Digital Release - Album Publishing.

Pyramix V12 (and above)

Master Settings

In the Generate Master window, the Image Format choice is important depending on the files you need to generate.

• Mixer Output => PMI 44.1 kHz 16 bit file (RedBook compliant) => Post Processing (including sampling rate conversion and dithering).
  That means that all subsequent Album Publishing files are generated using a 44.1 kHz-16 bit file.
  If your project is not running at 44.1 kHz, SRC Filter type option is available to generate the CD Image file, choose between Linear Phase (Default), Minimum Phase or Apodizing. 
• Dithering : Use Mixer settings refer to the Dither option set in the Mixer for the selected Bus.  Use 16 bits WLC/Noise shaper can be set to 1st/8th/49th order or no dithering.  This obviously bypasses the dither setting in the Mixer for the selected Bus. 
• Stereo file output only.

• Mixer Output => MTFF at the selected sampling rate => Post Processing (including sampling rate conversion and dithering if required).
  

• If your project is not running at the same sampling rate as the selected Sample rate, the SRC Filter type option is available to generate the file, choose between Linear Phase (Default), Minimum Phase or Apodizing.

• Resolution : 24 bit fixed / 32 bit float.

• Mono, stereo and multichannel file output.

• Lossless compression : only available for linear 24 bit PCM resolution, so not for 32 bit floating point PCM.
  

• For Mixdown Mode : Mixer Output => DSDIFF in DSD64 / 128 / 256 (involves DXD / DXDx2 / DXDx4 conversion) => Post Processing (including sampling rate conversion and dithering if required in Album Publishing).
• For DSD Rendering Mode : Timeline files rendering (bypassing the mixer) => Post Processing (including sampling rate conversion and dithering if required in Album Publishing).
  See Pyramix user manual for further details on the DSD Render feature.
  

  The Process DSD at original Sample rate option allows performing DSD Rendering processing (Gain and crossfades) at the original DSD material sample rate (i.e. 64 Fs for DSD64, 128 Fs for DSD128, 256 Fs for DSD256) (available since V12)
  

• If your project is in PCM, the SRC Filter type option is available to generate the file, choose between Linear Phase (Default), Minimum Phase or Apodizing.
  Only Mixdown mode is available.
  
• If your project is in DXD, you can choose between Mixdown mode (Mixer output), or DSD Rendering (if the audio files on your timeline are in the corresponding DSD type).
• If your project is in DSD64, 128 or 256, only DSD Rendering mode is available, set in the corresponding DSD type.
• In Mixdown mode, Sigma Delta Modulator let you select the required SDM (D or Trellis E).
  Note that Trellis E encoding requires more time than the default SDM D.
  Pre-processing gain can also be applied (to allow taking advantage of the extra [0, +3] dB DSD dynamic range allowed by the DSD domain)
• Stereo or Multichannel files (depending on the Area selected).

Post processing

Album Publishing

• Sampling rate conversion and dithering can be selected for each output format.
  Extra Gain parameter available in the Album Publishing output format is to be applied before Sigma Delta (DSD output format) or after SRC (PCM output format)

• Mono, Stereo or Multichannel files (depending on the target format). 

DST Encoding (DSD Edited Master only)

• Strategy can be chosen from :

  Same for all channels (00 / 000000)
  Stereo pairs (01 / 001233)

  Different for all channels (012345)

Album Publishing Processing Flow diagrams

Dotted line processes are only used if required.
As of Pyramix V12

PCM to PCM

Gain is only available if SRC and/or dithering is present.
SRC and Dithering modules are bypassed if not required (No Change)
Data conversion integer <-> float is used when the destination format does not support float bit rate.

DXD to DSD

DSDx to DSDx (same DSD rate)

DSDx to DSDx (same DSD rate) using MECO SDM (available since v12) with Gain (Extra Gain is different from 0.0 dB) adjustment 

(n = 64, 128, 256)

DSDx to DSDy (DSD format conversion)

1.With SDM D or Trellis E.

2.With MECO SDM (available since v12)

DSD to PCM 8Fs

Album Publishing Encoding dialog

Source summary displays information about the source datas, including Artwork, output jobs,...
At the end of the Album Publishing files creation, a peak metering is performed automatically.
See Pyramix User Manual for further details. 

FAQ

The Peak and TruePeak level are not the same between files with different frequencies

If you do a sampling rate conversion, you can expect to have different peak and TP values between the source file and the converted ones. 
Especially when upsampling, new samples are created, and therefore you can get more inter-samples peaks, peaks that were between two samples before the upsampling process.

Why is there a 32 bit Float and a 24 bit Linear resolution for PCM Digital Release ?

DSD production goes over 0dBFS PCM, which requires 32 bit float to describe. If you attempt to create DSD files from a 24 bit source file, when the program goes over 0 DSD (above Full Scale PCM) there will be clipping in the file.  This was introduced to allow publishing DSD files taking advantage of the extra [0, +3] dB DSD dynamic range allowed by the DSD domain.

Can I use the Digital Release 32 bit to deliver PCM files ?

Of course, but if you need to deliver 24 bit PCM file, please check the dithering settings.

When producing DSD files that does not match the source DSD format (DSD128 to DSD64 for example), is there any PCM conversion ?

With the SDM D or Trellis E, there is a conversion to DXD, DXD x2 or DXD x4 depending on the source / destination DSD formats.
The new Meco SDM (available in the SDM section of an output format) allows high sample rate (64 and 128 Fs) intermediate multibit stage when publishing from a DSD format to another (e.g. DSD256 to DSD128)
If the source and destination DSD formats are the same, there is no PCM conversion.

When producing one file per audio track DSD files, does the process introduce any fade to eliminate/reduce clicks when the newly created tracks are played separately?

If the "Include Pause " option is ticked, the SDM is reset for each track.  If this option is not ticked, the SDM is not reset and used in continuity accross all audio tracks.
No additional fades are introduced, the DSD Media Player should handle such case.

When are the metadata written to files during the Publishing process ?

The Metadata are inserted at the very end of the Album Publishing process.
Once each file has been successfully generated, each file is re-opened to insert the « text » metadata, and re-opened a second time to insert the cover artwork if present and allowed by the format.

What is the difference between DSD Edited Master (DSDIFF) and DSD Digital Release (MTFF)?

DSDIFF Edited Master is for SACD mastering and additionally contains markers (the TOC), some limited metadata, and a unique ID.
Since MTFF is not really supported outside Merging ecosystem, we allow generating DSDIFF Edited Masters at higher rates than SACD's 64Fs, so other software can read it.
Even if it does not really makes sense since SACD is DSD64 only, we even extended the DST encoding to 128/256 Fs for DSDIFF Edited Master, after noticing that some other software player can actually support them. That also could be handy for lossless archiving of huge DSD256 media.

We allow the users to use DSDIFF Edited Masters as input for Album Publishing since it is quite standards for old SACD mastering projects.

But, since DSDIFF has been designed for professional purpose only, much like raw DDP without CD-text, it is too limited in terms of  metadata (compared to DSF embedding ID3 tag) and is limited to ASCII character set for text metadata, so we encourage people to use MTFF for their master, since it covers all needs:

 - supports PCM data
 - supports DSD data
 - can contain authoring metadata (Performer, Artist, song titles, ISRC codes and so on)
 - can contain a Table of content (so it can describe a full disc)
 - can embed artwork
 - uses Unicode (UTF16-LE) character set for text metadata, so it can virtually support any language
 - supports up to 8 channels

Appendices

Album Publishing file naming

Authoring - checking your work

File formats summary

• MTFF : The MTFF Digital Release Master file supports sampling rates up to 352.8 kHz (PCM) and DSD64, 128 and 256.
  But the Album Publishing individual files are limited to 352.8 kHz.
  
  

Previous versions

Image formats

In the Generate CD Image window, the Image Format choice is important depending on the files you need to generate.

• Mixer Output => PMI 44.1 kHz 16 bit file (RedBook compliant) => Album Publishing process (including sampling rate conversion and dithering).
  That means that all subsequent Album Publishing files are generated using a 44.1 kHz-16 bit file.
  If your project is not running at 44.1 kHz, SRC Filter type option is available to generate the CD Image file, choose between Linear Phase (Default), Minimum Phase or Apodizing. 
• Dithering : Use Mixer settings refer to the Dither option set in the Mixer for the selected Bus.  Use 16 bits WLC/Noise shaper can be set to 1st/8th/49th order or no dithering.  This obviously bypasses the dither setting in the Mixer for the selected Bus. 
• Stereo file output only.

 
 (Since V11.1)

• New Digital Release 32 bit Floating point PCM support (For DSD delivery)

• Create a DSD Digital Release in 32bit MTFF to allow for the transmission of +0dBFS information for DSD delivery. 
  This was introduced to allow publishing DSD files taking advantage of the extra [0, +3] dB DSD dynamic range allowed by the DSD domain.
• Mixer Output => MTFF at the same sampling rate as your project => Album Publishing process (including sampling rate conversion and dithering).
  All subsequent Album Publishing files are generated using your project Native sampling rate.
• Sampling rate conversion and dithering can be selected for each output format.
  Extra Gain parameter available in the Album Publishing output format is to be applied before Sigma Delta (DSD output format) or after SRC (PCM output format)
• Stereo or Multichannel files (depending on the target format).
  
  

Mixer Output => DSDIFF in DSD64 => Album Publishing process (including sampling rate conversion and dithering). 
Stereo or Multichannel files (depending on the Area selected and target format).

Processing Flow diagrams

Dotted line processes are only used if required.
As of Pyramix V11.1 :

PCM to PCM

Gain is only available if SRC and/or dithering is present.
SRC and Dithering modules are bypassed if not required (No Change)
Data conversion integer <-> float is used when the destination format does not support float bit rate.

DXD to DSD

DSDx to DSDx (same DSD Fomat)

DSDx to DSDy (DSD format conversion)

DSD to PCM 8Fs

Album Publishing Encoding dialog

Source summary displays information about the source datas, including Artwork, output jobs,...
At the end of the Album Publishing files creation, a peak metering is performed automatically.
See Pyramix User Manual for further details. 

FAQ

The Peak and TruePeak level are not the same between files with different frequencies

If you do a sampling rate conversion, you can expect to have different peak and TP values between the source file and the converted ones. 
Especially when upsampling, new samples are created, and therefore you can get more inter-samples peaks, peaks that were between two samples before the upsampling process.

Why is there a 32 bit Float and a 24 bit Linear version ?

DSD production goes over 0dBFS PCM, which requires 32 bit float to describe. If you attempt to create DSD files from a 24 bit source file, when the program goes over 0 DSD (above Full Scale PCM) there will be clipping in the file.

Can I use the Digital Release 32 bit to deliver PCM files ?

Of course, but if you need to deliver 24 bit PCM file, please check the dithering settings.

When producing DSD files that does not match the source DSD format (DSD128 to DSD64 for example), is there any PCM conversion ?

As of today there is a conversion to DXD, DXD x2 or DXD x4 depending on the source / destination DSD formats.
If the source and destination DSD formats are the same, there is no PCM conversion.

When producing one file per audio track DSD files, does the process introduce any fade to eliminate/reduce clicks when the newly created tracks are played separately?

If the "Include Pause " option is ticked, the SDM is reset for each track.  If this option is not ticked, the SDM is not reset and used in continuity accross all audio tracks.
No additional fades are introduced, the DSD Media Player should handle such case.

When are the metadata written to files during the Publishing process ?

The Metadata are inserted at the very end of the Album Publishing process.
Once each file has been successfully generated, each file is re-opened to insert the « text » metadata, and re-opened a second time to insert the cover artwork if present and allowed by the format.

About the Dither, when and how should it be used?

If SRC then use the SRC in dither in Mixdown options and disable the Mixer one.
If Direct Mixdown then use the one in the Mixer.
If you use the Album publishing its better to have the MTFF Generate Master that is in 32bit float and from this one publish output formats that can have dither individually applied.

Note: All those are done in one pass, it's important not to SUM dithering such as the Mixer Dither with the SRC one.