From what ive read on the internet and on google, DSD is able to capture sound at 1 bit. From what i understand, this is essentially a way to capture audio at the most atomic level??

I'm not an expert on audio but if im correct, assuming that someone has an DSD audio file, would they even be able to hear the difference between DSD audio vs high quality (320kb/s) audio at 24 bit depth?

I mean wouldn't you need specialised speakers to actually hear the quality of DSD to begin with, not to mention the right shaped room, high neural efficiency in processing sound as well as perfect functioning Cilia (hairs in the inner ear).

Can humans practically perceive this difference in audio quality, and is DSD analogous to say, the frames per second in reality vs how many frames the human eye can process?

  • "From what i understand, this is essentially a way to capture audio at the most atomic level??" Can you elaborate that?
    – bipll
    Jun 14, 2018 at 20:48
  • I mean does capturing DSD audio at a bit depth of 1-bit, the most data consumptive standard to record audio because you can't capture data of any type in units any smaller than 1-bit? Jun 14, 2018 at 21:13
  • Hmmm? Bitdepth is not a strict prohibition to capture lesser units, it's about something else, you know. One bit used to represent a membrane position means a huge lot of samples, oversampled at tens of thousands higher rate. When converted to analog this whole mess is lowpassed to produce a much smoother wave in the usual, audiable spectral zone.
    – bipll
    Jun 14, 2018 at 22:00
  • And—since it finds real use in e.g. SACD, it's hard to be deemed purely theoretical.
    – bipll
    Jun 14, 2018 at 22:04
  • what im saying is that yes, practically you can capture/record DSD audio. However in in September 2007, the Audio Engineering Society published the results of a year-long trial, in which a range of participants 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. Jun 14, 2018 at 22:10

1 Answer 1


Partial answer :

Many Analog to Digital converters are actually implemented using a one bit ADC at a high sampling frequency (several MHz). The bitstream is then decimated to more common formats like 24 bits / 48 kHz).

This has the advantage that you can have a cheap analog low pass filter with a cut-off frequency at half the high sampling frequency (given the fact that your input analog audio shouldn't contain a lot of frequency component over 1 MHz for example), and then implement a digital low-pass filter with steep attenuation at half the end sampling frequency. Implementing this filter in the digital domain is cheaper and more stable that in the analog domain.

As for the question to know if there's an audible difference between a DSD stream versus a 16 bits / 44.1 kHz one, it seems you have yourself find a study that indicates that there's no audible difference. The AES unfortunately is not willing to publish freely it's work, so I cannot comment on that study. Were there any bias or methodological errors ? The double blind test is a good clue that there are minimum biases. One might discuss which kind of audio content was submitted to the panel.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.