Dear audio industry, fix mastering before bothering with high resolution

The audiophile world (small niche though it is) is buzzing with a renewed interest in high resolution audio, now to be known as HRA.

See, for example, Why the Time is Right for High-Res Audio, or Sony’s new Hi-Res USB DAC System for PC Audio, or Gramophone on At last high-resolution audio is about to go mainstream, or Mark Fleischmann on CD Quality Is Not High-Res Audio:

True HRA is not a subtle improvement. With the best software and hardware, a good recording, and good listening conditions, it is about as subtle as being whacked with a mallet, and I mean that in a good way. It is an eye opener. In lieu of “is that all there is?” you think “wow, listen to what I’ve been missing!” … The Compact Disc format is many good things but high-res it is not. It has a bit depth of 16 and a sampling rate of 44.1 kHz. In other words, it processes a string of 16 zeroes and ones 44,100 times per second. Digitally speaking, this is a case of arrested development dating back to the early 1980s. We can do better now.

As an audio enthusiast, I would love this to be true. But it is not. Fleischmann appears to be ignorant of the Nyquist-Shannon sampling theorem, which suggests that the 16-bit/44.1 kHz CD format can exactly reproduce an analogue sound wave from 20–22,050 Hz and with a dynamic range (difference between quietest and loudest signal) of better than 90Db.

Yes there are some ifs and buts, and if CD had been invented today it would probably have used a higher resolution of say 24-bit/96 Khz which gives more headroom and opportunity for processing the sound without degradation; but nevertheless, CD is more than good enough for human hearing. Anyone who draws graphs of stair steps, or compares CD audio vs HRA to VHS or DVD vs Blu-Ray, is being seriously misleading.

Yes, Sony, you are a disgrace. What is this chart meant to show?

image

If shows that DACs output a bumpy signal it is simply false. If it purports to show that high-res reproduces an analogue original more accurately within the normal audible range of 20-20,000 Hz it is false too.

As an aside, what non-technical reader would guess that those huge stair steps for “CD” are 1/44,100th of a second apart?

The Meyer-Moran test, in which a high-res original was converted to CD quality and then compared with the original under blind conditions (nobody could reliably tell the difference), has never been debunked, nor has anyone conducted a similar experiment with different results as far as I am aware.

You can also conduct your own experiments, as I have. Download some samples from SoundKeeper Recordings or Linn. Take the highest resolution version, and convert it to CD format. Then upsample the CD quality version back to the high-resolution format. You now have two high-res files, but one is no better than CD quality. Can you hear the difference? I’ve yet to find someone who can.

Read this article on 24/192 Music Downloads … and why they make no sense and watch the referenced video for more on this subject.

Still, audio is a mysterious thing, and maybe in the right conditions, with the right equipment, there is some slight difference or improvement.

What I am sure of, is that it will be nowhere near as great as the improvement we could get if CDs were sensibly mastered. Thanks to the loudness wars, few CDs come close to the audio quality of which they are capable. Here is a track for a CD from the 80s which sounds wonderful, Tracy Chapman’s debut, viewed as a waveform in Adobe Audition:

image

And here is a track from Elton John’s latest, The Diving Board:

image

This everything louder than everything else effect means that the sound is more fatiguing and yes, lower fidelity, than it should be; and The Diving Board is far from the worst example (in fact, it is fairly good by today’s standards).

It is not really the fault of recording engineers. In many cases they hate it too. Rather, it is the dread of artists and labels that their sales may suffer if a recording is quieter (when the volume control is at the same level) than someone else’s.

Credit to Apple which is addressing this to some extent with its Mastered for iTunes initiative:

Many artists and producers feel that louder is better. The trend for louder music has resulted in both ardent fans of high volumes and backlash from audiophiles, a
controversy known as “the loudness wars.” This is solely an issue with music. Movies, for example, have very detailed standards for the final mastering volume of a film’s
soundtrack. The music world doesn’t have any such standard, and in recent years the de facto process has been to make masters as loud as possible. While some feel that overly
loud mastering ruins music by not giving it room to breathe, others feel that the aesthetic of loudness can be an appropriate artistic choice for particular songs or
albums.

Analog masters traditionally have volume levels set as high as possible, just shy of oversaturation, to improve the signal-to-noise ratio (SNR). With digital masters, the goal
is to achieve the highest gain possible without losing information about the original file due to clipping.

With digital files, there’s a limit to how loud you can make a track: 0dBFS. Trying to increase a track’s overall loudness beyond this point results in distortion caused by
clipping and a loss in dynamic range. The quietest parts of a song increase in volume, yet the louder parts don’t gain loudness due to the upper limits of the digital format.
Although iTunes doesn’t reject files for a specific number of clips, tracks which have audible clipping will not be badged or marketed as Mastered for iTunes.

Back to my original point: what is the point of messing around with the doubtful benefits of HRA, if the obvious and easily audible problem of excessive dynamic compression is not addressed first?

None at all. The audio industry should stop trying to mislead its customers by appealing to the human instinct that bigger numbers must mean better sound, and instead get behind some standards for digital music that will improve the sound we get from all formats.

4 thoughts on “Dear audio industry, fix mastering before bothering with high resolution”

  1. “A violin should sound like a violin” is what I keep saying. But a violin has harmonics that go way above 20kHz, as do many, many instruments, and also the human voice. And the true thump of a bass drum, or the intake of breath of a singer go well below 20Hz.

    So 20-20 is not enough. Since it is easy and cheap to at least fix the electronic path between mic and LS we should start here, at least by mving up to 24/96 which with today’s internet bandwidths is easily downloaded, but makes the CD dead…

    After that we can start to judge the mic positioning, the mic quality, the LS performance ( in my opinion the weakest link) and the mixing techniques.

    But fix the transfer of signals from beginning to end, and that at least is one quarter of the problem tackled.

    1. Hi, are you saying that humans can hear these ultrasonics? The evidence for that is slim as I understand it.

      Tim

      1. The argument I have heard is that once in the analog medium the tones outside of your hearing will actually have an effect on the tones you CAN hear, or possibly even feel. So it should come across as more natural, more true to life.

        The biggest problem though is that affordable audio equipment is still optimised for the 20-20 range. You have to spend hundreds of pounds on headphones if you want them to go much below/above those limits, for a 7.1 home cinema configuration you would be spending potentially thousands on speakers.

        Portable devices are only just adopting DSPs capable of higher quality playback, but then they fall down because their headphones output is too weak to drive those expensive headphones properly. That is seems to have been ultimately the cause of the problem in the first place. Due to their top volume being much lower than home equipment and so people want the source material to be loud to compensate.

        My personal opinion? I’m not really sure. I have heard material in DTS 24/96 that sounded much better than CD, but my speakers are certainly not capable of far outside of 20-20. So its likely more due to the fact they went back to the masters and maintained the dynamic range properly.

        The only thing I can say for certain is that 80-90s CDs sounded much nicer than remastered/re-releases, or just about anything made since. There are already examples on CD where a re-release killed the dynamic range, leading to people clamouring for the original. (eg ABBA Gold) So at the end of the day you are right, they need to fix the mastering first before we can really make any judgements.

  2. Steven Wilson is one of the few modern artists that doesn’t have his albums mastered for maximum volume, preferring to retain the dynamic range. At one point, he wanted to include an explanation in the sleeve notes to explain why his records didn’t sound as loud as others in the listener’s collection (“Please, use your volume knob”).

Comments are closed.