Giving up on The Absolute Sound

I’ve been a long-term subscriber to both The Absolute Sound and Stereophile.  Even as I could find reviews sooner and cheaper online, I stuck with them both as they had access that other writers do not, and seemingly reputations that they would want to defend by not making crazy mistakes.  While I will still keep my subscription to Stereophile going, once my current one to The Absolute Sound lapses, I’m afraid I’m done with it for good.

This year, The Absolute Sound did a four-part series on computer audio and the differences they found in different tweaks.  Of course, writers from TAS subscribing to tweaks and adjustments won’t strike anyone as strange, as they’ve been that way for a long time.  This series hit me closer to home as I have a degree in Computer Science, and was a programmer for a long time for my career and still write code when I need something done.  Because of this, I have a pretty good idea how a computer works, right down to the lowest levels of the system, and so my BS receptors are more likely to be triggered here.

Reading what they wrote about tweaks that helped made my head spin.  They talked about putting the computer on a granite shelf with fancy cones to isolate it, about replacing the power cable with a PS Audio cable, using an external HDD instead of an internal one, using a fancy power cable with an external HDD and so on.  There is then an entire article on WAV compared to FLAC, different FLAC encoders, and how WAV sounds superior to all of them still.  At this point I decided I had read enough from TAS for the foreseeable future.

First, lets just say that using a fancy power cord on a PC is pointless, as it is with an external HDD.  Regardless of what you think about the effects of a cord on an amplifier or other audio component, a PC is not a component like that.  Everything on the PC eventually goes through the CPU and RAM, and when it is there it is simply a binary 1 or 0.  There isn’t a larger 1 or smaller 0 due to a better power source, it’s just 1 or 0.  Additionally you won’t have errors occur here as the transport mechanisms in use have error checking built-in, so if the signal level were to be so low that a 1 could be confused for a 0, the error would be caught and the issue fixed before it was output.

Just like this, WAV and FLAC are bit-transparent.  When you select a FLAC file to play back, it will be turned back into a WAV file, and then fed to the CPU and RAM just like the WAV file will.  The FLAC file might be less subject to error as the smaller file size means that a smaller buffer more likely won’t be an issue as it could with a WAV file when streaming over a network, and CPU overhead isn’t an issue with FLAC decoding.  FLAC encoders also have to follow a set of standards in the FLAC specification; so all the outputs from them will be identical to each other, provided you use the same settings with each.  Any differences you were to hear would be ones you imagine hearing, as your computer is not seeing any difference in the two files.

Here, I believe The Absolute Sound is suffering from “magic box syndrome”, where they don’t understand what is going on inside the PC, so they come up with all the explanations they want.  I see this with Blu-ray players as well, where people want to see a difference in them, but we can check and know for a fact that they are producing the exact same image.  Instead of trying to isolate for every variable in the system, perhaps starting by using SSDs and a RAM disk to make sure latency isn’t an issue, TAS decides to assume that there are differences between the two.

When a reader complained about this, they compared the issue to different CD transports producing different levels of jitter.  A publication the size of TAS easily has access to test equipment to measure the jitter in the different setups, and compare WAV to FLAC jitter, but they didn’t choose to run those tests.  Similarly the Weiss DAC 202 that they have raved about has a bit-transparency test, that lets you use included test files to verify that the DAC is getting a bit-perfect signal from the PC.

Did they use this test with the variety of configurations that they tested to see if the DAC always saw the same data?  No, they chose to not run a single objective test that could give them an idea as to what is happening and instead rely on guessing and assumptions to rationalize what they think they are hearing.  They also say it’s too bad the reader won’t be around when they discover the cause of these differences they hear, once again treating the PC like a magic box, but they appear to have made no scientific attempt to determine if there even were differences.

Now I think the writers at TAS could think they are hearing differences, and they are certainly convinced of it.  Similarly I’m not going to say that I can do a measurement of an audio device and say with absolute certainly that it is good, or that I know exactly how it performs.  I don’t as that test is covering a specific set of parameters, and a specific set of devices, just like a listening session does.  I do know that TAS could have easily thought, “I wonder if we are hearing things, or if there is something else going on here”, brought in the objective test equipment, and quickly determine if they were actually producing any differences.  If they see no change in objective measurements, they could then think about what they are hearing and try to decide if they need to change their methodology or go back and start again.

Instead of taking the questions and concerns of their readers to heart, and running more tests, TAS seems to have decided that they are certain of these findings and can not be convinced otherwise.  As I’ve mentioned in my reviews, while my subjective opinion is subject to personal bias, visual and auditory acuity, and associated equipment, objective measurements should be universal and free from individual bias.  They can be published and then the reader can determine how to interpret them.

By choosing to completely ignore the objective, even in the face of ample evidence that the subjective thoughts have no actual basis, TAS has abandoned their idea of helping their readers find equipment and accessories that can improve their listening experience.  They have instead provided them with pages of illogical, ill-informed, and frankly ridiculous suggestions that don’t help them at all.  I’m no longer sad to see my subscription lapse, as I now know enough to seek out publications and writers that are secure enough to question their own thoughts and findings in the name of helping their readers, instead of thinking of themselves as impervious to mistakes.

, , , , , ,

  • Charles Campbell

    No offense, but I have to disagree.  Power cords make a difference on the computer and WAV does sound better than other file formats.  I found this to be true with no help or influence from TAS.  I know it is just ones and zeros but something else is happening that is definitely creating a different presentation.  I don’t like it because it’s next to impossible to explain by any science I know but there it is.

    • Chris Heinonen

      You’re committing the same logical error that The Absolute Sound does here. You hook up a new power cord and say “Oh, I think I hear something better, so it must be the power cord!” Instead of finding a way to actually AB the setup (which The Absolute Sound could afford to do) to isolate that variable, or measure the output to see if anything has actually changed, or do any number of things to figure this out, including talking to those that design computers.

      It often seems that only in audio do these magical properties come up. When I review a display for AnandTech, no one asks if I used a silver and gold DisplayPort cable, or a special power cable, to get the most out of the monitor, as they realize it won’t impact performance. No one asks Anand if he used special power cables when testing servers, or charging his laptop, since it makes no difference. With certain cables you have the potential to have a small change, but the power cord on a PC isn’t one of them. My issue with TAS is more that they take their singular subjective opinion as fact instead of taking it as a starting point to investigate what might be going on, which doesn’t help anyone.

    • Karel

      FLAC is a lossless codec. The data fed into the encoder is identical to the data coming out of the decoder, down to the bit level. Please read up on the difference between lossless and lossy codecs first.