What is SACD?
SACD is an abbreviation for Super Audio Compact Disc, an optical music carrier that was designed to replace the sonically outdated audio CD format introduced in 1983. It is designed to provide drastically superior sound quality, and optional multichannel sound, while maintaining backward compatibility with CD.
Who invented SACD?
Who invented SACD?
SACD was developed by Sony and Philips. It was invented by Dr. Yoshio Yamasaki who also invented Super DVD a next generation DSD audio/video format that Sony rejected in favor of the lesser quality BluRay. The trademarks are owned by Sony. Philips is the licensor of the disc format and the trademark.
What does the beautiful logo mean? The logo simply shows SACD - the S and A in the upper half, the C and D in the lower half. The logo conveys a sense of fluid, organic, natural curves as opposed to the straight, angular shapes of the CD logo, representing natural, analog sound in contrast to the imperfect digital sound of CD.
What is a Hybrid SACD?
A Hybrid SACD is an SACD disc that can be played on regular CD players. On a Hybrid SACD both the CD layer and the high-density layer are 'read' from the same side, the other side has a printed label. The high-density layer is partly reflective, partly transparent. At the wavelength used by regular CD layers (780 nm) the SACD layer is invisible so a CD player will just 'see' the CD layer. At the wavelength used for DVD and SACD (650 nm), the SACD layer is reflective.
Are all SACDs hybrid?
In the early days of SACD, many titles from Sony, Exton, and others were released as 'single layer' i.e. SACD-only but nowadays virtually all SACD releases are hybrid discs. Single-layer SACDs look 'plain silver' while hybrid ones have a goldish shine to them. Dr. Demery of Sony stated this is due to an optical illusion caused by the dual-layers.
Dual DSD-layered SACDs
There is a second option for the second layer, rather than a low resolution CD version of the music, the second layer can be high resolution DSD and provide double the music capacity and in the future this may be combined with video, lyrics and/or graphics. Dual Layer SACDs are sometimes used for special surround demo discs or for long classical works. They are relatively rare so far, however after the overdue death of the low resolution CD, recording companies will no longer see a benefit of putting an unused CD layer on their SACDs will be exploring all the possibilities of dual DSD-layered SACDs.
There is a second option for the second layer, rather than a low resolution CD version of the music, the second layer can be high resolution DSD and provide double the music capacity and in the future this may be combined with video, lyrics and/or graphics. Dual Layer SACDs are sometimes used for special surround demo discs or for long classical works. They are relatively rare so far, however after the overdue death of the low resolution CD, recording companies will no longer see a benefit of putting an unused CD layer on their SACDs will be exploring all the possibilities of dual DSD-layered SACDs.
Are all SACDs multichannel?
No, especially in the beginning many SACDs released were stereo only. Most news SACDs rare Multichannel plus Stereo. Exton has recently switched back to Stereo only from Multichannel.
Is the stereo signal derived from the multichannel signal?
No. Unlike DVD-Audio, the SACD format does not support 'down-mixing'. When an SACD contains stereo and multichannel sound, these are stored separately on the disc. However some new multichannel players will play a mixdown of a multichannel mix if the stereo mix is absent from an SACD if 2 channel stereo playback is selected from the set up menu.
Do all SACDs contain a stereo mix?
Nearly all SACDs contain a stereo mix. There are a few examples of Hybrid SACDs that contain a multichannel mix but no stereo mix on the SACD layer. See SACD dishonor roll – No 2-channel Stereo high resolution program If you listen in 2 channel stereo make sure your SACD has a stereo program by looking up the title at sa-cd.net under Content and if it says "Stereo" or "Stereo/Multichannel" you can rest assured it will have a high resolution 2 channel stereo program.
What is DSD?
DSD is an abbreviation for Direct Stream Digital. It was originally invented by Sony for archiving their studio master tapes. Sony challenged their designers to come up with a truly transparent medium to transfer their aging analog master tapes to as even the highest resolution PCM offered a coloration that Sony found unacceptable plus they wanted this medium to be future proof and able to be downsampled to any consumer format to come along in the future. This would be the very last playing of many of these very old prized analog master tapes as many were literally falling apart.
How does DSD work?
DSD is basically a 1-bit Delta/Sigma conversion scheme. DSD being a 1-bit signal means that every bit represents a sample - a measurement of the amplitude of the sound wave at a certain time. Since a bit can only have two values, every bit in a DSD stream only tells weather the amplitude of the sound signal was higher or lower than the previous sample held in the buffer. Because it doesn't tell how much higher or lower the amplitude is, you need a lot of samples to accurately describe the signal. DSD uses 64 times the sampling frequency of CD: 2.822 MHz vs 44.1 kHz. The factor 64 is not randomly chosen. It's a power of two, meaning that it's relatively straightforward to upsample from typical PCM frequencies including 44.1 kHz and multiples like 88.2 kHz. For a more complete discription of the process see: Super Audio Compact Disc: A Technical Proposal
What is the difference between PCM and DSD?
PCM (Pulse Code Modulation) is a very abstract way to describe an analog signal in a digital terms. In PCM, every sample consists of a combination of bits describing the amplitude of the signal. The number of bits determines the resolution of how finely the signal can be described, where every added bit doubles the number of levels that can be distinguished.
Converting from analog sound to PCM and back to analog sound involves a number of processing steps, such as quantization. Every step can cause further distortions such as quantization noise, which has to be filtered out, thus deteriorating the sound quality.
DSD is an extremely simple way of converting from analog to digital and back again. The entire process is extremely transparent. In fact, the DSD bit stream is so closely related to the analog signal that if you were to play a uncompressed DSD signal as is you would here audible music.
Wasn't CD supposed to be "Perfect Sound Forever"?
Actually no even back in 1982 when CD premiered in Japan it was a step backwards from the 50kHz Soundsteam process used by the American recording companies. CD was highly compromised from the very beginning. Technology has progressed and so has insight into human perception of sound. For instance, it only became apparent later that although the human ear cannot directly pick up frequencies above 20 kHz they are actually of importance for the way we hear sounds. SACD with DSD extends the frequency range towards 100 kHz.
Actually no even back in 1982 when CD premiered in Japan it was a step backwards from the 50kHz Soundsteam process used by the American recording companies. CD was highly compromised from the very beginning. Technology has progressed and so has insight into human perception of sound. For instance, it only became apparent later that although the human ear cannot directly pick up frequencies above 20 kHz they are actually of importance for the way we hear sounds. SACD with DSD extends the frequency range towards 100 kHz.
A possibly more important difference between CD's 44.1 kHz 16-bit PCM sound quality and SACD's 2.8 MHz 1-bit DSD sound quality is the accuracy in the time domain. As it turns out, the human ear is extremely sensitive to minute timing differences. In fact, of the various cues our brain uses to determine the direction of sound sources, probably the most important cue is the difference in time it takes for a sound to reach our left ear versus the right ear. With a sampling frequency of CD, 44,100 times per second it's very difficult to reproduce a good 'sound stage', this is an area in which DSD excels. Furthermore, at 120 dB the dynamic range of SACD is much improved over CD. SACD also provides the option of multichannel sound where Audio CD only offers stereo.
What does it mean when an SACD is recorded in PCM?
The audio on the high resolution layer of an SACD is always DSD or DST, however the original recording may have been made using PCM technology, which is then converted to DSD for the SACD master.
What is DST?
DST is an abbreviation for Direct Stream Transfer, a losslessly compressed version of DSD. Lossless compression means every single bit from the original input stream is delivered at the output after decompression, it is said to be acoustically transparent however I have read that many Stereo only SACDs were released without using DST compression as the engineer's thought they sounded better that way and Exton's new HQ Stereo SACDs emphasize they are not lossless compressed. DST is used on multichannel SACDs in order to fit the whole program on a single disc. Lossless compression basically doubles the playing time a multichannel SACD would otherwise have.
What is DSD128?
DSD128 is DSD at twice the default sampling rate: 5.6448 MHz. DSD128 is used in some studios for editing. The normal DSD format is also called DSD64. Dr. Yoshio Yamasaki's the inventer of DSD whose's proposed Super DVD format was rejected by Sony used DSD128. At current there is not consumer format being proprosed that uses this double speed sampling rate.
What is DXD?
DXD is short for Digital eXtreme Definition, an encoding scheme for professional use that was developed for editing high-resolution recordings because DSD was orignally not ideally suited for editing, however this is changing, check out the newest Sonoma editing consoles. DXD is 352.8 kHz 24-bit PCM.
I acknowledge the help of Sony, Philips, the Stradivari Society, PS3, SA-CD.net and everyone else who supplied the needed answers to our FAQ's. This FAQ represents many years of research and is offered to all who want a clearer understanding of how SACD works. For more detailed explanations please download Sony's free PDF Super Audio Compact Disc: A Technical Proposal
Under the heading How does DSD work?, you stated the following:
ReplyDeleteSince a bit can only have two values, every bit in a DSD stream only tells weather the amplitude of the sound signal was higher or lower than the previous sample held in the buffer. Because it doesn't tell how much higher or lower the amplitude is, you need a lot of samples to accurately describe the signal.
[end of excerpt]
After I wrote the following, it occurred to me that you might have meant the same thing, but expressed it differently:
Although I don't know for a fact, it seems to me that DSD was designed so each step up or down is greater than the amount that the waveform being recorded can possibly change during a sample interval. If the waveform COULD change at DSD's limit, and it did so in a positive direction over a certain interval, then the corresponding data would consist of all 1's over that interval. But since the signal can't change that fast, any slope can be represented by a combination of 1's and 0's. If the waveform remained constant, the data would consist of alternating 1's and 0's; a positive slope would be represented by more 1's than 0's, and vice versa for a negative slope. All the converter knows is whether the input is greater or less than the value at the output of the integrator which "keeps track" of the value represented by the output-data stream.
This is one of the great benefits of using a one-bit ADC for audio: since the step size is constant, the linearity is (theoretically) perfect, and the only conversion error is quantization error, which is buried by noise/dither. So, small signals are preserved, and not affected by changing low-level nonlinearities across the amplitude range. This is probably why DSD sounds so much like good analog, which also has excellent low-level linearity across the amplitude range.
The illustration under the heading "How does DSD work?" gives the impression that DSD is always converted to analog via an analog low-pass filter ("audio direct," "source direct," or "DSD direct" mode), and you apparently forgot to mention on this page that this isn't always the case. Often, DSD is converted to PCM and converted to analog through a DAC, for various reasons, possibly to help conceal the DSD's advantages over PCM. Ideally, there would be a control to allow switching between these two modes, and the reasons for doing so would be explained. Some players have HDMI outputs, and you can make up for the lack of a switch by getting a separate processor (with an HDMI input) that will switch between direct and PCM modes, and much more.
ReplyDeleteThe direct mode offers the greatest transparency, since DSP is essentially already analog, but it doesn't allow any DSP to be performed (as of this writing) since there are no consumer-grade signal processors that can process DSD. So, if you want to perform any DSP, you need PCM mode. But at least with DSD, you have the choice between extreme transparency or performing DSP. With a PCM source, the extreme transparency simply isn't available, as can be seen by comparing the impulse responses shown in this blog. DSD's impulse response is almost indistinguishable from analog, whereas all other impulse responses are clearly different.
Based on some favorable reviews of inexpensive units, I suspect that DSD-direct is not limited to high-end units. But it's typically advertised only on high-end models, which might be intended to give well-heeled audiophiles the impression that they need the high-end unit or an external processor to get direct mode. It seems to me that Sony and Philips would be prone to offer source direct on their inexpensive units to demonstrate the advantages of their SACD format, which would be especially pronounced in an inexpensive unit. They could do a lot more to popularize the format by making it available at a more reasonable price.
The illustration shows how pure DSD works with NO PCM inter-steps before conversion to analog. Which is the ideal as DSD cures a lot of PCM's problems which are defeated if PCM is used anywhere in the chain from the microphones to the speakers.
ReplyDeleteIt is true not all SACD players keep the datastream pure DSD but convert to PCM. In the least expensive players this is for cost reasons, in the some megabuck players this is for design reasons. However most SACD players over $400 decoded DSD directly to analog with the analog RCA outputs.
The "Pure DSD" option has to do with HDMI not the analog outputs as the original specification of HDMI did not allow SACD out at all. HDMI was upgraded for high resolution PCM, thus DSD was converted to PCM. The latest version of HDMI allows DSD Direct. The "DSD Direct" button you refer to lets you choose DSD or PCM over HDMI to match your receiver or preamp. This has no effect on the analog out.
If you are interested in an SACD player and are unsure if it stays pure DSD until conversion to analog over the analog outputs search the internet or write to the manufacturer for confirmation.
You are correct DSP is not yet possible in DSD, to do so you must convert to PCM. There are analog signal processing devices and room neutralizing devices such as tube traps.