Help with PC based loudspeaker system, ground up.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
First off, I want to preface with the fact that I'm a bit embarrassed that I'm starting this thread because I can tell that there is a huge amount of knowledge already posted on these forums and I know that with enough searching I could probably find what I'm looking for. The problem I'm facing is that I'm so new to the world of home audio, I really don't even know the correct terminology to point me in the correct direction.

I have a fairly discerning ear and would like to put a quality "starter" loudspeaker/amplifier/DSP set together myself. This would need to be a "source to ear" setup, with the source being my PC. I'm only looking for two "towers" and a subwoofer or two with the required amplifiers and signal processors/crossovers. I don't have a specific budget in mind yet but I know that any advice you may give me without one is rather useless so I'll peg my initial investment at ~$2000. I'm very new to this and have no problem building speakers, but I think I want to minimize any electronics work this go around. I also want this to be something that if I decide to expand later, I don't have to toss components. I really hate buying the same thing twice because I bought a poor version of it.

My desire is for the system to be extremely accurate in reproducing sound as it was recorded, and hopefully a quality "sound stage" They will end up being used for music, movies, and games; however, accurate music reproduction is the most important piece to me.

I started looking into dipole/open baffle speakers as I frankly just didn't understand them, having previously understood the options of either sealed or ported enclosures only. I find them intriguing and may want to pursue an open baffle speaker design. Siegfried Linkwitz has some interesting arguments for that setup on his website, and I think if I were to purchase high quality speakers, even if I can't make the open baffle work system work for me, I could always put them in larger enclosures as well.


Here are my questions though:

A) If I'm starting with nothing (no receiver, EQ, speakers, etc) How much of the signal processing could I do digitally? I hate soldering, and all of my sources are digital and from my computer, so I would think it should be possible to use the computer to fully separate and control all signals out to a simple amplifier, or set of amplifiers which could then just go directly to each speaker. If not with my motherboards 7.1 outputs, at least with an add on soundcard I would hope. Ideally, I would build zero crossovers, and yet still be able to fully control the signal level output so that all I need to do is place amps in between the computer and the individual speakers.

B) I know I'm going to end up tinkering... So I'd like to know what cost effective tools will give me the ability to really fine tune the system? For example, I'd ideally like to be able to really fine tune frequency response based on the actual listening environment, and I'll probably do it several dozen times as I tinker with baffle designs and crossover parameters. Having software that utilizes a microphone to help me do this would be ideal. Any suggestions?

(just found this software Bodzio Software which may help me with both... perhaps.)

C) If I can't fully separate outputs natively, could I use the optical SPDIF output from my computer and some other singular device to do the same? For example, could I purchase an "all in one" receiver that will allow me to use the computer to set up active crossovers...

EDIT: It seems like www.minidsp.com has at least the DSP options covered here, with the SPDIF input and a variety of options for outputs... The price is a bit steep, at $500 for the 4in10out in a box version, but this really does seem like it would do everything I might want as an external solution...

D) Actual Speaker selection/ building etc: I'm somewhat drawn to the strange and, in my opinion, counterintuitive design of the linkwitz LX521/Orion and would like to perhaps put my own "mini" version of it together, however probably as a 3 way, or even two way with separate sub. I may want to upgrade to the real deal at a later date, and from what I've read, I believe you can't really go wrong with well matched SEAS speakers. I haven't really spent time researching whether the speakers spec'd for open baffle designs would also work well in enclosures if I decide to go that route later.

Are there other SEAS based open baffle DIY designs out there that I can reference?


Thank you for your time, and I'm sorry for the long post. I appreciate it greatly that you made it this far.
 
Last edited:
Ideas...

Hi, I will watch this thread with interest. I am an advocate for good and cheap, and also that a claim can be defended at least somwhat rationally. As in money, religion and politics, there is an enormous amount of plain old ******** in the audio hobby. At what point does science (acostics, electronics, etc) end and subjective opinion, perhaps with a little bit of false pride and self deception ("i have gear recommended by an authority, it must be superior.") thrown in? Only you can say what is right for you. I shall be glad to argue my points and provide links if you ask. Good luck and as the cattle breeder said, "buy the bull, not the ********!"
 
I re-read your original post. You want to build your own speakers, am I correct? Aside from "toy" projects and refinishing Bose 901s, that is above my pay grade! I am pretty low-end already, but I think you will find it a challenge to assemble a basic system including speakers for < $2000 (US) IF YOU BUY NEW. The age-old advice is put the bulk of your money on the speakers, and I agree. If you can settle on a DIY project, you can potentially save the enormous mark-up of most hi-fi bought new. I too have used (mainly) a Windows PC as my source for years. Even if you could use the onboard DAC (good luck! even I have my limits !!!), you still need an amp. Unless you are going for high efficiency, you need a few hundred watts per channel available. Certainly for a sub. I am a fan of Behringer ("pro" stuff, reasonably priced.) Even so, a DEQ2496 processor and INuke3000 would cost you about $700 new. The DEQ2496 will take analog or digital inputs. There is your processing in the digital domain. I'm sure all this can be done in PC software now, but I have zero experience there. You may still need an outboard DAC (even a cheap one is an improvement over a hissing and buzzing mini-line-out on the PC). Or a USB/Toslink converter? A DAC taking a USB input? You'll be lucky to have $1100 left for buying speakers AND sub. Even my used Bose 901's are worth maybe $400 the pair (with restorations). If you can DIY you can probably make wonderful speakers and a sub for $1100 in parts. Happy hunting on selecting a project, there must be hundreds of them. If I were starting from zero with $2000 US, I would probably do something like this:

PC (music source) -- free, also cost of music service or collection not factored in.
DAC: a USB capable model such as a DAC01a (<$200).
Amp: Behringer NU3000DSP: PC controllable Parametric and other EQs, more power than any reasonable person will ever need, $269 delivered. (Con: extra ADC/DAC steps, noisy fan many people replace with a quiet one; I just took mine out and it works ok unless pushed really hard.). I'd still have $1500. Based on my lack of skills, I'd try to find a quality used deal on Ebay or similar. The speakers are really the hardest thing, because common sense (notably absent in this hobby!) dictates that this is the greatest influence (along with the room) upon the sound you finally hear. When you get right down to it, most of the electroncs specs pale in comparision to the mangling the signal takes when it enters a voice coil, even in the best units :)
Speakers:
 
Last edited:
Thanks for the comments.

I guess a bit more background may be helpful. I used to be really into car audio and understand basic speaker parameters, but it's been a while and I've forgotten a lot of the details. I'm also an mechE and pretty competent with electronics as well, however I really don't enjoy building circuits, so I would like to avoid that. Just in general I learn fairly quickly. I've put less than 8 hours of internet reading about HT speakers to date, and will be doing much much more research before I make any purchasing decisions.

As a random side note, my dad has owned Bose 901s (two versions of the basically identical speaker) for nearly 30 years, those are the speakers I grew up listening to.

First off, I have no problem buying used for most things. I think I would almost definitely purchase amplifiers used as I see no additional value in that equipment being new.

Sreten,
Thanks for the link, looks interesting. I think that Nao note with the miniDSP as suggested in it's description would be solid option. MiniDSP actually offers their own microphone as well so I may be able to use their software for the tuning side. From a pricing side, The drivers, a minidsp 4x10, software, wood etc should all come out to ~1500$ with out amplifiers.

This seems like a pretty solid option, and then the only question left is do I build it exactly as described, or do I end up going with the seas components...


My motherboard itself has an S/PDIF optical output, which I would assume should work for my needs. Any good links for legit improvements on the computer output side (Motherboard is Gigabyte UD7TH)
 
More questions: will you be content with 2.1 (stereo + sub) or must you have home theater (5.1 or more)? HT will likely cost you 2-3x as much in speakers alone. With an engineering background, you may want the freedom and flexibility of components vs. an integrated HT receiver. The Ultimate EQ product is interesting, and while I did not read it in depth, seems it would allow ultimate flexibility for tweakers and set-up. Still, you need at least a DAC and and amp for each channel. These can be had as cheaply or as expensively as you choose. Even worse if you are actively EQ-ing all drivers!!! There are so many options, it really does help to do some research and ask yourself what your requirements are first. If you like HT I recommend

Home Theater Forum - Home Theater Systems - HomeTheaterShack

The "owners" of the free and excellent Room EQ Wizard software. There is a DIY section but I still like here the best.

As for "legit" improvements on the digital side, well you are a trained engineer I assume, and in any case you can wander into the marshy thicket of subjective opinion. Good luck whacking your way through jitter arguments, asynchronous, and a dozen other things. Do any of these impact the final sound you'll hear? Even if a difference can be measured with equipment, I say that in many cases an honest test (almost never happens!) would reveal no audible difference between (say) a $100 DAC and a $1000 DAC. Decide for yourself. And may your $2000 be wisely spent.

Here is my favorite audio skeptic:

Why We Believe

For the skeptical and budget conscious, there is (for instance) the $100 Ebay DAC. For the well-heeled gullible audio-fool, there is an endless array of audio components ranging in price from the expensive to the astronomical, such as audio purifiers, magic cables, isolation feet, speaker cones made out of unicorn hide, tweeters made with the tweets of rare birds, and capacitors handmade by gnomes who live in an ancient Oak tree in the Black Forest :)
 
More questions: will you be content with 2.1 (stereo + sub)

I have no interest in more than 2.1 (or 2.2 maximum)

With an engineering background, you may want the freedom and flexibility of components vs. an integrated HT receiver. The Ultimate EQ product is interesting, and while I did not read it in depth, seems it would allow ultimate flexibility for tweakers and set-up. Still, you need at least a DAC and and amp for each channel. These can be had as cheaply or as expensively as you choose. Even worse if you are actively EQ-ing all drivers!!!

I think the MiniDSP serves as the only DAC I would need, as well as the fulfilling the activEQ/crossovers roll. I also like the fact that it seems infinitely useful, for a wide range of potential future applications. It also seems to integrate with the room EQ wizard software and an offers a microphone which seems like it would give me the ability to properly adjust the sound.


The "owners" of the free and excellent Room EQ Wizard software. There is a DIY section but I still like here the best.

As for "legit" improvements on the digital side, well you are a trained engineer I assume, and in any case you can wander into the marshy thicket of subjective opinion. Good luck whacking your way through jitter arguments, asynchronous, and a dozen other things. Do any of these impact the final sound you'll hear? Even if a difference can be measured with equipment, I say that in many cases an honest test (almost never happens!) would reveal no audible difference between (say) a $100 DAC and a $1000 DAC. Decide for yourself. And may your $2000 be wisely spent.

Here is my favorite audio skeptic:

Why We Believe

For the skeptical and budget conscious, there is (for instance) the $100 Ebay DAC. For the well-heeled gullible audio-fool, there is an endless array of audio components ranging in price from the expensive to the astronomical, such as audio purifiers, magic cables, isolation feet, speaker cones made out of unicorn hide, tweeters made with the tweets of rare birds, and capacitors handmade by gnomes who live in an ancient Oak tree in the Black Forest :)

I'm well aware of the ridiculousness that is high end audio products, which is why I'm here as I assumed most of the DIYers would avoid all of that. I don't understand the "jitter" or pretty much any reason why the S/PDIF optical out on my motherboard wouldn't serve my purposes and I think I will plan on using that without any additional sound cards.

I think the NaO Note II RS might be the right answer for me, but I may have to do something about the lower frequencies. I don't particularly to add an additional subwoofer to the existing towers, so I may consider using different speakers for that.
 
The NaO speaker looks like a good value (it is DIY right?). You seem on the right track -- spend most of the funds on a good speaker system. I am a MiniDSP fan, I have their basic 2x4 and have used it easily for an EQ and x-over. Do share about whatever you decide upon! This forum is great; they tolerate everything from half-assed projects like my glue-together tube woofer to expertly fabricated systems done in CAD by engineers with access to high-tech processes to make their unicorn hide speakers with laser-cut military grade aerospace stainless steel structures (kidding, but only slightly).

Addendum: if you are not doing theater, and are considering an old-school ("stereo") do you even need a sub? I've owned two, and very little music goes down that low. Like many other things, you can add or delete one to your system as your tastes change. When you're selling that expensive sub on Ebay for half what you paid for it, you will sadly remember my sage advice :)
 
Last edited:
Hi,

Like I said don't try anything of your own imagination, like different drivers.
Build it exactly as described, forget about trying to cover future bases.

NaO Note II RS Details

Read this page and contact M&D about the details about basing
it on a PC music Server with the Bodzio UE and PC based DSP.

It fits your criteria to an absolute T. Simply do not start fretting
about the bass, in all real HiFi's it is the least of your worries.
It has to be icing on a very good cake, and it simply is not
shabby in the first place. Subs should not be needed for
most quality music, but by definition if you read the
above, really low AV LFE will not be reproduced.

rgds, sreten.
 
Last edited:
So then here is my next question... Why is it so hard to find a brand that produces quality amplifiers for a somewhat reasonable price?

The only brand I've seen that comes close would be Emotiva (at least by recommendations)

I'm thinking that a 6 channel amp (60 or 80watt x 6) would handle the upper three channels, and then something around ~150x2 for the 4 subwoofers.

Any other recommendations?
 
Hi,

I estimate you would need near 200W into 4 ohms for the bass,
near 100W into 8 ohms for midbass, near 50W into 8 ohms
for the midrange and near 30W into 4 or 8 ohms treble.

Welcome to the wonderful world of amplifiers, where cost
usually makes little sense with hifi brands, and reasonably
priced "studio" stuff is severely sneered at as being too cheap.

Going full 4 way active clearly creates a big problem in amplifier
costs, and having a cost effective philosophy for amplifiers.

FWIW I'd use "studio" amplifiers for the low bass and mid bass,
the former about twice the power of the latter, probably new,
and then carefully choose the midrange and treble used.

The other option is building and here I'd halve all the power
numbers above. Overdoing building is pointlessly expensive.

rgds, sreten.
 
I have gone down the ultra-cheap route myself:
- PC is the source and crossover
- 7.1 channel sound card gives Nx analogue outputs
- Three stereo amps off eBay (Denon) for three way speakers
- Two high quality caps for tweeter protection
- Old speaker enclosures 're-purposed' with new baffles (sealed, 12" woofers in 90 litre cabinets, 4" mid and 1" tweeter in sealed 7 litre bookshelf cabinets sitting on top)
 
Things to look out for:

There is an issue that it is not always possible for a single sound card to act as the destination for your audio apps, source of audio for the crossover software and multi-channel output all at the same time. It is hard to tell whether this is going to be possible before buying a particular type of card. I had it working for a while using a Creative X-Fi card and Win XP, but found that in Windows 7 this functionality had been removed from the X-Fi drivers! Remaining options are to use two sound cards or a software based 'virtual audio cable'. I have opted for the former because the software-based solution involves re-sampling I believe. I am using an M Audio Delta card as the destination for audio apps and source for the crossover software, and a Creative X-Fi (PCI version with drive bay connector panel) for the multichannel outputs. These are both available secondhand for not much money. The crucial point to note is that the two cards' sample clocks must be locked together, and with the two cards I mention this is as easy as linking the SPDIF output of the X-Fi to the SPDIF input of the M Audio with a phono cable. You can then set the M Audio as the default audio device in Windows control panel so that all your audio goes to it. Your crossover software uses the M Audio as its source and the X-Fi as its destination for the 4, 6 or 8 output channels. (Probably as easy as selecting them from drop-down lists).

Once set up and the crossover software is running (you can make it run automatically at startup), the system is 'transparent' in that all your PC's audio comes out of your active speakers without any other messing about.

A 'gotcha' that I have encountered is that some 'professional' sound cards seem to be very cavalier with regards to glitches (an understatement!) at power on and off, and I think could certainly damage the drivers as well as being bloody irritating. I have seen some people advocating turning the amps on only after booting the PC etc. but you really don't need that hassle, and the power to the PC could die at any time anyway. Again, I have to report that the Creative X-Fi is exemplary in this regard, and I simply don't worry about what order things get turned on in.
 
Assuming the PC side is taken care of (it may only involve plugging in a couple of cards as described above, installing drivers and setting up a few things in menus - but is it ever as easy as that?) all you have to worry about is designing and building your speakers(!).

I have to report that, unless I was very lucky, there's nothing to it. Really, given the control of the crossover that you're going to have, you can optimise any combination of drivers, and provided you don't try anything too outlandish, it's going to be fine. I mean, if you're just combining a woofer, mid and tweeter in a box, like thousands of similar designs before you, how far wrong can you go? In my case I stuck to sealed enclosures (ports are a bit too clever for their own good IMO) and fitted low cost polypropylene woofer & mid, and silk dome tweeter (£25, £9 and £15 each as I recall, special offers and surplus stock etc.). I fancied trying big woofers, so went with 12" and fitted them in large boxes as suggested for those drivers by the free program WinISD.

Now you could fret about lobing and beaming and directionality. You could worry about breakup modes, and kid yourself that you're an artist and selecting particular drivers to blend together perfectly. But basically the physical sizes of the drivers determine most of that stuff anyway, so you're going to get what you're going to get. In my case I went the measurements route (which can be a bit involved) and was able to correct the phase and delay for each driver individually, but I seriously expect that it should be possible to get fairly close by ear. (I didn't go in for room correction, however). It's something you can develop once you're underway anyway. A measurement mic and pre-amplifier is probably <£100.

I gradually homed in on a setup that sounds perfect, finding that really subtle changes are audible. 'Baffle step correction' is a frequency response curve designed to counteract the change in directionality of a driver mounted on a baffle that occurs with frequency. The required correction varies with speaker positioning relative to walls etc. and I found that a change in gain of 1dB in this is clearly audible. It is this sort of thing that suggests to me that passive speakers are always a shot in the dark, and where the cult of the speaker designer as tortured artist comes from. For the active speaker builder it's all very straightforward.

I can report that active speakers only sip power compared to their passive ancestors, so I really don't think you need massive amplifiers. All three of mine are around the 40-50W mark.

But here's the thing: the sound is distinctly 'high end'. They're certainly the best speakers I've ever heard. If we're talking stereo imaging: superb. If we want musicality: beautiful, rich sound (no hint of 'clinicality', no unpleasant edginess). Listening fatigue: none, unlike any passive speaker I've owned. And they'll go loud without anything unpleasant happening - and because they're so clean you tend to play them louder (at 'realistic levels' as they say) without realising it, which adds to their general astounding-ness.
 
Last edited:
Also, forgot to mention that a really simple way to get going with all this is to convert some existing sacrificial passive speakers to active. In fact, if the speakers have ports you can do it without permanently modifying them at all: pass your cables in through the ports and connect directly to the drivers using the small crimp spades that they usually use. Do put caps in series with the tweeters to protect against low frequencies and DC (I'm using 80uF film). I started out converting some floor standing Mission two-ways like this, with spectacular improvements, but really wanted to try sealed enclosures and large drivers hence building my own - but even then only replacing the baffles on existing donor enclosures.
 
Thank you very much for the information..

I'm still trying to figure out what is going to be "worth it" and I know that until I have some constants, and have listened to some systems, that's going to be an extremely difficult question to answer.

I think my initial plan will be to use the onboard sound (which has decent specs) and play around with some of the free software options.

I think I'll eventually move to the minidsp, but in the mean time I want to build some mini-dipole speakers for next to the computer (Like http://gainphile.blogspot.de/2009/09/s11-ob-bookshelf-open-baffle-with-seas.html, and pick up a par of NHT bookshelfs to compare them against. I can always use a quality speaker, and I've heard good things about these for the cost.

Then make a decision as to whether I want to do the NaOs, or the LX521s... Anyway, thanks for all of the feedback.
 
And always remember Beran eck's Law (I think it is) that basically says that if you design a speaker system, it will of course sound better to you than all other speaker systems. So take other people's ravings with a grain of salt, trust your ears, and good luck leaving your ego out of the equation :)
 
And always remember Beran eck's Law (I think it is) that basically says that if you design a speaker system, it will of course sound better to you than all other speaker systems. So take other people's ravings with a grain of salt, trust your ears, and good luck leaving your ego out of the equation :)

In my ravings am I claiming special abilities? No, the opposite. merely that the clever thing is to have chosen to go active in the first place and that success is almost assured. I think the myth that speaker design is difficult deserves to be busted! There are many amateur and commercial speakers that are complete disasters but they, of course, have been developed around a 1930s configuration that precludes 'iteration' and cannot fully compensate for positioning in the room - and they sap power from the amp, have poor cone damping etc. The active designer (DSP especially) can hardly go wrong if he uses a bit of rigour in measurements, but may even achieve success by using just his ears. IMO.
 
Hi drivemusicnow. The difficulty with my software is that it's homebrew and a bit rough round the edges (the 'GUI' wasn't designed but just happened!), but I'd be happy to send you a copy to see if you could get it to work on your system. Its big advantage is that it allows you to fiddle with the settings in real time. (As, presumably you can with a Behringer..?)

Looking at the software, the essentials are:
- set up sample rate, audio devices
- select two, three or four way
- set crossover frequencies and slopes (N-1 frequency/slope settings for an N-way system)

At this stage you could have a working active speaker, setting your frequencies/slopes to something 'about right' and adjusting the levels on your amps to give you equal levels from both drivers at each crossover. However, I would suggest that it wouldn't sound great. The following refinements are very helpful:

- baffle step correction to counteract the relative loss of bass energy from a speaker in open space. This depends on baffle width and speaker placement in the room and can be an essential refinement, but you can start with something sensible and adjust it by ear quite happily.

- delay correction to compensate for driver offsets relative to each other e.g. the acoustic centre of the woofer may be further back than the tweeter. Also, the necessary delay may depend on the height of the listener's ears relative to the speakers.

-phase correction based on measurements compensates for each driver's gradual phase lag that occurs at higher frequencies. This is a good thing anyway, but ensures that adjacent drivers are in phase at the crossover, thus avoiding a frequency response 'suckout'.

- amplitude correction based on measurements. Possibly not necessary as the drivers may be pretty flat anyway, but you may wish to refine the crossover slopes to give an accurate slope incorporating the driver's natural rolloff. (Less critical the more drivers you use).

- room correction. I haven't done this, but it is possible to partially correct for reflections, based on in-room measurements. It can look good in measurements but can sound worse, especially to listeners outside the 'sweet spot'.

Combining all the corrections you should end up with the ability to reproduce a square wave, and obtain a perfect impulse at the listener's ears - but only at one spot in the room, probably!

All of these refinements would not be possible or easily adjustable with a passive speaker, and after a while you realise that they all contribute to the sound in subtle or not-so-subtle ways.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.