Random String Records discloses the details of its releases by displaying codes of the engineering techniques used in rendering the digital audio file. Descriptions of techniques appear below. Codes are displayed in parentheses.
Random String Records defines the following release types :
Raw (R). This release type includes all the files used in the recording without change. Tracks may be mastered as mono or stereo and likely produce limited sound stage. File size is small. File format is likely MP3 or AAC with low bit rate and bit depth. No dithering is applied when rendering the release.
Streaming (S). This release type includes all the files used in the recording with limited track volume balancing and peak limiting. The file size is smaller and quality lower as determined by the streaming provider. File format is likely AAC with 44.1k bit rate and 128 kb/s bit depth.
Aggregate (A). This release type includes all the files used in the recording with some engineering. Typically, track volumes are balanced and peak limiting is performed. The sonic signature is analyzed, a sound stage is generated and tracks are mixed in stereo. File size is larger and quality higher than the Streaming release type. File format is likely MP3 with modest bit rate and bit depth, likely 48k and up to 320 kb/s respectively. Dithering is applied when rendering the release.
Note : MP3 is a deprecated audio file encoding superceded by modern ISO-MPEG codecs with higher audio quality at lower bit rates like AAC for streaming. The Fraunhofer Institute terminated MP3 licensing in 2017. We offer but do not endorse MP3 as a Hi-Res MUSIC encoding.
Curated (C). This release type includes all the files used in the recording with significant engineering. Typically, track volumes are balanced and peak limiting is performed. The sonic signature is analyzed, a sound stage is generated and tracks are mixed in stereo, quad, or ambisonic. File size is large. File format is likely FLAC or WAV with high bit rate and bit depth. Audiophile quality dithering and noise shaping are applied when rendering WAV files. Dithering or noise shaping may optionally be applied to FLAC files.
Some Background. Recording digital audio requires capturing discrete snapshots of the continuous stream of music. It is widely accepted under the Nyquist-Shannon Sampling theorem that under limited circumstances the continuous stream of music can be reconstructed by interpolation from the snapshots given a sufficient sampling rate. The Whittaker-Shannon Interpolation Formula establishes the limits of reconstruction. Perfect reconstruction is successful only when the snapshots have a band-limit less than the Nyquist frequency, otherwise frequencies above the Nyquist frequency cause distortion below the Nyquist frequency through a process called folding. Further, Chueng and Marks showed that an arbitrarily small amount of noise renders reconstruction of the continuous stream unstable. Read the original paper called Ill Posed Sampling Theorems, Kwan Cheung and R. Marks, "III-posed sampling theorems," in IEEE Transactions on Circuits and Systems, vol. 32, no. 5, pp. 481-484, May 1985, doi: 10.1109/TCS.1985.1085735.
Dithering is the process of introducing controlled noise into a reconstructed continuous stream of music reducing harsh, inharmonic quantization distortion of an imperfectly reconstructed continuum.
Noise shaping filters the distortion resulting from quantization and dithering by passing it to frequencies such that the distortion is lower at higher frequencies.
We apply dithering and noise shaping to some, but not all releases. We encode the release with the following codes to disclose the dithering and noise shaping used in the release, if any.
None (N). No dithering or noise shaping are applied to the release.
Narrow Band (NB). Narrow band dithering is applied which adds some low level distortion.
Wide Band (WB). Wide band dithering which may resemble analog hiss is applied.
