Ardour and DXD

Hello,

i am an ardour’s user for many years, it’s an amazing sofware, very powerfull and stable. I have tried to do DXD (352.8/24bit) editing with ardour 3 and jack with my usb3 dac, all work perfectly but i can’t export in this format, it’s limited to 192khz even i choose “same as session”. Why ? it would be great if the DXD format was totally supported, i think it could be become one of the future standard.

Thank’s for all

julien

Are you producing music for bats?
Sure, some believe in the “extra samples at editing stage” but even there, 96kHz is all you need if the final target is sampled at 44.1kHz or 48kHz. If it isn’t, you’re a bat with a fine set of ultrasonic audio reproducing kit, or you need to ease up on the snake oil intake.

You don’t need to be a bat to hear in-band aliasing artefacts or pre-echo from compromised ADC and DAC filter designs. Both are reduced by using higher sample rates than may appear necessary if you only take a simplistic view of converter design.

seb jrigg maybe you should first understand what DSD/DXD is all about !!

http://www.merging.com/products/pyramix/dsd-dxd

jrigg, sorry, not aimed at you.

@chrisg: From Merging's page: It is a departure from normal PCM digital audio which allows an enormous jump in quality from the normal 44.1 kHz to 192 kHz universe. . In other words, largely bullshit.

from wikipedia:

In September 2007 the Audio Engineering Society published the results of a year-long trial, in which a range of subjects including professional recording engineers were asked to discern the difference between SACD and a compact disc audio (44.1 kHz/16 bit) conversion of the same source material under double blind test conditions. Out of 554 trials, there were 276 correct answers, a 49.8% success rate corresponding almost exactly to the 50% that would have been expected by chance guessing alone. When the level of the signal was elevated by 14 dB or more, the test subjects were able to detect the higher noise floor of the CD quality loop easily.

More background reading on the origin of DSD in purely commercial considerations:

  • http://www.ayre.com/insights_dsdvspcm.htm
  • http://www.realhd-audio.com/?p=629

The latter suffers from the usual “we didn’t do a blind test” but makes some interesting anti-DSD observations based on studio work.

Going wider, a well conducted (though not perfect) double blind test showing that 24 bit isn't detectable by most people under most listening conditions: http://archimago.blogspot.co.uk/2014/06/24-bit-vs-16-bit-audio-test-part-ii.html

My point was that if you are using DSD (which is a choice some people make regardless of such tests) and you want to edit or process it then conversion to DXD is required, particularly if you want to then return to the DSD world. It wasn’t about the relative audible merits of such systems.

You don't need to be a bat to hear in-band aliasing artefacts or pre-echo from compromised ADC and DAC filter designs. Both are reduced by using higher sample rates

If you are using an ADC with a flawed design, wouldn’t it be more logical to replace it with a better one?
For some years now, audio ADC design has been a fully researched and solved problem.
Dan Lavry’s paper on optimal sampling rates says 60kHz would be ideal, and 96kHz is more than enough. Above that, signal quality is actually compromised in real-world systems, increasingly so as sample rate goes up.

In the real world the DAC is the least to worry about. Transducers, amplifiers and speakers will produce larger distortions. Any nonlinearity shifts some of the ultrasonic content down into the audible range.

All of this misses the OP’s point, which was to make a reasonable feature request to support an audio format which is increasingly common in high end (mostly classical) recording circles.

It will be added. I just didn’t anyone left with any impression that DSD or DXD is anything other than marketing fairy dust.

Hello every body,

thanks for the futur add Paul. It’s not only a question of frequence range but more a dynamical question, and for the numerical traitement (filters) a high frequency is more appropriate. I can make a difference between a CD and a DXD easily with a good headphone (specialy with the transients), but it was not the subject of my request… i think that it is a good things that ardour was compatible with DXD (not DSD… it’s another things… DXD is a PCM format not DSD).
Thanks for all.
Julien.

I often post this link in threads about sample rates etc.
http://people.xiph.org/~xiphmont/demo/neil-young.html

One of the conclusions there is that you never need anything over 48kHz sampling rate.

I nearly posted that link too.

48kHz is the most widely-used standard in professional recording studios, followed by 44.1kHz. (see http://www.pro-tools-expert.com/home-page/2013/8/16/music-producers-what-sample-rate-do-you-record-at-poll.html)
It’s sufficiently higher than 44.1k to make the antialiasing filter requirements a lot easier (for a 20kHz audio limit the transition band is twice as wide, 20k-28k instead of 20k-24k). It goes straight to video without S/R conversion, and of course converting down to 44.1k for CD isn’t a problem when needed.

Certain dynamics processing and distortion plugins benefit from a higher sample rate because they are liable to generate sidebands above Nyquist frequency which get aliased back into the audible range, but the better ones upsample and downsample internally anyway to mitigate that problem.

I was dismayed to see this, from one of the industry’s most highly regarded figures:

(from about 45:40)

Conflating (albeit in very authoritative terms) resolution (bit depth) with sample rate, and perpetuating the notion / myth of “digital stair-step” waveforms… sigh.

OMFG. That’s appalling. Neve should be ashamed of himself.

On the other hand, it was 14 years ago so maybe we’ll forgive him just a little bit.

On the other hand, it was 14 years ago...
Yes... but it is appalling :)

Silly statements aside, I could go on about how design compromises in most of the commonly used pro audio converter chips can result in audible artefacts which are mitigated by using higher sample rates, but fortunately I don’t have to:

http://www.hypex.nl/docs/papers/AES123BP.pdf

Reading converter chip data sheets (something rarely done in online discussions) can also be quite instructive.