Again a little out on a limb, I'm a fan of Single Ended Triode amplifiers.

Welborne Labs had a pair of Yote 300B amplifiers in need of a little TLC listed in their bargain bin at $1000. The simple addition of tubes and a couple of plate chokes was enough to get them both working again and they sound sublime. The 96dB sensitivity of the Supravox driver marries in well enough with the 8W output of the 300B tubes. While 8W does not sound like a lot compared to the power output of most solid state amps, it's more than enough to drive the Supravox drivers to ear splitting levels. I seldom need more than 85dB or so at the listening position, so can barely account for 1W of power from the amps.

The remarkable simplicity of SET amplifiers means that they reveal every nuance of detail from the source and throw an utterly convincing soundstage that's full of texture and tone. While SETs are not everyone's cup of tea, partnered with the right speakers they manage insight and dynamics that other topologies seem to smear. I have found the addition of these amplifiers driving the 3D Sonics speakers to be revealing of almost every change I have made to upstream components, making them a perfect base to use as a test system.

A Stevens & Billington transformer volume control provides volume attenuation rather than using software level volume controls in Vista. Unfortunately, software based volume controls attenuate the signal in the digital domain by dropping bits, which can lead to a loss of resolution as soon as the volume is moved below maximum output. I've found transformer based volume controls to be very close to sounding neutral, even outperforming shunt volume control made up entirely of 0.1% tolerance laser-cut metal film resistors.

Lastly, I suppose I should mention cables. Despite my attitude towards other components, I don't do funky high-cost cables and all the fuzz associated with them. For speakers, I use a single strand of solid core 24-gauge silver covered in a simple cotton jacket. The whole shebang costs a few dollars per foot from most good audio DIY outlets and sounds fine to my ears. Interconnects are made up using either Cat 5 cable or suitable solid core coax. Power cords are all generic off the shelf types, no special plugs or dielectrics needed.

The PC

There's nothing remarkable here: a Gigabyte X48T, 4GB of OCZ Platinum DDR3, a Corsair modular PSU, and an X6800 dual-core processor. A 250GB Western Digital Caviar hard drive stores audio files in WAV format. The operating system is Vista 64, which is not ideal for audio due to some of the open source software failing to function properly if at all. There have to be compromises somewhere and current study requirements keep me on the Vista coach just to keep up to speed with its administration. Anything using a dual-core processor running at over 2GHz is likely more than sufficient for an audio server. 4GB of memory is pretty much mandatory for Vista 64 to work efficiently when placed under any kind of load. Another caveat is that I have not managed to locate any USB-ASIO drivers for the Vista 64 platform. ASIO drivers are preferred to help lower latency by routing signals directly to hardware where possible. Vista 32 and XP are fully supported by USB-ASIO, so that's something to bear in mind if you'd like to keep latency as low as possible.

The Test System Outboard DAC solutions
Comments Locked


View All Comments

  • phusg - Monday, December 15, 2008 - link

    > It'll please you all to know I have no plans to write another audio article.

    Doesn't please me at all Raja! As I said I'm very grateful for you opening my eyes to the possibilities of DRC. Shame you (or your editor?) don't fancy taking all the criticism and making a kick-arse follow-up article.

    All the best,
  • Rajinder Gill - Monday, December 15, 2008 - link

    I do have some work at hand that I've held back from posting mainly because it opens up another huge can of worms. The testing involved is arduous and takes up copious hours. I am in a position to make spectral measurements, although I need to confirm a few things before I'd even come close to disclosing all of the results. Initial tests show a marked difference between the 2 DAC's used in the article at equivalent levels of gain (between 1-3db at certain freq points). There's a pattern to some of it. Although, I can attribute much of this to noise generated by the amplifier due to the increased gain required by the TDA1543 (ac related hum). I have a few things I can do to cross-correlate those results though, so it's not a complete dead end.

    Blind testing has shown that human can discern down to .75db of gain. I'm not sure on the consistency of all of that as some of the blind testing involved in the research did show an element of placebo in places. Once I’ve eliminated some of the variables and more importantly tested an internal soundcard output in comparison, I might be in a position to post something worthy of reading.

    I'll probably blind test a few of the solutions I have at hand before I add anything else here though. It's very difficult to keep all camps happy, and to keep it interesting enough for me to want to commit more time to it

  • DorkMan - Sunday, December 14, 2008 - link

    Raja, sorry if I had the Sarcasm dialed up to "11" when I made my previous post. It takes a lot of work to create an article like yours, and you did a great job. My beef was not with the quality of your work but with some of the "religion" that seems to surround this hobby. Whether it's tubes, coaxial speaker wire, or whizzer cone transducers, there seems to exist a group of people with a religious zeal for these things and for whom specs and double-blind tests are irrelevant. Whatever.

    But I commend you for your article. It was well-written, and I have no issue with your conclusions. I, too, calibrate my video-editing workstation position for response flatness.

    If you're not too turned off by all the sniping, why not do a series on audio myths? Double-blind testing of tubes and solid-state, and speaker wires?

    Anyway, hope you have a great Christmas season.
  • audionewbieyao - Wednesday, December 10, 2008 - link

    If you're a true PC audiophile, I'll say do you best to stay away from gaming sound cards, 'cause they're just not made for audiophiles.

    I suggest try out Asus Xonar, Auzentech, Onkyo, and if possible, the new Xonar Essence STX would be the best.
  • thietavu - Tuesday, December 9, 2008 - link

    I have myself tried to solve this "problem" for years: to create a good quality audio system around my computer. Since all my CDs and even a few of my vinyls are now digitized into almost 100 GB of data on the hard disk, the challenge was to get some good sound out.

    I kind of made a mistake by buying "too high-end" (or too good anyway) headphones (Grado SR-325) and a headphone amplifier (Musical Fidelity X-Can v3). After trying two high quality studio sound cards (M-Audio AP2496 and later AP192), I realized that the sound just wasn't what it should have been. There was always a slight "harshness" in what I heard, making the system sound a bit unpleasantly cold and mechanical.

    After a long consideration I then bought Musical Fidelity's X-DAC v3 D/A converter, plugging it to AP192's SP/DIF digital output. That made the change. The slight but disturbing harshness disappeared and some originally cold-sounding albums (like Poco's Inamorata) suddenly became both listenable and even enjoyable. And excellent recordings now sound excellent.

    Where did the difference come from? Judging by normal measurements, there should be no audible difference at all, since M-Audio card's analog output is of high quality. To me, it appears like the "magic" happens in digital->analog conversion somehow. In cheaper devices like most sound cards, that conversion seems to be somehow "rough", a bit like using a digital camera with too few real megapixels available (imagine a 10 Mpix camera with bad optics etc, and the result becoming maybe equal to a real resolution of only 3 Mpix or so).

    I admit that the differences aren't that big, and only very good quality headphones or loudspeakers etc. can reveal them. But after it is revealed, music simply sounds more like "music" in one system compared with another. After that, it's unpleasant to go back to more "mechanical" sound. But as said, everything is relative. I can enjoy an old, noisy C-cassette in an old car stereo as well, as long as the music is good! That's what it is all about: creating feelings. :)
  • DorkMan - Monday, December 8, 2008 - link

    Sorry guys, I know many of you are really into this nitpicky level of Audio Purity. I've been into audio since the late 1960's, when I, too, was fanatical about specsmanship. Now I see the error of my ways.

    Bipolar loudspeakers have been around since, say, the 1950's. They're a decent solution as long as you can deal with room resonances, as this article does.

    But a whizzer cone speaker? When did we fall off the truck? Please tell me that the whizzer is driven by a separate coil subsystem. If not, you'll get decoupling resonances as various frequencies and these effects can change as the transducer ages.

    I went to the manufacturer website. Not embarrassed about selling speaker "ribbon" cable for $1,000? Tells me something about the company--long of PR and a bit short on electrical engineering.

    Okay, on to the amps. Tubes! The retro look! I have familiarity with tube amplifier designs and think it's great if you want odd harmonics near clipping, a bulky box that actually glows and a chance to heat a cold room at the same time. Hey, it's okay; odd harmonics sound nice--but they're not an accurate reproduction of the input signal.

    Finally, I'm all for balancing out a room--as long as the listener sits in the same spot. In a living room, every guest is going to get a different response curve, with some probably getting socked by boost when there should have been cut for that location.

    Finally, trust your microphone. If it says the room is flat, you're done. By putting in a slight bass boost, you're making the music sound good to "your" tastes, but be honest--the room then isn't flat. Might as well use tubes as amps.
  • ackcheng - Monday, December 8, 2008 - link

    Not sure if this is mentioned before, other than Audiolense, there are a few other programs we can consider for room correction

    They are DRC which is free
    and Acourate">

    I myself use Acourate and found it a very good tool for DRC!

    But I do not think the so called DRC in Windows Vista a real DRC software!
  • RagingDragon - Saturday, December 6, 2008 - link

    Certainly not what I expected at AT, but I enjoyed the article, and I'm OK with AT expanding it's range of covered topics (i.e. audio, digital photography, and home theatre).

    The DRC software is certainly interesting, though not really relevant to my setup (headphones). Based on manufacturer specs the TDA1543 isn't very impressive; however, the Wolfson WM8741 looks much more promising. The article's conclusions suggest the WM8741 has a wider dynamic range, flatter frequency response, and is more accurate/precise. Which is exactly what I would expect from the DAC spec sheets. I'd like to see these external DAC's compared to some high quality PC sound cards - like the Asus Xonar series, or even the M-audio revolution/stereophile cards (though if I recall correctly the DAC's in these and equivent cards from Terratec are inferior to the Xonar DAC's).

    I was looking at prebuilt external DAC's a couple months ago, and was extremely disappointed in what I found. Despite high prices most of them used cheap DAC chips which would have a hard time beating integrated audio, let alone a discreet sound card with high quality DAC's. The author's Twisted Pear Opus kit has a better DAC (WM8741) than *any* of the prebuilt external DAC's I looked at, and units equalling the Doede Douma kit (TDA1543) were ludicrously expensive.

    I'm not conviced that external DAC's and amps sound better than high quality PCI cards - and even if they do, I still don't want the extra clutter.

    Since I'm living in an appartment I use headphones rather than speakers to avoid annoying the neighbours, so the upcoming Asus Xonar Essence seems perfectly suited to my needs: stereo only (I'm using stereo headphones so surround sound is irrelevant), an extremely high quality DAC (Burr-Brown 1792A), built in head phone amplifier, swappable opamps, and "professional audio capacitors" (probably snake oil, but they won't make the sound worse).
  • prd00 - Friday, December 5, 2008 - link

    I'd pick Onkyo USB sound card up anytime compared to those stupid creative cards. I've tried all sound cards, but right now, I've ended with onkyo.
  • htgts350monaro - Wednesday, December 3, 2008 - link

    Thank you Rajinder for this, and what looks like will be, a great series of articles.

    Please be aware that despite all those who profess this to be an inappropriate article for this site that I, and I think many others, feel they are incorrect.

    Being an amateur audiophile and computer enthusiast myself this is a great mix of both my interests. I have been wondering for a while whether it would be possible to combine these interests and simplify my current combination of source components for my home theater system which I also use for critical 2 channel listening. (not ideal I know but limited space and budget regretfully denies a separate 2 channel system for me unfortunately)

    It looks like you may be on the path to solving my problem - please keep going with this. One thing I have gained from this article at least is a new interest in DRC software...

    For all thos who say that all audiophile claims are 'snake oil' I ask this - have you actually gone into a high end hifi store and LISTENED to any of the products available and done your own comparisons? Have you ever compared different DAC's/OP-AMP's in the same circuit and system with components of a high enough quality/resolution to expose those differences YOURSELF? I have and the differences can be VERY obvious, to the point that double blind listening is not needed - you would not need to double blind test two LCD's if one had half the contrast range of the other - you would be able to detect it quite easily and this can be true of audio components no matter how many tests say that the components measure the same frequency response/ dynamic range/ s/n ratio etc.

    But please go out and try it for yourself BEFORE you claim it is all bogus. (and no - not being able to hear a difference between components on your $300 minisystem or $50 PC speakers does not count!)

Log in

Don't have an account? Sign up now