DSD Master is a tool for creating the highest possible quality PCM
conversions from DSD originals. We know of no way to substantially
improve upon this process. The converter can work with originals in
standard DSD (DSD64), DSD2 (DSD128), and DSD4 (DSD256). It works with
DSD files in DFF (DSDIFF) or DSF formats. It is expected to work with
multi-channel DSD files, but this aspect has not been tested.
quality PCM copies of DSD files are useful for many reasons, including
multi-system home environments where not all DACs support DSD, and the
transfer of DSD content to mobile devices.
DSD Master's algorithms are very intensive users of CPU power. It recommended not to use BitPerfect to play music at the same time as DSD Master is being used to make conversions.
This is the consumer
version of the product, and is licensed for consumer use only. Pro
Audio customers wishing to use DSD Master to make the highest quality
PCM files for commercial distribution should contact BitPerfect Sound
Inc. to obtain a commercial license.
DSD Master is a product of BitPerfect Sound Inc. If you ever encounter any problems with DSD Master, please send an email to firstname.lastname@example.org (support only available in English).
This User Manual is written for the v1.2.2 version of DSD Master.
Output File Formats
DSD Master will produce PCM output files in WAV, FLAC, AIFF, ALAC (Apple
Lossless), and "Hybrid-DSD" file formats. All standard PCM sample rates
can be selected, including 44.1kHz, 48kHz, 88.2kHz, 96kHz, 176.4kHz, 192kHz, 352.8kHz and 384kHz. Output files at all
sample rates are created with 24-bit depth, with the sole exception of
44.1kHz which is dithered down to 16-bits using a TPDF dithering
DSD Master is also able to produce a unique and useful file
format which we term a "Hybrid-DSD" file. Please read carefully the
separate section at the end of this User Manual on the Hybrid-DSD File
Format before using this option.
standard metadata contained in the DFF, DSF, or DST files will be
transferred to the resultant PCM or Hybrid-DSD files. If the DSD
originals contain no metadata, then there will be no metadata in the
created files. Be aware that not all file formats support the same
metadata standards, and that no standards exist for what information is stored as metadata, or, in some cases, how it is stored. For example, DFF and WAV files have very limited
metadata support capability, whereas FLAC, AIFF (& DSF), and ALAC each have very comprehensive but dramatically different methods of storing and defining metadata. Customers with exacting metadata
requirements are advised to use third party software (we like Yate) to groom the
metadata in the output files to their specific requirements.
How to Run DSD Master
the DSD Master icon on your desktop. Double-click on it to run it.
The DSD Master icon should appear in the System Tray and its Preferences Window will appear,
where you can set up how DSD Master behaves. If you close the Preferences Window, it will not cause DSD Master to quit. You can click on the
System Tray icon, and "DSD Master" will appear in the top left of the
Menu Bar. Clicking on this opens a drop-down menu which allows you to re-open the Preferences Window.
the desired configuration has been set (see later), there are four ways to
initiate conversion. The preferred way is to select the file or files
to be converted in a Finder window (multiple selection is supported). Right-click and select "Open With" and then "DSD Master".
Alternatively, just drag-and-drop your DSD files either onto the DSD Master application icon, or onto the DSD Master system tray icon to
initiate conversion. Finally, you can use the "File | Open" dialog from the Menu Bar to select files from a Finder Window.
Additional files can continue to be click-selected or
dragged-and-dropped, one at a time or in groups, and will be queued in
the background. Newly added files are queued even while previous
conversions are in progress. If you change the conversion parameters
after a conversion has commenced, it will have no effect on files which
are already "in the queue". The conversion parameters applied to each
file will be those which were current at the moment the file was queued.
Output Format. This is the desired file format of the PCM output file. You can select from AIFF, FLAC, Apple Lossless (ALAC), WAV, and Hybrid-DSD.
Output Sample Rate. This is the desired sample rate for the PCM output file. You can select any of the standard PCM sample rates from 44.1kHz all the way to 384kHz.
Do Not Compress FLAC. If you select FLAC as your output file format, this option produces a so-called "uncompressed" FLAC file (FLAC compression '0').
add to iTunes: This will cause the output file to be automatically added to iTunes. If
you do not have iTunes, or if you ask to add FLAC
files to iTunes, this will be ignored.
If you have checked “Automatically add to iTunes”, then the resulting file will be automatically
imported into iTunes. Depending on how
you have set up your iTunes Music Library, this may have unintended consequences. Many users allow iTunes to
manage its own Music Library, and have it set up to copy all tracks to its own
iTunes Media Folder as part of the import process. If that is the case, then you will end up with
unnecessary duplicate files on your computer. Since many of these files (particularly Hybrid-DSD files) are extremely large, this
can be very wasteful of disk space.
Therefore, in such cases, it is recommended to delete the converted
files from the output location you selected as soon
as they show up successfully in iTunes.
Additionally, for users who manage their own iTunes Libraries,
for various reasons prefer to create your output files in a different
folder from the one from which you wish to import them into iTunes. In
that case, you should not
set DSD Master to import them automatically into iTunes, and instead
them manually after you have moved them to their final location.
This is the folder in which the output files will be located. The output files will have the same names as
the DSD originals, but with the appropriate new extension. If an output file of the same name already exists in the specified output folder, conversion of that file will be abandoned.
Important Note: Due to restrictions imposed by Apple's App Store regulations, Hybrid-DSD files can only be output using the "Specify:" output location option (which can nevertheless be configured to be the same directory as the input file)
DSD files are supposed to follow the SACD protocol which requires that the music
signals be encoded 6dB down from maximum theoretically encodable (0dB)
level. This is because DSD encoders have the potential to go unstable
under certain conditions, and if the signal level is kept at least 6dB down from
the 0dB level this problem is effectively eliminated. However, the
onset of instability is graceful, and many DSD recordings can be found to contain
passages which the recording engineer has allowed to exceed this -6dB
limit without any obvious sonic penalty. Others, however, are
conservatively encoded, and never actually approach the -6dB limit.
Compare this to PCM, which also has a maximum encodable
limit of 0dB, but can make use of the full dynamic range, all the way up to 0dB. However, any attempt to drive the input signal above 0dB is
not treated gracefully and will be abruptly clipped. It is therefore very
important that PCM recordings should not be encoded with levels which
exceed 0dB. However, if the peak levels are encoded too conservatively,
this results in a failure to take full advantage of the dynamic
headroom offered by the format.
The theoretically 'correct' procedure when
converting DSD to PCM is to apply 6dB of gain to compensate for the
'SACD 6dB Level'. This is quick and easy to do, but will result in
clipping whenever the DSD recording was driven into its 'graceful'
overload state. Soft clipping is routinely employed to alleviate this, but at the cost of undesirable dynamic compression. The ideal solution is to read through
the file before conversion, and establish in advance what the exact
correct gain setting should be, so that the peak of the DSD recording is
encoded precisely at the 0dB level of the PCM format. The downside of
this approach is that it is slower. This process is called 'Normalizing' and is the approach taken by DSD Master.
DSD Master provides two strategies for automatically normalizing the
encoding level. It can be done either on a per-track basis, or
on a per-batch basis. On a per-track basis, each track is pre-scanned
prior to conversion to find the loudest signal level in the track. A
gain setting is then determined that will raise this point to precisely
0dB in the resultant PCM file. Done this way, each track will receive a precisely
tailored amount of gain, which may vary from track to track.
Sometimes, however, consecutive tracks are part of a contiguous
performance, and you don't suddenly want the gain to be lifted during a
quiet passage. For these situations, the setting "Normalizing over file groups"
is provided. Using that setting, the entire file group will be scanned to find
the loudest signal level in the group, and then a common gain setting is
applied to every file in the group. Therefore, only the loudest point
in the loudest file within the group will be encoded at 0dB in the group of PCM files.
Starting with v1.1 DSD Master now uses what we term "Analog Normalization" which means that the 0dB point we use is the level of the actual implied peak of the encoded waveform, and not just the largest sampled value (which will always be lower if it does not coincide precisely with the actual peak itself). This means that we avoid encoding so-called "Inter-Sample Peaks" which can cause audible defects (i) in certain DACs; (ii) if the PCM file is subject to downstream sample rate conversion; or (iii) if it is re-encoded as a compressed AAC file for synching with your mobile device.
more than one file is dragged-and-dropped as a group onto DSD Master,
these are considered, for volume normalization purposes, to be a single
"file group", regardless of their musical relationships to each other.
If there is some reason why you would wish to stipulate a specific gain, then this can be done using the "Manually specify gain (dB)" setting. However, note that if this induces clipping, then the clipping will be hard.
file conversion a progress window appears. Each track being converted
shows a progress bar with an X button on the right hand side. If the X
button is clicked at any time before the conversion is complete,
conversion for that file is abandoned. If the X button is clicked after
conversion is complete, it removes the progress bar for that completed
item from the progress window.
The progress bar is blue while the
conversion is in progress, green when the conversion has completed, red if the conversion has been abandoned by pressing the X button, and orange if the conversion could not be initiated in the first place. The conversion proceeds in
two stages, a first stage where the DSD file is pre-scanned to determine
the automatic gain level (if so selected), and a second stage where the
actual conversion takes place. The first stage is much, much slower
than the second stage but a lot of pre-processing is also performed in the first stage.
Above each progress bar the gain level is shown.
If the gain has not yet been determined, this will show as "N/A". If
the gain is greater than 6dB, then the difference is the unused dynamic
headroom of the original DSD recording. If the gain is less than 6dB,
then the difference is the amount of dynamic overload present in the
original DSD recording.
Hybrid-DSD File Format
of DSD files requires the use of dedicated playback software that is
able to read native DSD files (.DFF and .DSF files) and transmit
the music data to the DAC. It also requires the use of a specialist DAC
capable of playing music in the DSD format. At the time of writing,
the large majority of DACs do not have this capability, and this includes the
built-in outputs of all Mac computers. Software such as iTunes does not
support DSD playback at all, and so users of BitPerfect have not been
able to play DSD files natively, simply because the files could not be
imported into iTunes in the first place. This is a disappointing
limitation, because BitPerfect has internally long been able to play
Hybrid-DSD ("DSDh") is a new technology introduced for the first time here in DSD Master. A DSDh
file utilizes the Apple Lossless Audio Codec (ALAC) file format and has
the file extension ".m4a". It contains both PCM data and DSD data.
iTunes sees it as a regular ALAC file and will therefore import it without difficulty. During playback, BitPerfect recognizes that it is reading a DSDh
file, and if the selected audio output device supports DSD it plays the
DSD content, otherwise it plays the PCM content. BitPerfect version 2.0 or later is required to support this feature. iTunes, on the
other hand, interprets the DSDh file as an ordinary ALAC file and only ever plays the PCM content.
When creating a DSDh
file, DSD Master allows the user to select any available sample rate for the PCM
content, but the DSD content is always stored at its native sample
rate. For example, DSD128 files may not be converted to DSD64, or
have been designed to be compatible with any audio player which
correctly supports the Apple Lossless format. The Hybrid-DSD files
created by DSD Master are given a file name of the form
"filename.DSDh.m4a" where "filename" is taken from the original DSD
file. The ".DSDh" is inserted purely for convenience, to enable easy
identification of Hybrid-DSD files by the user. It has no other purpose, and can be
edited out manually by the user if preferred. Additionally, Hybrid-DSD files are created with the words "Hybrid File created by DSD Master" in the comments section of the file's metadata, as well as the words "DSD64", "DSD128" or "DSD256" according to the format of the DSD content. Again, this is for convenience only, and allows iTunes' "Smart Playlist" feature to identify DSD files. These comments may be edited out and will not affect playback in any way.