiFi really is just an endless money pit, isn't it? There's never a point where you couldn't continue to spend money on improvements.

[mwillems]

I thought it would probably be easier to just have one integrated unit, but that does sound like a good approach. I assumed with Media Center handling playback, it wouldn't matter what each channel is assigned as.

If multi-channel receiver has a multi-channel digital input and you like the DAC in the receiver you should be fine.  The problem comes if you want to use a DAC outside the receiver; many receivers don't have analog inputs for all channels (but some do), which is what I meant above by a potential lack of "user-configurable inputs."  Obviously power amps mostly don't care what you feed them and have analog inputs for each channel.

Thanks for the recommendation, I'll keep it in mind. I'm likely not going to have anything larger than bookshelf speakers, so I should only need six channels. (but an extra two could probably come in handy)
And really, if you have subwoofers, do you need a speaker with multiple mid-range drivers?

It depends on the drivers and your goals.  Some highly regarded speakers are four-way set ups, and I've seen several 3-way speaker setups that still need a sub.  Especially with the kind of steep crossover filters that can be easily created through digital filtering, four-way becomes increasingly attractive (although it can be done in analog, either through an analog active crossover, or creative passive crossovers).  There are about ten octaves we can hear, and if you want flat, well-behaved frequency response with uniform directivity it's hard to find speaker elements that do more than four octaves well (they exist of course, but you'll pay for them). One moderately famous three-way design is Linkwitz's Orion, but depending on which version, it still needs a sub for the lowest bass extension.

Just to be clear, and to make sure I'm not misleading you: if you have more than one matched element outputting in the same range, there isn't much benefit to amping them separately.  For example, I run a 2x2 midbass array in each speaker for my low frequency stage and all four of the woofers on one side are on one amp channel because of the efficiency gains involved; but they're all reproducing the exact same signal and there's no analog components in between the amp and the elements so I still gain the benefits described above.  In a normal bookshelf speaker, though, each element is operating in a different range and so would need a separate amp channel.  The idea is that each frequency band is amped separately for each "channel" (i.e. HF left, HF right, LF left, and LF right each have a separate amp channel).
 

HiFi really is just an endless money pit, isn't it? There's never a point where you couldn't continue to spend money on improvements.

More or less, although I think there is a point of diminishing returns, and only you know when you've reached it.  For my part, I've managed to push my system distortion below -40dB at normal listening levels, and my frequency response is flat +/- 3 dB across eight octaves of the audio band. I've reached the WAF limit on room treatments, which is the only really the meaningful thing left to do (it would help level out those last ripples and the two octaves on the bottom).  At this point I'm down to replacing my HTPC fans to eliminate the ambient noise from the PC.  So I'm not sure I'm going to spend much more for the foreseeable future (fingers crossed that some new idea doesn't occur to me   ).

One piece of advice I'd give you before you start down this road in earnest:  if you don't have one, get a decent condenser microphone, whether one of the USB ones Parts Express sells, or a phantom powered one like the Behringer ECM line (just make sure you have a sound interface that will supply phantom power first).  You can probably find one for between 30 and 70 dollars.  Once you've got a mic, do some system measurements using free software like Holm Impulse or REW and puzzle over those for a while, learn measurement technique, etc.  It will sensitize you to what the "problems" are with your current system, and will demonstrate some of the benefits you might or might not be able to achieve by bi-amping, etc.  It will also equip you to actually do the bi-amping if you want to, as it requires lots of measurement to get right.  

If this is all something you're already doing, forgive me for preaching to the choir, but if not, it's a relatively low-intensity and low-investment way to "break in" to this side of DIY, and will be time well spent, if, for no other reason, it will tell you what kind of EQ your current system may need (or what kind of EQ "sounds flat" to you, i.e. "ear calibration").  If you have any questions about measurement technique or anomalous results, let me know, I've spent entirely too much time engaged in measurement and tweaking over the past few years (which is inevitable when you home-build speakers).  

[Listener]

That basically sounds like the type of benefits I was imagining from driving the speakers this way.
You say that this is bi-amping/tri-amping - does that mean you should be using separate amplifiers (e.g. three power amplifiers) or simply that a multichannel amplifier is powering each driver independently? (I suppose there are technically six amplifier circuits in there?)
 
Really glad that I decided I'd never want anything more than stereo, and bought that nice Stereo DAC now.
Well, I shouldn't complain - I have use for more than one high quality zone anyway, but I would probably have done things differently if I had known about this in advance.

In your OP, you talked about avoiding the crossover in the speakers.  Bi-amping or tri-amping won't do that.  Each amp will still see part of the crossover.

To get the crossover in the speakers completely out of the path, you need an active crossover that splits the signal before it goes to the two or three amps for that channel (and no crossover in the speakers).  That usually requires speakers and an active crossover designed to work together.

I have used Waveform Mach 17 three way speakers with their own active crossover for about 13 years. I like them very well.

Lots of pro audio studio monitors have an active crossover and amps built in.  Very cost effective compared to high-end audio gear.

I'd be a bit cautious about using multiple DACs.  Even if the delay within the DAC is consistent from unit to unit, the application software and driver software may introduce variable delays between channels.

Google and you'll find discussions of the merits of active crossovers and active speakers that include amps.

Bill

[mwillems]

In your OP, you talked about avoiding the crossover in the speakers.  Bi-amping or tri-amping won't do that.  Each amp will still see part of the crossover.

To get the crossover in the speakers completely out of the path, you need an active crossover that splits the signal before it goes to the two or three amps for that channel (and no crossover in the speakers).  That usually requires speakers and an active crossover designed to work together.

I have used Waveform Mach 17 three way speakers with their own active crossover for about 13 years. I like them very well.

Lots of pro audio studio monitors have an active crossover and amps built in.  Very cost effective compared to high-end audio gear.

I'd be a bit cautious about using multiple DACs.  Even if the delay within the DAC is consistent from unit to unit, the application software and driver software may introduce variable delays between channels.

Google and you'll find discussions of the merits of active crossovers and active speakers that include amps.

Bill

Bi-amping is splitting the signal before it gets to the amps and cutting out the passive crossover in the speaker.  If you just wire the existing passive crossover to two amps you're bi-wiring, not bi-amping.  See http://en.wikipedia.org/wiki/Bi-amping_and_tri-amping :"benefits of bi-amping cannot be realized if passive crossover networks of a speaker system are not removed" etc.

Regardless, I think he's imagining using JRiver as his active crossover and cutting out his passive crossover entirely.  I can confirm that JRiver makes a pretty great active crossover replacement if you have a multi-channel DAC, and would second your point about using several different DACs.  Syncing can be hard if the hardware doesn't support it.

And as you say, there are also some excellent active monitors that are effectively pre-bi-amped (or bi-amp-ready) from the factory and/or give you the option of biamping or not.  

[mojave]

And really, if you have subwoofers, do you need a speaker with multiple mid-range drivers?

It depends on your room size and volume requirements. I'm helping a friend with his new home theater. It will be run by an HTPC with JRiver and feature 15 channels. The mains/center speakers have dual 12" drivers and the -3dB of the speaker at around 60 Hz. We will be crossing over to 18" midbass units for each main/center at 150 Hz with those handling down to around 40 Hz. The two subwoofers each have dual 18" drivers in a sealed cabinet. The system will play from 5Hz to 24kHz with peak capability of about 130 dB.

the application software and driver software may introduce variable delays between channels.

The software has no way of affecting the clock of the hardware. Any DAC's clocked from the same master clock will be in sync.

[Listener]

Bi-amping is splitting the signal before it gets to the amps and cutting out the passive crossover in the speaker.  If you just wire the existing passive crossover to two amps you're bi-wiring, not bi-amping.  See http://en.wikipedia.org/wiki/Bi-amping_and_tri-amping :"benefits of bi-amping cannot be realized if passive crossover networks of a speaker system are not removed" etc.

Regardless, I think he's imagining using JRiver as his active crossover and cutting out his passive crossover entirely.  I can confirm that JRiver makes a pretty great active crossover replacement if you have a multi-channel DAC, and would second your point about using several different DACs.  Syncing can be hard if the hardware doesn't support it.

And as you say, there are also some excellent active monitors that are effectively pre-bi-amped (or bi-amp-ready) from the factory and/or give you the option of biamping or not.  

I responded to your post because I felt that you provided misleading information.  Your use of the term bi-amping (and that in the wikipedia reference) does not match the general usage that I have seen in the audio world.

For decades, many audiophile speakers have provided two sets of terminals for connection to amplifiers.  One set connected to the low pass side of the crossover and the low frequency driver. The other set of terminals connected to the high-pass side of the crossover and the high frequency driver(s).  Providing three sets of terminals ha sbeen less common.  The manufacturer provided metal straps that connected the two sets of terminals so that the speakers could be used with a single pair of speaker cables.

Audiophiles could remove the strap and run separate pairs of speaker cables to each set of terminals.  At the amplifier end, you could connect the speaker cables to the same amplifier.  This has been called "bi-wiring". 

If you added more amplifier channels, you could go farther and connect separate amplifier channels to each set of speaker cables.  This has been called "bi-amping".  Each amp channel seens one part of the crossover.

That is the usage of "bi-amping" that I've seen for many, many ears.

---
Removing a passive crossover from existing speakers and implementing an active crossover in the PC will involve some work and may not produce a better result.  The speaker manufacturer knows more about the performance of the raw drivers, the cabinet and the filler material than an individual user can.  The manufacturer probably spent more time and effort on optimizing the crossover (and had better test facilities) than a user has.

---
Multi-channel receivers may not be that good at delivering (full) power to all channels at the same time.  Another thing to be cautious about.

Bill

[mwillems]

I responded to your post because I felt that you provided misleading information.  Your use of the term bi-amping (and that in the wikipedia reference) does not match the general usage that I have seen in the audio world.

For decades, many audiophile speakers have provided two sets of terminals for connection to amplifiers.  One set connected to the low pass side of the crossover and the low frequency driver. The other set of terminals connected to the high-pass side of the crossover and the high frequency driver(s).  Providing three sets of terminals has been less common.  The manufacturer provided metal straps that connected the two sets of terminals so that the speakers could be used with a single pair of speaker cables.

Audiophiles could remove the strap and run separate pairs of speaker cables to each set of terminals.  At the amplifier end, you could connect the speaker cables to the same amplifier.  This has been called "bi-wiring".  

If you added more amplifier channels, you could go farther and connect separate amplifier channels to each set of speaker cables.  This has been called "bi-amping".  Each amp channel seens one part of the crossover.

That is the usage of "bi-amping" that I've seen for many, many ears.

I've never seen that usage before, but I believe you that it's extant.  I was using the definition Rod Elliot offered in his paper "Bi-amplification, not quite magic but close"  here (his site is back up again!): http://sound.westhost.com/bi-amp.htm

He says: "The most common question I get is ... 'Do I need to disconnect the passive crossover in my speakers?'  The answer is ... Yes, otherwise you are not really biamping at all," His paper does contemplate that a user might "go active" for part of the setup but retain a passive crossover for part of it too, but generally his definition of bi-amping is using an active crossover to split the signal before the amps.  

But the definition isn't important as long as we're all clear on what we mean here.  To clarify, what I meant by bi-amping (or tri-amping) above was doing all signal shaping prior to the amplification stage, and amplifying each frequency band of each speaker separately.  

---
Multi-channel receivers may not be that good at delivering (full) power to all channels at the same time.  Another thing to be cautious about.

Bill

I agree completely.

---
Removing a passive crossover from existing speakers and implementing an active crossover in the PC will involve some work and may not produce a better result.  The speaker manufacturer knows more about the performance of the raw drivers, the cabinet and the filler material than an individual user can.  The manufacturer probably spent more time and effort on optimizing the crossover (and had better test facilities) than a user has.

This is a very good point, and I agree that bi-amping requires a lot of work and may not produce a better result.  But with time and application I think it's likely to produce a better result in many systems.  The manufacturer spent a great deal of time with the passive crossover, no doubt, but they were building to a price and certain types of filters are impractical (or impossible) in analog at speaker levels in the space available to them.   Bi-amping with digital filtering is more flexible and can produce better results because it isn't constrained by the cost of parts and the physical limitations of those parts, and because it can solve problems easily that are comparatively hard to solve by using a passive crossover (damping, precise delay tuning, etc.).  But you're right that it's labor intensive and can produce worse results (in part by virtue of its flexibility), especially if the labor is neglected.  That's part of why I stressed the importance of getting used to speaker measurement above.

More from Rod's article:

I have seen many claims that loudspeaker manufacturers often go to extraordinary lengths to design the best possible crossover network for their products. I do not doubt that for many high-end systems, this is certainly the case. It must also be considered how much extra this costs, and we can be assured that many systems have a less than ideal network, simply to keep costs reasonable. Several times, I have seen reviews and speaker crossover schematics where expensive speakers use ferrite cored inductors for the low frequencies, and bipolar electrolytics are also common.

I do not consider these to be optimal or appropriate for a high quality system, and nor do many others. The truth is that cost considerations are nearly always made in any system, and much more so when the selling price becomes a consideration.

As I stated at the beginning, if you spend $25,000 or more for a pair of speakers, then we are into the "cost no object" area. Most people cannot afford such luxuries, and as a result they must settle for something they can afford. Only a very few systems will be as good as they can be, and you will pay dearly for it.

Biamping is not a simple tweak, and is not to be taken lightly. Make no mistake though, its application will improve almost any loudspeaker available, with very few exceptions.

[Ninouchka]

I had the chance to try out the new Yamaha AVENTAGE CX-A5000 and MX-A5000.
Balanced connections. I used my powered pro audio Yamaha HS80M,not too big, has two great
flat amps inside, added xlr connections to use them without crossovers and use the crossovers
from the cxa preamp outs. Lots of options with the 11 pre outs. This for trying out the preamp.

For testing the mxa power amp (11 balanced xlr inputs), I connected my Monitor AudioRX8's, hooked on a pc with
Avid firewire mBox Pro 3 card and a Steinberg MR816X.

Wow, I was realy amazed by the sound and price.
I just am thinking of an Aurora 16 with AES, instead of my firewire cards for my home theater room,
but this is in the same price range..

This new line of Yamaha is great for bi or tri amping.

I tried the MXA and the CXA together in a treated room, with a Cambridge Audio BR universal player,
via HDMI. with MA RX8 and the surrounds, the quality of the DACs inside is superb.
I think I know what to ask for Christmas. These two babies and a Lynx aurora 8,because I
need something with really low latency to record guitar and vocals.
I sell my recording studio and move everything to my home theater and compose/record in the nice
seats in front of a big screen.

A question : do you know the real time latency of the steinberg UR external audio card?
I need to be able to record 96 kHz/24 bit at least at 64 samples, at 7 channels.
The aurora does it, but is way higher in price than the steinberg  UR.

thanks

[6233638]

Lots of pro audio studio monitors have an active crossover and amps built in.  Very cost effective compared to high-end audio gear.

That is something I had also been thinking about, especially if all you need to do is put a nice DAC in front of them. The issue that I saw with that, was that if you're doing room correction, it's still only working on a stereo signal rather than isolating each driver.

The biggest problem with Studio Monitors for me though, is that most of them are designed to be functional units and are typically rather ugly.

Still, it's one thing to have a pair of speakers that are corrected, but I still think I'd want a pair of subwoofers in the room, and those would likely not be corrected to work together. (unless they were sold as a set)

I'd be a bit cautious about using multiple DACs.  Even if the delay within the DAC is consistent from unit to unit, the application software and driver software may introduce variable delays between channels.

Oh, I don't plan on going down that route, I know that if you are using multiple DACs you can run into problems with clock drift. (though Benchmark claim it's not an issue with their DAC2 units, and some DACs like the Myteks let you slave the DAC clocks off a "master" DAC)

[Listener]

That is something I had also been thinking about, especially if all you need to do is put a nice DAC in front of them. The issue that I saw with that, was that if you're doing room correction, it's still only working on a stereo signal rather than isolating each driver.

Some studio monitors have provisions for measuring response and doing DSP based room response correction in the speakers.  For example, JBL's 4300 and 6000 series and Genelec's 8200 series.  Such monitors often provide an SPDIF or AES digital input so no DAC is needed.  Of course, they may be limited to 24/96 i nput and provide no support for DSD.

The biggest problem with Studio Monitors for me though, is that most of them are designed to be functional units and are typically rather ugly.

Beauty is in the eye of the beholder.  The large Wilson audio speakers can't be beat for sheer ugliness in my opinion.  (Cylons?) The alien lifeform shapes of Vivid speakers are hard to top either.

Functional units for $ 1500 - 3500 versus elegant shapes and furniture finishes for $ 10,000 plus (for audiophile speakers + amps).  I lean toward value for money these days.

Still, it's one thing to have a pair of speakers that are corrected, but I still think I'd want a pair of subwoofers in the room, and those would likely not be corrected to work together. (unless they were sold as a set)

I'd agree that subwoofer frequency ranges are the most important for room correction.

The JBL and Genelec systems integrate their subwoofers in the measurement and room correction.  They support multiple subwoofers as well.

However, why can't you do room correction in MC before the subwoofer signal is split out?  For a single subwoofer, that should be pretty effective unless you have a lot of overlap between main speakers and subwoofer.

Bill

[mojave]

A question : do you know the real time latency of the steinberg UR external audio card?
I need to be able to record 96 kHz/24 bit at least at 64 samples, at 7 channels.
The aurora does it, but is way higher in price than the steinberg  UR.

thanks

Gearslutz.com has a latency database thread. The Steinberg UR28M is listed, but the UR824 should have identical latency. It is among the highest in/out latency. However, this only matters with live playback. The recommended maximum latency for not detecting any lip sync issues between video and audio is 22ms. The in/out latency of the UR series is lower than that.

[Ninouchka]

22ms is too high for me as guitarists and also for drumpads.
It has to be below 10, the ur22 doesn't score well, but I hoped that the bigger one would be better.

[daveca]

I only ever plan on having a stereo system, and ideally I think it would be a pair of bookshelf speakers and subwoofers.
 

From an EE who is designing such a system at the moment.... thats called a "flawed initial assumption"

System design always starts at the output. If the output is a cheap set of headphones with a range of 200 Hz to 4.5 KHz, its useless to design an amplifier with a max frequency of 40 KHz.

That will be about the result of building such a high quality system with bookshelf speakers.

Bookshelf speakers are totally oppsite of the goal. Bookshelf speakers may/will have crossovers inside (or a single speaker) and most of the point of this kind of design is to eliminate that.

Bookshelf units tend to be lower powered and far from ideal sound quality; if not they would be floor standing speakers.

A "good" set of speakers is $ 5-10K and I appreciate where youre coming from, my goal is to build such a system, including high quality speakers for less than $3000.
That is not easy...

Since the goal is to eliminate crossovers and have several channels, then the FIRST place to start is not the DAC, its speaker (driver) selection.

Ive spent three DAYS, on-line, selecting drivers for such a system, its not a simple task. Choose them first then work backwards into channel amps.

 Im looking at 6 speakers, low, mid and tweet that are around $1000-1200, not including enclosures. Surprisingly, the equipment with better specs is not the most expensive.

Start by selecting drivers you can afford and looking at their frequency graphs, such as this:

http://www.madisoundspeakerstore.com/approx-12-subwoofers/sb-acoustics-sb34swnrx-s75-12-woofer/

The graph is the key...notice the very poor frequency response, its not flat. The useful passband for that speaker is from 50- 90 HERTZ
Look at teh flatness of the curve, the width of the flat region and how much change in dB (decibel, unit of sound pressure) there is over that range.

Think of it as "dollar per hertz.."

[JimH]

Welcome to Interact.  Why don't you post your speaker candidates?  I might make a couple myself.

[daveca]

Here or new thread?

[astromo]

You could start fresh and in your first post, include a link from here to provide continuity / linkage. Include a link in this post to your new thread.

Not that I'm about to build a set of speakers. I'm with Jim, I'd be interested to follow what's involved.

[daveca]

AS in:

http://yabb.jriver.com/interact/index.php?topic=86405.0

?

[6233638]

From an EE who is designing such a system at the moment.... thats called a "flawed initial assumption"

System design always starts at the output. If the output is a cheap set of headphones with a range of 200 Hz to 4.5 KHz, its useless to design an amplifier with a max frequency of 40 KHz.

That will be about the result of building such a high quality system with bookshelf speakers.

Bookshelf speakers are totally oppsite of the goal. Bookshelf speakers may/will have crossovers inside (or a single speaker) and most of the point of this kind of design is to eliminate that.

Bookshelf units tend to be lower powered and far from ideal sound quality; if not they would be floor standing speakers.

Well, the reason I started out "backwards" with this, is because the DAC is something I have been able to take advantage of now in my current setup. I had also been thinking that it would be driving a much better speaker setup at some point in the future, which is why I chose that model.
If I had thought that I'd need something multichannel to drive stereo speakers, I would probably have bought a lower-end DAC/amplifier for my headphones.
 
However, I then started looking at speakers, and noticed that a number of the higher-end designs (often $10K and up) had eliminated the passive crossover and were now driving each component independently.
As much as I would like to, I don't think there's any possibility of me being able to justify spending that kind of money on speakers.
 
This is what got me thinking that you could probably do something similar with room correction software, and a multichannel amplifier/AVR if you ripped out the passive crossover. (or were able to order speakers without one)
 
 
DIY is something I had considered, but I know nothing about cabinet design, and I don't really have access to the space or tools that would be required to do a professional job.
 
I would be considering bookshelf speakers and not floorstanders for aesthetic reasons rather than performance reasons.
I know that it doesn't matter for the majority of people that frequent AV sites, but looks are just as important as performance for me.

I used to have a really nice set of JBL bookshelf speakers (honestly, they were probably too big to really fit in the "bookshelf" category - but they were not floorstanders either) which I got rid of in my last move because they were simply too big for the space I'm in now, and I just didn't like how they looked.
 
 
I do still wonder if it's going to be worth the trouble to have a setup like this though. It seems like it adds a lot of complexity and other dependencies (like only being able to be driven from the PC) compared to adding a really nice stereo amplifier to go with my DAC (Benchmark's new amplifier with almost 130dB dynamic range seems very impressive) and buying a nice pair of bookshelf speakers.

[daveca]

"
This is what got me thinking that you could probably do something similar with room correction software, and a multichannel amplifier/AVR if you ripped out the passive crossover."


Exactly, youre on the right track. Set the drivers up, test the accoustics with a software tone generator. See what the result is, maybe an EQ can correct it. Tone generator software, a mic back into the PC sound card and a simple PC oscilloscope program. I dont know much about cabinets either...

Since you already have the DAC thats fine, but just avoid spending too much on high quality amps for bass and midrange, if possible, because even cheap amps can drive those at such low frequencies.
Spending $50 each on bass and midrange amps saves hundreds of dollars on amps that can be spent on drivers.

[daveca]

"adding a really nice stereo amplifier to go with my DAC (Benchmark's new amplifier with almost 130dB dynamic range seems very impressive) and buying a nice pair of bookshelf speakers."

The bookshelf speakers totally invalidate all the previous improvements. Useless.
Why have multiple channels (six) just to combine them, in the end, into a relatively cheap set of speakers?

Theres no such thing in the high quality audiophile realm of good bookshelf speakers, they are a huge compromise in quality for less expense and space.