I built my latest Digital Audio Workstation machine in the summer of 2006. It was an interesting process, and I thought that writing about it might be helpful to others, so here goes. Watch out, though: this one’s going to ramble a bit.
For a guy who learned recording on the old analog technology, what some of us call the “big iron”, the move to current digital technology and methods has been a major adjustment process. I have both had to learn new things and unlearn some of the old ways. When I started learning, most recorders used tape, and all of them were analog. For that reason, we didn’t even use the word “analog”: it was the only game in town. In a professional studio, you had a big multi-track tape recorder (usually with 24 tracks), a smaller 2-track tape recorder, and a very large mixing console. You also had a rack filled with extra equipment that could be connected for various “special” things like compression, reverb, etc. You expected to keep most of what you bought for a number of years. In a lot of places, there is equipment of this sort that is still in operation after anything from 20 to even 50 years.
In today’s world, so much has changed that it takes less time to list what hasn’t changed. For example, we still use microphones, preamplifiers, power amps, and speakers. Those items are still things that you might keep for a long time. Most everything else, though, has a pretty short life. Software is constantly being updated, so that even if you still use the same “product”, there always seems to be a new “version” of it just around the corner. Much of the hardware we now use is basically “throw-away” technology. Any hard drive that has been running for 5 years is probably on borrowed time. An entire computer that is 5 years old is very unlikely to be still in use for studio production. Since I first fully committed to using computers for recording, I have been moving to a new machine every 2 or 3 years.
Until recently, the main reason for a computer upgrade has been the need for more power. Sometimes this is driven by the arrival of new software that has more features and power, but demands better hardware to run it. At other times, it is just a matter of knowing that the next “faster” machine will enable you to do more.
I have been using recording software from the same company for over ten years now. Their first product, called Software Audio Workshop, ran well on a 486 processor with 8 Meg of ram. It would allow you to record, edit, and play up to 4 tracks of 16-bit audio, and was far and away better than any other recording software available for the PC at that time. I built a new computer then just to accommodate that program (the old machine didn’t have enough memory). Since then there have been several other products, each bigger, better, faster than the last, and each requiring more processor power and memory to serve its greater capabilities. The current product is SawStudio, and before “stepping up” to it I did yet another in my series of computer upgrades, this time to a Pentium III 866 with 512 meg of ram (that one is now my “web” machine, on which I am writing this article). The next machine, which I built myself (the previous two were custom built for me by The Computer Guy, who is still one of the top system builders, IMO), used a P4 1.8 GHz processor on an Intel D845GRGL motherboard with 1 Gb of ram, updated a while later with a 2.8 GHz processor so that I could run more plugins on my mixes. I was pretty happy with that machine for some time, and even now it could still do any session I throw at it.
So why did I upgrade this time? Well, there were reasons, but for the first time raw power wasn’t really one of them.
My original idea, in fact, was to use an old “spare” sound card I had to build a sort of downscaled machine that I could use for remote recording work. I would use this machine to record the tracks in the field, then bring the files home to mix on my “main” machine in the studio. I expected to be able to do this very cheaply because I only needed a P3 machine for this job. Unfortunately, though, I was not able then to find a cheap used machine that would work with my old Sonorus StudI/O sound card, which was rather finicky about the systems it would work with.
I also explored the idea of using an older laptop to control a desktop computer so that I wouldn’t have to lug around a heavy CRT display. After some preliminary testing, I abandoned that idea as not very practical. At that point I left things as they were, because I did not have the budget to do anything better.
Circumstances then stepped in to get me back in action. Intel was just launching their new CoreDuo processor, and in the process got into a price war with AMD over some older processors. Certain models that had been more than $300 in price dropped to near $100, and I started investigating again.
This time I decided to put a little more thought into exactly what I wanted in a machine, in particular what I expected in terms of features and performance. This should always be your first step when you think about a new machine, and too many of us overlook it. I decided that I wanted the best compromise among a list of requirements: These included:
- Small size and light weight for easy portability
- Low cost
- A modest increase in processing power
- Low noise from the machine so that it would not need a special box or closet
- No need for direct access to the box in normal use
- The ability to use my most recent sound card
- Compatibility with new technology
- Compatibility with “legacy” technology
Today the most obvious solution for easy portability is a laptop computer. For many of us, however, it is not the best solution. The reasoning behind this deserves an article of its own, but here it is enough that I wanted to use the PCI sound card I already had, an RME HDSP 9652, in the new machine. For me this meant that the best compromise for size would be a microATX motherboard in a “mini tower” case.
My last few computers all used Intel motherboards. These all gave me good service in systems I felt that I could trust. I still consider them to be a solid choice, but a chat with a tech advisor at a computer store persuaded me that this time AMD deserved a closer look.
The first reason for this decision was that I had heard that the AMD 64-bit processors actually provided better performance gains for 32-bit software than Intel’s first 64-bit devices did, IOW, the AMD processors offered better backwards compatibility. This was important to me because, for various reasons, I still use Windows 2000 sp4, and may continue to do so for some time to come. Windows 2000 is NOT going to be “upgraded” as a 64 bit OS, and the later Windows versions are not on my list of things I want. In addition, SawStudio is still coded as a 32-bit application (which makes sense until 64-bit OS setups completely take over, which won’t be for a while yet).
Next up was the fact that Athlon 64 processors actually run a bit cooler than current generation P4 devices, which allows the use of quieter CPU fans. For my machines at least, this is definitely true, as the processor fan on my new Athlon 64 machine is MUCH quieter than the one on my old P4 machine. In both cases the processors were purchased as “retail box” versions (you get a better warranty that way), and therefore both came with the recommended “stock” fans from their respective manufacturers. I would say that makes for a very fair comparison.
Finally, the AMD processor line offers some interesting future upgrade options with the right choice of motherboard. With the mobo I bought, even though I chose a fairly cheap processor, I have the option of simply plugging in something MUCH more powerful later on without having to make any other changes to the machine, which is one more factor in making the machine useful for longer before I have to build a whole new one (I have done “midlife” processor upgrades in at least two of my former machines, an approach which can, at the right time, give very good value for the money spent). I chose the Athlon 64 3200+ processor.
The absolute key decision in building a new computer is your choice of motherboard. I put more time and effort into this choice than any other single decision in the process. This can be a scary step, as there is an absolutely bewildering variety of choices available. The way to make the choice process manageable is to narrow down your options by first deciding what features you must have. I had already decided that the mobo had to be in a microATX form factor to allow me to use a decently small case. This makes the difference between a machine you have to wrap both arms around to move and one that you can easily tuck under one arm (leaving your other hand free for such trivial tasks as opening doors).
I had already decided to use an AMD processor, and in particular a Socket 939 unit (Although the Socket 940 was “latest and greatest”, the Socket 939 allowed a greater variety of processor options). Since Intel does not make motherboards for AMD processors, I had to find another manufacturer. I had heard from several different people that ASUS has a good track record with their motherboards, so I decided on that brand.
ASUS offers a wide range of motherboards, and there were still several choices even in microATX AMD processor Socket 939 designs. I had to have at least one “legacy” PCI slot to work with my chosen sound card, but I wanted the rest to be as “current” as possible. I also needed to have onboard LAN and USB2 ports. I decided that SATA capability was important because I don’t know how long the old parallel ATA drives will still be available. I also wanted onboard video. I finally settled on the A8N-VM mobo as the best compromise for my needs.
After processor and motherboard, you need to decide how much of what type of memory to buy. Although it is not the cheapest, for some time now I have exclusively chosen Micron memory. I have never yet had a stick of it fail. I bought 2 512 Mb sticks to give me 1 gig of ram while taking advantage of the dual channel memory capability of the motherboard.
I was not particularly picky about the case I bought. I found the cheapest microATX tower case that was in stock locally, for which I paid just over $50. It was made by Foxconn, which I recently found out is now the world’s largest electronics manufacturer.
I deliberately chose NOT to buy either a floppy drive (when was the last time you used one?) or a CD-rom drive. Not including a CD-rom may seem odd at first, but this gets back to the noise control issue. No matter how quiet a computer is, it always makes some noise. Having it sit right next to you may not always be the best idea, especially if you are going to use a microphone in the same room. In portable use, it is wise to assume that this is a real possibility. You will probably want to keep the computer tucked somewhere out of the way. It is pretty common to enclose the computer in a box or even in a closet, which makes any direct access to the box difficult. For remote use, I will often put the computer under the work table I use rather than on top of the table, where the noise may be in the way, and the box itself may block vision. In that case I do not want to have to reach under the table for any normal operation.
There is also the fact that this is not my only computer. I use a different machine in the studio for all my online stuff (My DAW is NOT used for surfing the Web!), and that machine already has a CD burner drive. On the occasion where I need to move something from CD to my DAW, I simply share that drive across the network to access files on CD from the DAW. When I built my previous DAW machine, I installed a DVD reader drive in it, which I used for installing the OS and other software. After that, the old DAW was closed up in a box, and whenever I needed a file from CD I just used the drive on my Web machine. One day I wanted to use the DVD drive again, and I discovered that it was no longer recognized by the computer. All I could figure was that this drive must have died of neglect. That money-wasting experience made me decide that there was no need for more than one optical drive among a group of networked machines, and the only time an optical drive absolutely must be directly connected to a given machine was for the original operating system installation. For installing Windows 2000 on my new machine, I used an external CD drive connected to the USB port, and this worked just fine.
There was one interesting feature in my new computer that I did not know about until after I built it: I can “turn it on” from a PS/2 keyboard. Because of this neat feature, I almost never need access to the front of the computer case. In the studio, I keep the machine pushed nose-first into a big wooden box, which helps reduce noise from it. I usually cover the back with a piece of acoustical foam to further reduce noise. I do not, however, completely close off all outside air to the machine, as I want it to keep cool.
The initial construction of this machine cost me just over $300, but I found out that I wasn’t quite done with it yet. There was still a process of “discovery and improvement” ahead of me.
I originally intended to use a hard drive from an older machine for it, but realized that I wanted to be able to do a full, clean install, and I had a lot on the old drive that I did not want to lose. I spent about $80 on a 160 gig Seagate SATA HD that I found on sale. One of the really neat changes in computer technology in recent years has been the overall improvement in the performance of hard drives. Not all that long ago, if you were serious about multi-track recording and mixing, you had to very carefully select hard drives, and in the beginning only SCSI drives were suitable for that use. Today, almost any HD that runs at least at 7200 RPM is capable of streaming more than 24 simultaneous tracks at 24 bits and sample rates of up to at least 48 kHz. For me, that meant I could get away with the cheapest drive I could find and not worry. In fact, this one is the quietest and fastest HD I have ever had in a system.
My next unexpected problem was about video. With my old DAW, I used dual 17” CRT monitors, each running at 1024 x 768 resolution. I had gotten used to the added display room, but this new machine had to be portable, and even lugging one CRT like that up and down stairs was more than I wanted to deal with. About the same time as building the new computer, I had made the change to a single 19” flat panel display. Using it with the old DAW and its ATI dual monitor card, I found I could create a single 2048 x 1536 desktop, of which any 1280 x 1024 portion could be seen in the display. I could scroll to whatever was off-screen just by moving the mouse, and this let me have all of the display space I needed while only having to carry around a single relatively lightweight display.
I had hoped that the built-in video in the new computer would be able to do this, but I discovered that I would have to install a dual monitor card to get this capability. I could not take the card from the old machine, because the new machine did not have an AGP slot. I had to buy a PCI Express video card, and I found one for about $40 that used the ATI Radeon X300 chipset. One important feature of this card was that, instead of a fan, it had a large metal heat sink. I did not have to worry about added fan noise from this video card. By using the included ATI Catalyst software, I was able to have the large desktop area that I wanted, just the way I had with the old machine.
There was still one more improvement I felt was needed. Part of my original design goal was to have a machine that would be as quiet as possible. I now discovered that the dominant noise source in the new computer was the power supply fan. I replaced the power supply that came with the computer case with a 400 watt supply made by Cooler Master that cost about $60. The computer as it now runs actually uses less than 100 watts, but I deliberately chose a fairly big power supply. This supply has a fan that can run at low speed if it does not get too hot. By deliberately using an “oversize” power supply, I cause it to be very lightly loaded, so that it can very easily dissipate what little excess heat it does generate, which means the fan will probably never have to speed up and become noisy to keep the supply safely cool.
I have been very pleased with the results of this project. This is the quietest, yet also the most powerful, computer that I have ever had. It is light and easy to move around. Its more advanced features make it a great technology “bridge” that lets me use valued older hardware a while longer while letting me also take advantage of some the latest features and capability, and this also gives me a good chance that I will be able to re-use a fair amount of the hardware from this computer in the next one. I did not, this time, try for the “most powerful that I could afford”, because raw power was not a particular issue this time around. Going for a “less powerful” option allowed me to save money while still satisfying my most important goals and needs.
I decided to write about this project in hopes that some of what I have learned might be useful to others. When it’s your turn to “turn over” to the next computer, start by making sure you know exactly what your goals are, take time to do your research, and choose your parts wisely. I hope, in this regard, that you find my experience useful. Building your own DAW can seem a bit scary at first, but you know what? “It ain’t exactly rocket science!”