Audiophiles review fiber optic digital audio cable

by philg

As a recovering audiophile myself, I was heartened to see the spirit alive at Amazon. I’m getting a 6′ optical digital cable to connect a desktop PC with an outboard digital audio converter. This is mostly because the speakers are making some funny noises and I’m too lazy to debug whether the extraneous sounds are coming from the sound card or the old M-Audio powered speakers.

Note that the function of this cable is to transmit 1s and 0s, with an error correcting code (essentially some redundant data) included so that the digital data can be completely recovered by the receiver even if there are some errors in transmission. Asking “is the sound different?” is equivalent to asking if you would get a different version of nytimes.com in a Web browser by connecting via wireless or a CAT5 cable. Yet there are 541 reviews and we learn from one review “Sound quality: The quality of the sound was on par with some of the more expensive cables I’ve had previously. I didn’t notice any interference and it sounded fairly clear.” and that “If you bought this $450 cable [competitive product], I cry for you.” Numerous other reviews refer to the sound quality of the cable.

Speaking of sound quality… what do people like for desktop digital audio converters? I don’t need a headphone amplifier (and probably wouldn’t want one powered by USB if I did), but Audioengine D1 looks pretty good.

20 thoughts on “Audiophiles review fiber optic digital audio cable”

  1. Update: I decided to use my PhD in Electrical Engineering to figure out if the popping noises were coming from the sound card or being generated by the M-Audio speakers internally. After approximately 30 years spent on the MIT campus, I came up with the idea of unplugging the powered speakers from the RCA cables coming up from the PC. The popping noises persisted. Still… I think it would be nice to have an outboard DAC. The Windows 7 Device Manager control panel indicates that Realtek High Definition Audio may be involved. This was a high-end HP desktop computer so it presumably has some decent components but the sound quality has to be higher if the signal is converted out of the electrically noisy PC case, no? [Asking the question another way… has anyone really noticed a sound quality difference in moving the DAC out of the case? I wouldn’t imagine that a $100 DAC would have a better chip than what is on the sound card board. And the noise inside a PC is all much higher frequency than audio frequencies. So really what is the point of moving circuits out of the case?

  2. Your assumption that HP used decent components is probably false. Realtek chips, whether sound cards or Ethernet controllers, are bargain basement commodity components used to tick a marketing checklist. Not that Macs are any better, the audio circuitry in my $4500 Mac Pro is nothing to write home about. The cost difference in the DAC IC chips themselves, e.g. from a Realtek special of the month to a higher quality ESS Sabre, Wolfson, AKM, Cirrus or other quality DAC ,is not that high at wholesale prices, under $10, but when amplified by markups and the cost of audiophile pixie dust, it can easily go in the 4 figures.

    What playback software you use makes a difference. Many programs (e.g. iTunes have shockingly poor audio buffer management, specially in this era of machines with multiple gigabytes of RAM, and will try to load audio samples from disk too late, potentially causing clicks or glitches if the system is busy when the disk fetch occurs (and inefficient modern operating systems have an absurd amount of behind the scenes busy work going on doing useless things like swapping out desktop background textures).

    An outboard DAC will make an improvement over consumer audio gear, but not over a pro audio sound card, i.e. what you’d get at B&H or Sweetwater, not Best Buy – the latter have proper galvanic isolation and power regulation, along with higher quality output op amps with higher impedance. To get the best output on Windows, you will want to bypass all the mixers, sample rate conversion, sound effects and other knick-knacks Windows “helpfully” injects in the audio path (shades of Clippy). That’s why many pro audio devices use special ASIO drivers that bypass the Windows multimedia subsystem altogether. You may find this guide useful:
    http://www.benchmarkmedia.com/wiki/index.php/Windows_7_Audio_Playback_-_Setup_Guide

    I personally use a Benchmark DAC2 connected using Toslink, but you don’t need the headphone amp. I can recommend the Cambridge Audio DACMagic and the Arcam rDAC and rPAC.

  4. “has anyone really noticed a sound quality difference in moving the DAC out of the case”

    I experienced much better audio quality when using a $30 sound card instead of the audio output of my computer’s motherboard. The difference is amazing and I’m using modest speakers.

  5. I experienced significantly higher sound quality using a USB based DAC, about $20. Something about the DAC being inside the case lends itself to all kinds of extra noise being introduced. So I would say your premise is correct moving it out of the case will make a big difference. Get whatever outputs you need for your setup.

  7. Do you remember the 80’s audio phenomenon of “jitter”? In an inexplicable feat of bad electrical engineering, all the early CD players generated a clock for the digital-analog converter by attaching a phase locked loop to the raw output bit stream. Instead of converting samples at a perfectly metronomic rate, the DAC would speed up and slow down in response to fluctuations like irregularities in the disc motor speed or even because the hole in the CD wasn’t punched exactly at the center. Tiny variations in a DAC’s time-per-sample introduce very interesting distortions into an audio signal.

    So that’s why different digital cables used to give different sound quality…and for all I know, maybe they still do.

  8. Johnny: that doesn’t make sense to me. The CD’s rotational speed is supposed to fluctuate, starting at 500 rpm the center and slowing down to 200 by the edge.

    The explanation given to me how to make that work is to have a DAC with crystal clock read out a buffer; when the buffer got too full or two empty it was time to adjust the speed accordingly.

    Doing the speed adjust any other way (which could introduce jitter) seems very hard to me.

    I also don’t think holes are punched; they are created as part of the moulding process.

    Your explanation also doesn’t make a lot of sense in explaining different cables giving different results, as it claims in the last sentence.

  9. Johnny: If there is jitter in the output of, say, a sound card, I don’t think that any cable is going to de-jitter the signal.

  10. Phil: The point is that some early equipment put the phase locked loop on the receiving end of the cable. So any smearing of the digital signal by the cable would influence the timing of the clock signal produced by the phase locked loop. If a square wave came out less square, that would jitter the timing.

    Bas: I don’t know how CDs are made today, but in the 80’s the audiophile magazines explained that the foil was etched before the hole was cut.

    Yes, a CD’s bits per rotation varies from beginning to end; I don’t know the details of how the early CD players dealt with that, though I believe the buffers were very small at that time, and part of the phase locked loop’s job was to adjust the motor speed to keep the small buffer from ever getting completely empty or full. I think memory was still expensive then. I was staggered when I learned that it was standard practice to use the PLL instead of a crystal to generate the DAC clock.

  11. iTunes used to have perfect buffering. I have no idea why it got worse, but it certainly is bad now.

    CD changes rpm, but it does that to maintain a constant bit rate. However, the bit rate coming off the CD is not perfect for many reasons.

    Because the receiving DAC cannot control the flow of data from a CD Player, there is no option other than a PLL or other method of changing the rate to match whatever comes in. The only way to fix this is to send the clock from the DAC back to the CD Player, an option that 99.99% of CD Players do not allow. This is a bad design from the start, making fiber optic a bad, albeit convenient choice.

    The best option is a computer CD drive, where the software can buffer data and alter drive access to match the DAC rate. But then you need a cable like FireWire or USB that allows the DAC to communicate the desired clock and data rate back to the controller that is reading the media. There are still many opportunities for error, but at least it’s feasible to solve (unlike one-way fiber optic).

  12. Johnny: Maybe they were talking about how the masters were made? I was in a factory in the early 90s and it was all injection moulded. I assume it has always been that way.

    I imagine the buffers being rather small, so I can see jitter being caused if it did run dry and bits ended up being sent at “best effort”.

  13. Re: Caustic’s analysis. Yes, I’ve heard the same thing, though I highly doubt the part about CD rotational speed ever coming into play. But yes, receivers get their decoding clock from the incoming digital signal. This is ALWAYS how digital signals are received. You need to PLL sync to the incoming bit stream, as you obviously can’t rely on your internal clock being in phase (or even sufficiently frequency matched) with the source. And if you think about it, there’s really no other way to do it. Having a perfect clock is pointless if the incoming data is coming in a bit faster or slower. After a long period of time, you’ll either have to pause or you’ll run out of buffer, depending on how your clock differs from the source. So, the receiver clock HAS to be phase locked to the source clock. The question is how well that phase lock loop suppresses jitter (assuming the jitter of the local clock is negligible itself). By definition, SOME of the jitter in the bit stream will make it through the PLL’s loop filter, and so Caustic is absolutely right that any jitter in the bit stream will have an effect on the sound. He is thus also right that the cable could, in THEORY, have an effect on the sound quality by shaping the signal.

    So, it’s physically possible for the cable to matter in a digital audio system because of the fact that the digital audio stream is not just sending over binary data, but is also serving as a clock. However, I’d be really surprised if anybody could hear that difference in a well-designed receiver, which should do a good job suppressing jitter by using a proper PLL design.

  14. Asynchronous Sample Rate Conversion is used to address the issue of transmission introduced jitter. In that case offboard DAC has own local jitter free oscilator. The difference in the frequency of the local clock (i.e. conversion rate) and incomming S/PDIF stream is overcome by intense math calculations performed in chips like Cirrus Logic’s CS8241.

  16. Re: “However, I’d be really surprised if anybody could hear that difference in a well-designed receiver, which should do a good job suppressing jitter by using a proper PLL design.”

    When CD players first came out, a lot of the audiophile magazines complained that they were unlistenable compared to vinyl records. They sounded really good at first…but some reviewers said that the treble seemed cold and brittle compared to vinyl. Some said that they couldn’t tell the difference in a 30-second A/B test, but that listening to vinyl all day gave them pleasure, whereas after 15 minutes of CD listening they had an inexplicable urge to turn the damn thing off. There were lots of arguments between people who couldn’t stand CDs and people who said the “GoldenEars” were just imagining a nonexistent difference.

    After a few years, the audiophiles and engineers finally figured out that jitter was responsible for the subjective difference, and some of the audiophile manufacturers managed to produce some very expensive CD players that even the GoldenEars enjoyed.

    Probably only the most Golden of the GoldenEars can hear the smearing in a digital cable, and they probably wouldn’t perceive it immediately; rather, they’d feel more irritated after an hour of listening.

    But “proper PLL design” is difficult, and there’s not much incentive for low-end manufacturers to bother. I doubt that things have changed much, except that fewer people now than then have ever heard truly good audio or know what they’re missing. The cheapest way to avoid jitter is to use a crystal clock, but as Jonathan says, that’s not an option if the crystal can’t control the rate at which the bits are coming in. Modern computers have very large buffers, so for them it should be an option. I would be curious to know if they’re taking advantage of that option, or still mostly using PLLs.

  18. Phil,

    Here’s a list of DACs with a wide price range:

    Shiit Bifrost
    Wadia 121
    Audioquest Dragonfly
    Musical Fidelity V-DAC 2 (I use the original version)
    NAD DAC 1
    Halide Design DAC

    If you are looking for the simplest solution there are several powered speakers out there that accept the digital outputs from iPods and computers.

    PS Have you heard anything about Tufte’s space in Chelsea? I haven’t been able to make it over to that part of town yet?

  20. Just get the Behringer UCA-202. Also, the popping sounds could be coming from static electricity.

Comments are closed.