From Acousence Records
SACD
We have not yet produced any SACDs at ACOUSENCE to date because we were initially far from convinced of the musical result when we performed our own experiments using the DSD (Direct Stream Digital) recording procedure for SACD, not to mention a strange tone colouration and unnaturalness in the timbre and tone structures of individual instruments. External productions that we have heard so far also gave no occasion for us to believe that anything else could be achieved with this medium at all.
As can be read in various professional articles written by ACOUSENCE's label chief, Ralf Koschnicke, temporal resolution plays a very significant role in music transfer, and in this respect, the DSD recording method of SACDs is open to criticism. Although the transmission range in this method is not sharply restricted by filters, as with a CD, the high frequency noise inherent in the system starts at 20 kHz already, so that the signal parts that are so essential to the musical fine structure are easily masked by signal noise, which ultimately leads to a similar result as the one effected by filtering in a CD.
The concept behind the sound carrier itself, however, continues to be very interesting; it is simply a breeze to operate and presumably the SACD is indeed the last, honest-to-goodness sound carrier that we will get for the foreseeable future - if not for ever. New experiments have now been performed, in particular on the basis of the current theoretical considerations with regard to the relevance of the HF information for musical accuracy, that were aimed at having the signal noise, if it cannot be avoided in the first place, start as high up as possible (with respect to the starting frequency) and thereby having all parts of the useful signal above the signal noise, if possible, at least at CD-level, that is -90dB. We have succeeded in doing so by recording the sum output of the analogue mixer directly via DSD converters by the EMM Labs company in DSD format and by using the dynamic range, which is significantly higher compared to conventional PCM audio, up to the limits of the allowed SACD specifications in the process. Incidentally, it became very clear that the temporal resolution of detail suffers considerably and that the strange tone colouration mentioned above occurs already if the level is only 8db lower, for example.
Now that our first SACD has been mastered, the results are convincing all around. The chosen method in particular - recording in 4fs (24bit/192kHz) multi-track, analogue mixing, then converting it very puristically 1:1 in DSD on the SACD - appears to be the ideal way for this format. Based on the experience made, pre-production in DSD is just as inexpedient – for that, the 64fs DSD of the SACD is simply unsuitable due to the noise problem – as the commonly used path via PCM digital audio in 48Khz or even 44.1kHz – in that case, the resolution is limited far below the potential of the SACD already.
On the two figures shown below, one can see a passage from our "DER SYMPHONISCHE RING“, digitalised both times for the purpose of graphical representation in 24/192 PCM but sent down the DSD ADDA route first. The measuring range is calibrated to 90dB=Black.
http://bbs.hifi168.com/showtopic-172246-4.aspx
