Qobuz Lies about High-Resolution Music

I got an ad on Facebook today from Qobuz, the beleaguered French music service that offers both streaming and downloads.

Qobuz lies

You may have seen similar graphics before about high-res music; I think Sony was using this in the past. But it’s simply a lie. Digital music files do not create stairstep waveforms, no matter what the resolution. Encoders extrapolate the data in the files to create fully fluid waveforms. Using this sort of illustration is just another way of trying to get gullible people to pay more for music.

And even if it weren’t true, it wouldn’t be a correct visual interpretation of the differences between music files. If a high-resolution file is 24 bits – as Qobuz claims its music is – then where does that show up in the graphic? This stairstep graphic has been used to illustrate differences in sample rate, not bit depth. (And it’s still a lie.)

11 thoughts on “Qobuz Lies about High-Resolution Music

  1. Absolutely right. The whole subject is a bit of a nightmare. What adds to the stupidity is that some people download, or otherwise acquire, allegedly Hi-Res music and then play the same through cut-price speakers or earphones.

  2. Absolutely right. The whole subject is a bit of a nightmare. What adds to the stupidity is that some people download, or otherwise acquire, allegedly Hi-Res music and then play the same through cut-price speakers or earphones.

  3. It’s modern marketing… misleading either accidentally or by intent. If you have a high resolution audio system you might notice the difference, if not you won’t. I do have such a system so I’m hoping that Qobuz will survive but the number of people who have such systems is too small alone to sustain Qobuz. So of course they have to provide normal resolution streaming too. What a pity though that they mislead to try to persuade owners of regular audio products to subscribe to something from which they won’t benefit.
    But then life isn’t fair.

  4. It’s modern marketing… misleading either accidentally or by intent. If you have a high resolution audio system you might notice the difference, if not you won’t. I do have such a system so I’m hoping that Qobuz will survive but the number of people who have such systems is too small alone to sustain Qobuz. So of course they have to provide normal resolution streaming too. What a pity though that they mislead to try to persuade owners of regular audio products to subscribe to something from which they won’t benefit.
    But then life isn’t fair.

  5. Something which only a handful of people in the world understand is that the low-pass filter at the DAC’s output DOES NOT “regenerate” or “recreate” the original waveform. In a properly sampled signal, the complete original waveform is always present, and is never lost.

  6. Something which only a handful of people in the world understand is that the low-pass filter at the DAC’s output DOES NOT “regenerate” or “recreate” the original waveform. In a properly sampled signal, the complete original waveform is always present, and is never lost.

  7. 5 yrs later, and Qobuz keeps lying about this: same graphs, same misleading marketing.

    – Their files may -and normally do- have a higher sample rate, but those additional samples don’t densify the aural region (20-20 Khz) as they show in their graphs; they instead extend data captured to supra-aural, above 22.05 Khz (this is basic digital theory, Nyquist and then Shannon discovered this decades ago in the dawn of digitalization!), up to 48 Khz in the case of a 24/96K file. A straight line in a vector drawing app isn’t more straight because you describe it with some additional middle points besides the first and last ones.

    – The 24 bits claim also aren’t doing what they say (ie: rounding the reconstructed wave, that it is, as you say in the article, already soft by the effect of the sum of sin(x) pulse functions that reconstruct that waveform): they achieve a better S/N ratio, but with 16 bits that noise was already lower than anything you can hear.

    Both issues (higher sampling and bitrate) aren’t completely useless though, as MQA (Tidal) has brilliantly shown (and could be easily heard in proper recordings, even compared with the very Qobuz hires files): Extended sampling rate allows much softer antialiasing filters in a nyquist frequency moved further away, thus leading to a significantly better resolution in time domain (MQA further softens this filter even more in other very clever ways); and those 24 bits allows to bury a data channel masked in the noise floor, which in MQA contains that extended supra-aural region (24 Khz to 48, 96, or even up to 192Khz) that allows not only that extended filter previously mentioned, but also a much more compact file for streaming, almost the size of a standard CD.

    Of course Qobuz do neither of both: they don’t work with the antialiasing filter, and they don’t take advantage of this potential data channel under the noise floor. Nevertheless, there may be a marginal aural benefit in time domain by moving the brick-wall antialiasing filter of a standard rebook pcm to a higher frequency, which is possible with that higher sampling rate.

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.