½ a decibel louder

I rather like audiophile music brands. It’s nice to have recordings captured with integrity, often with higher than usual resolution, sometimes with surround sound. I like Blue Coast Records for just that reason. I’ve downloaded one of its DSD files for testing purposes, and on the rare device which can play it this sounds wonderful.

But I stumbled across something which casts into doubt any and all of its claims.

It sells some of its recordings on ‘MQD’ CDs. MQD stands for ‘Master Quality Disc’ and they are carefully constructed and use a 24 karat gold data layer. Blue Coast says that they are archival quality and can last for 300 years.

All that’s well and good, if true.

The company will also sell you blank MQD CDs, which you can burn yourself (does the dye on a recordable disc last as long as the gold substrate? Who knows!)

This is where things go off the rails. It makes a number of claims which are just plain idiotic:

When used for mastering music, we have noticed an increase of approximately ½ db in volume along with better frequency response and enhanced imaging when compared to our former mastering discs. We suggest you compare with your regular blank discs by recording the same music on to both and do a blind listening test.

Now they say that these discs have a lower error rate, by which I assume they mean a reduced number of incorrectly recorded bits. But this ought to be low anyway. What it cannot do is change either the volume or the frequency response, let along the imaging. It simply can’t.

To increase the level by some portion of a decibel means to alter the actual samples. Not just a few, but tens of thousands of them on a wholesale basis by a significant amount. If this is the more accurate medium, then the replaced one must have been on average reducing the absolute value of the samples. By how much? Half a decibel reduction is a 5.6% reduction.

But why would errors be biased towards the zero value? Bit errors should be random around the correct value.

And how would this affect frequency response? I assume they mean the top end, but bit accuracy has no effect on any particular frequency. The stream of samples is pretty indifferent to frequencies, and all the different frequencies are plastered over the top of all the other frequencies.

Still, they give themselves the traditional out: ‘Many of you may not hear the difference, which is fine. But a few of you will.’

In fact, if a half decibel of volume were a reality, it would be clearly revealed in properly conducted double blind tests. Not necessarily as volume difference, but as some qualitative difference. In fact, what this statement does is prime some people so that they are more likely to imagine a difference.

The problem with these silly claims is that they lead me to doubt the reliability of their other claims.

This entry was posted in Audio, CD, Mysticism, Rant. Bookmark the permalink.

3 Responses to ½ a decibel louder

  1. Jonathan says:

    A belated comment but I think you might have missed the point they are trying to make. I am certain they meant the increase and loudness and the frequency response claims are from subjective listening tests. I am also certain that Blue Coast Records are perfectly aware that if you ripped a title burned to one of their CD-Rs versus another brand of CD-R, the resulting files would be absolutely identical and would null each other out if you loaded them into a DAW.

    However my own experience is that CD-Rs do sound different to each other. I’m not imagining it because I have burned exactly the same CD master to different brands of CD-R using precisely the same burner and produced exactly the same disk such that if you rip it the files are identical. These CDs also all have exceptionally low error rates and typically manage a 97% score in the Nero quality test. Yet I can hear the difference. And to prove I am not imagining it, I can record the analogue output from my Rega ISIS CD player to my DAW at 24/96, then trim the files so they all start and finish on exactly the same sample, the successfully double blind test them in Foobar.

    The fact is that different CD-Rs can and do sound different and subjective frequency response and perceived loudness are two of the major indicators, along with the sound of an orchestra. If people can’t hear the difference then it is simply down to the lack of a highly resolving playback system and / or listening skills that are not as highly developed as those of others.

  2. Hi Jonathan. Perhaps that’s what they mean. I think they are in error. I’d be convinced by some carefully conducted double blind testing. I’m not sure what the 97% figure from the Nero quality test could mean. It certainly does not mean that three out of every 100 bits are wrong!

    The point remains: how can errors in bits effect volume level or frequency response, even subjectively? As discussed above, imperfections in the data carriage are random with respect to the music, even if not with respect to the digital data stream. If one sample is wrong and can’t be repaired, then it will manifest as a ‘tick’. If your DAW supports individual sample manipulation, try it. If a significant proportion of bits are wrong, they will drag the waveform this way and that, but not in a particular way that maps sensibly onto the ways that are used to characterise analogue sound quality.

    I am profoundly sceptical of claims that people can, blinded, hear all the differences they believe they can. See, for example, this post, particularly the fourth point.

  3. Pingback: AudioQuest wants you to buy $44,000 speaker cables | Gadget Guy Australia

Leave a Reply

Your email address will not be published. Required fields are marked *