MY HOLIDAY DREAM

of a „DAC”, that can really do it all

NBEARABLE heat of 36° C in the shade, it’s muggy, there's no air to breathe - so I will close my eyes and start dreaming. These are small dreams, not big ones like "peace in the world", and to be honest, they are quite selfish ones. Namely I dream of a D / A converter that can do anything. But not in the sense that it can do everything that modern, multifunctional devices can, but what interests me the most: to convert a digital signal to an analog one for every type of digital signal known from the history of music signal recording. It is an impossible device, so these are only dreams, nothing more.

⸜ DMP was one of the first labels using digital Mitsubishi X-80 recorders, recording 16 bit, 50.4 kHz signal

The impracticability of this design is that such a device would have to consist of many separate products in one housing. It would not be possible to make one universal DAC, each of them would have to be a separate circuit. This is because the companies involved in the early digital recorder projects each did it their own way, and their recordings are usually not compatible with others. Moreover, they are not compatible with currently used converters, decoding 16, 24 or 32 bits, with sampling frequencies of 44.1 and 48 kHz or their multiples (88.2, 96 kHz, etc.). This forces recording labels to compromise.

Initially, digital recorders were designed as replacements for analogue ones, only to reduce noise and distortion levels. The medium which the music material was released on was analog anyway, so such a digital tape recorder was used in a mastering studio, where it fed signal through a digital delay line to lathe i.e. a turntable with a head cutting varnish, from which matrices for LP records were made.

⸜ Archie Shepp album recorded in November 1977 using Denon DN-034R with sampling frequency of 47.25 kHz and in 14 bits.

There was no problem with the fact that - for example - the signal had a sampling frequency of 50.4 kHz (MITSUBISHI X-80), 50 kHz (SOUNSTREAM, M3 ), or a bit depth of 12 (DENON), 14, 16 or 18 bits (DECCA). The tape recorders reproduced the signal via the built-in D / A converters without any conversion - today we would say: "natively".

The problem arose when the owners of the music material decided to release it on CD. From the mathematical point of view, the matter was not particularly complicated, although it did require a certain sensitivity. The conversion of the signal with a frequency other than the basic one for the CD 44.1 kHz was performed in devices called - no surprise - sampling rate converters (SRCs) with the change of the bit depth. The conversion was asynchronous, so the output sampling frequency had nothing to do with the input sampling frequency.

⸜ Yuri Tashiro Piano Trio album recorded in September 1979 using PCM processor, the Optonic RX-1 in 14 bits, 44.1 kHz

And now - the mathematically simple operation from the sonic point of view turned out to be extremely problematic. There is no "standard" converting algorithm, and each company used (and applies) its own solutions. And while the input and output of any SRC is mathematically identical, the way it is done changes the sound. Everyone who was fascinated by upsamplers could experience it - each one sounded different, didn’t it?

⸜ Horovitz live performance recorded in November 1981 using Soundstream recorder with 16 bit, 50 kHz signal

I think that was the main reason for the poor sound quality of CDs with material recorded digitally on tape recorders prior to the Compact Disc format. Although digital recorders with tapes on which the material was mastered were still used in mastering studios, between them and the computer controlling the glass matrix cutter used for pressing metal matrices for CDs there was an intermediate element in the form of a SRC system.

Today it is done differently. On the one hand, sample rate converters are more advanced, they give better sonic results. Nevertheless, CDs and LPs with material recorded on a non-standard digital device are pressed today without the use of these machines. They simply disappeared from studios and apart from a few "crazy" ones, they are used only by specialized companies that provide services in copying their content to hard drives, already with classic sampling and word length values. Almost always the signal is upsampled and the label receives PCM 24/96 or 24/88.1 signal.

⸜ A unique M3 demo disc with one side recorded in the M3 mastering system, using a 32-track recorder and 2-track mastering tape recorder in 16 bit, 50 kHz

So with each step we move further away from the source of the signal, and this always degrades the sound. Interestingly, labels wanting to release this type of material on SACD discs encountered similar problems. You cannot convert such a signal synchronously and you have to use asynchronous circuits. The changes in the sound turned out to be so big that many companies decided to recreate the original tapes with their original sampling rates, decode them natively to analog form and then convert the analog signal to a DSD signal.

This is why I dream of a DAC that would accept any PCM signal, even from audio files, that without their conversion and in a native form, would convert them into analog signal. For the first time music lovers would hear the same sound as people working in the studio, and not just a shadow of these recordings, which are often breathtaking. It could be nicely "packaged" in terms of marketing, referring to the growing popularity of phono preamplifiers with different equalization curves. Each of the digital signals would light up a nice, specific for the type logo and it would look beautiful. The decoding of DSD, HDCD and MQA signals would be added to that, so that we would have a complete, do-it-all device.

⸜ One of the most famous discs recorded using the M3 (16 bit, 50 kHz) mastering system, Ronald Fagen's The Nightfly

However, such a device will never exist. This project was born in my head because in a series of articles I took a deeper look at the first digital recorders and I was delighted with their complexity, amount of work and involvement of their designers - mostly audiophiles and musicians - as well as excellent fault tolerance which was never repeated again. However, I have an impression that I am alone in my owe and that most people when thinking "digital" automatically associates it with a computer and the Pro Tools system. And that's not true, completely untrue ... ■

WOJCIECH PACUŁA
Editor in chief