DSD (direct stream digital) is almost like a mythical creature that only certain audiophiles have confronted before.
These are people who come back from their adventures with DSD telling wild stories about how they heard this extra-terrestrial beast murmur a sound that nobody has ever noticed before (besides them).
If you’ve heard of DSD, you may be familiar with these stories. In this article, we’ll uncover if they are true or false; and we’ll separate reality from fiction.
But First, Some Primer
It may sound obvious, but most mythical stories are not real. If someone were to say that DSD sounds better than PCM, they could be right.
However, the difference between the two is splitting hairs when it comes down to sound quality.
The drawbacks with DSD are high, and the commercial viability of DSD is non-existent compared to PCM.
Before we look at why this is, we need to make sure we understand some of the fundamentals of digital audio first…
What Are Samples?
The common answer to this is usually something like “samples are pieces of music used in other pieces of music.”
While this is correct, in this context a sample is something much smaller than a piece of music.
A sample is the smallest piece of a waveform possible, describing the wave’s exact position at a precise moment in time.
When all these samples are put together, you get a sound wave. You can think of samples as like pixels but for audio.
In some DAWs and wave editors, you can zoom right in and see little dots on the waveform.
These are samples and you may even be able to move them around, though it’s very hard to edit audio this way unless you are removing a few peaks here and there.
What Is Bit Depth?
Audio is stored in computers with bits – 1s and 0s. So if you have some binary code such as 1001010101000110, since it has 16 digits, this means it has 16 bits.
So bit depth is simply a way to store more information per sample in an audio file in x bits.
- 16-bit means there are 65,536 different possible values per sample
- 24-bit offers more than 16 million possible values per sample
- 32-bit has more than 2.1 billion unique values
But it should be noted that 32-bit float is technically 24-bit with an 8-bit mantissa, and there are no major practical benefits to using 32-bit audio over 24.
What Is Sample Rate?
The sample rate is the number of samples in one second of audio. So if you have a 48 kHz sample rate, then 48,000 samples of the audio are measured per second to recreate the sound.
So, if you are digitally recreating a sine wave at 48 kHz, then you are measuring the amplitude of the sine wave at 48,000 different points per second.
The most common sample rates are 44.1 kHz and 48 kHz, which are used in most DAWs.
What Sample Rate Should You Use For Music?
What sample rate you should use will depend on the audio quality you prefer, but the differences are very hard to notice. 48 kHz is the safest option but most consumers can not tell the difference between 48 kHz and 44.1 Khz.
In this case, we look to other reasons to pick between 48 kHz and 44.1 kHz. Simply put, 44.1 kHz is a leftover from the CD era, and though it sounds perfectly fine, 48 kHz is more compatible with video formats.
From an engineering perspective, calculations involving 48 samples per millisecond are neater than 44.1 samples per millisecond.
If you want to use Max, Reaktor, or make your own VST plugins, this is worth considering.
What is PCM?
PCM stands for pulse-code modulation and is the standard digital audio format used to encode sound waveforms digitally.
With PCM, the amplitude of an audio signal is sampled at regular intervals, creating a waveform.
History
In 1937, Alex Reeves conceived of PCM for voice communication in Britain while working for a phone company in France.
He filed a patent in 1938, describing it in theory with its advantages, but there was no practical application that resulted.
PCM technology was mainly used as a better approach to send telephone calls in the early 1960s, but the quality was not good enough for high-fidelity music until the 80s.
It was finally brought into the consumer market in 1982 when the Compact Disc (CD) brought PCM to the mainstream market and took off for its convenience and fidelity.
What is DSD?
DSD is a niche, 1-bit audiophile digital audio format used to represent analog samples digitally in a different way than PCM.
Instead of sampling the amplitude of audio signals at regular intervals, each bit is an interval that changes depending on the amplitude of the signal.
Does DSD Have a Higher Sample Rate Than PCM?
Yes, DSD (direct stream digital) has a much higher sample rate of over 1000kHz. In comparison, PCM is between 44.1kHz-192kHz.
Does DSD Sound Worse Than PCM?
Not necessarily. The problem with DSD is that you cannot edit, mix, or master it because there’s no proprietary software to do any of these things.
DSD also uses extremely high levels of noise shaping, meaning that at lower frequencies (those we can hear) the noise is very low, but once you get into ultrasonic frequencies the noise level rises extremely quickly.
What Is A 1-Bit Format?
A 1-bit format is about as simple as it gets. Whereas, a 16-bit format has 16 different bits that could either be on or off at the same time, and the combined bits add more complexity.
History of DSD
In the early 1990s, Sony collaborated with Phillips who had begun developing the DSD technology, and together the two made the SACD format, which is the same as DSD but in a consumer format (like PCM is to CD).
However, while the two had finalized the development of the product, the industry had made a giant step forward.
DAC manufacturers switched from 64fs to 128fs, which is twice the sampling rate, as well as a 5-bit format rather than the 1-bit format that they previously created.
So, they were essentially faced with the dilemma of filming a black-and-white film right when color television was invented.
However, DSD didn’t pan out to be commercially viable enough to be used as a mainstream source of digital audio encoding.
PCM became the dominant analog-to-digital encoding format in the early 80s when CDs were invented.
What Are Some Drawbacks of Using DSD?
The problem with DSD is that it is noisier and offers a more limited frequency range than comparable 24-bit PCM of a sampling rate >88.2kHz.
As a 1-bit format, there isn’t enough space in DSD for it to be dithered properly. As a result, you end up with an elevated noise floor.
The same issue applies to the DAC designs that originally inspired the format.
Which Sounds Better? DSD or PCM?
A lot of controversies have been made over which encoding system sounds better, but one can never really know unless they hear for themselves.
Not really easy when almost every digital device uses PCM audio-only.
Furthermore, many argue that DSD is not suitable for high-end applications and high-resolution audio because of its high distortion, but there are still audiophiles who swear that DSD DEFINITELY sounds better.
How Do I Listen To DSD?
You can listen to DSD by using an external Digital to Analog Converter, or DAC. A DAC can handle the high sample rates that are used in DSD format and can be connected to your computer through USB.
You will also need software to be able to listen to DSD as well. Programs like HQPlayer and JRiver work for both Mac and Windows, while Audirvana works for Mac exclusively, and Teac HR Audio Player works for Windows exclusively.
Can You Do Post-Production With DSD?
There’s been no method to edit, mix, and master DSD files like you could PCM files.
Therefore, most “commercially available” DSD recordings are recorded directly to DSD with no mixing/mastering, or are simply converted to and from PCM!
There are a few new software packages that can mix, master, and edit using DSD, but most are from very small niche companies.
Why Does DSD Need To Be Converted To PCM?
PCM does not have noise in the higher frequencies like DSD so it allows non-linear processing effects such as saturation and distortion.
DSD also cannot be dithered like PCM audio, because dithering applies randomness, the 0s and 1s become meaningless and the result is simply noise.
What Is Dithering?
Dither works by adding a bit of random noise to the signal: the sample values are shifted around a tiny bit in an unpredictable way.
This gets rid of the nasty distortion that results when decreasing the bit depth from 24 to 16 or lower.
Even though this raises the noise floor, the result is still less harsh than without dithering.
Dithering is somewhat antiquated and should only be used in the final stage of the production process in mastering when preparing a 16 bit master. Do not apply dithering to anything else!
What Is Quantization?
With PCM, the amplitude of the signal is limited to just one of a set of fixed values, determined by bit depth. This limiting process is called quantization.
Dither is needed on the signal to avoid quantization distortion. For example, quantization occurs when a 24-bit recording (with 16 million possible values per sample), gets converted to 16-bit CD resolution (which has only 65,536 possible values).
What Are The Different Rates Of DSD?
DSD comes as the standard DSD64, double-rate DSD128, Quad-rate DSD256, and Octuple-rate DSD512.
The number at the end of the acronym signifies that it is 64, 128, 256, or 512 times the sample rate as a CD.
Are DSD Files Smaller than PCM files?
DSD records a 1-bit data stream at 2.88 MHz. This amounts to roughly 22 MB of disk space per minute. So, if you have a typical song around 3 minutes in length, that would only equate to 66 MB.
PCM files such as WAV use less space, however. For example, a 24-bit 48 kHz WAV file at the same length is about 40 MB while an MP3 file of that length would be about 4-8 MB.
Playback Options for DSD
What Is DSD Disc Format?
DSD discs are available through the use of specific recorders and appropriate tools.
These discs can be listened to via certain Sony audio hardware devices such as the PlayStation 3, as well as certain Sony laptops.
Can DSD Be Used With USB?
Yes, USB is an alternative to using discs with files burnt onto them for playback. In 2012, a lot of companies teamed up to develop a standard that detects DSD audio in PCM frames they titled DSD over PCM, or DoP.
DSD vs PCM
When comparing DSD vs PCM, it’s important to note that DSD audio has a higher noise floor than PCM audio, a more limited frequency range, and was based on an approach to DAC/ADC design that was substantially improved right after its release.
The other problem is it’s incredibly difficult to work with. In fact, to perform any kind of serious work with DSD, you have to convert it into PCM.
It really seems that in 90% of cases, your average SACD was recorded as PCM audio, mixed as PCM audio, and then converted to DSD audio.
Contradictory Results
The consistent theme when comparing PCM and DSD is contradictory beliefs and opinions, and this goes for scientific studies as well as personal tests.
One double-blind study in Germany found that hardly anybody could distinguish the difference between PCM and DSD audio.
However, in a 2014 study in Tokyo, the results concluded that listeners could distinguish 192 kHz, 24-bit PCM with DSD at 2.8 MHz and 5.6 MHz, but not between 2.8MHz and 5.6Mhz DSD.
Consumer Market
DSD never achieved any level of success in the consumer market because post-production (which includes editing, mixing engineering, and mastering engineering) is extremely difficult due to the lack of necessary software.
DSD is still used, however, as a format for studio equipment in an archival manner as a possible replacement for analog tapes.
Conclusion
The quality of production, mixing, and mastering in most cases is 99% in the music itself, with that last 1 percent arguably being reliant upon the encoding format (in this case, DSD vs PCM).
Producing music in DSD format is extremely tiresome and near impossible. So unless you have a point to prove, stick to PCM.
If you’re only interested in listening to music, then DSD is worth investigating if you’re really fussy about audio formats.
If we had to pick a clear winner between the two, it would be PCM, largely thanks to its wide compatibility.
FAQ’s
What is SACD Versus DSD?
SACD (Super Audio CD) uses DSD encoding, except it is compatible with CD as well to make it more commercially viable.
It was intended to replace the Compact Disc (CD) in 1999 when it was introduced, with its multiple audio channel functionality.
The SACD layer uses a 1-bit DSD with a 4.7 GB disc capacity while the optional CD layer uses a 16-bit PCM with a 700 MB disc capacity (just like a regular CD).
What Is DST?
DST (Digital Stream Transfer) is a lossless data compression method used to reduce space and bandwidth within DSD.
DST compression reduces the file size by twice or three times and carries eighty minutes of sound.
What is ADC and DAC?
ADC is an analog to digital converter while DAC is a digital to analog converter. They are basically what the name implies.
An ADC converts an analog signal into a digital signal, while a DAC converts a digital signal into an analog audio signal.
DACs and ADCs are commonly used with music players but are also seen in televisions and phones.
What is Nyquist Theorem?
The Nyquist Theorem states that a digital sampling system must have a sample rate at least twice as high as the highest frequency of the audio that is being sampled.
Simply put, this is because you need at least two samples to generate an oscillation.
So with a 48 kHz sample rate, repeating two samples of opposite values would create a 24 kHz tone.
This is why for high-quality audio we use at least a 44.1 kHz sample rate.
Because the highest frequency we can hear is 20 kHz, this gives us just
enough room to fit the entire audible spectrum of sound.
Source: https://producerhive.com/ask-the-hive/does-dsd-sound-better-than-pcm/
