Zaphod’s Guide to PGGB.

Choose a mode and appearance.

Choose Offline in the header to upscale your library, or Live for real-time playback. Stop a live session before switching modes. DAC selection is available in both modes under Settings → DAC → DAC Selector, subject to the current device and license limits.

Choose Light or Dark in the application settings. A new installation starts with the theme that matches your operating system. The sun and moon buttons in the header select Light or Dark. There is no switch to the old layout.

What PGGB does.

PGGB remasters audio files by converting them to a higher or lower sample rate, i.e., upsampling or downsampling. Its algorithms are designed for reconstruction accuracy close to the theoretical limit. It can also apply EQ filters to adjust tonal balance or apply Crossfeed and Room effects for use with headphones. PGGB converts between two digital audio formats: PCM (pulse-code modulation) and DSD (Direct Stream Digital). It saves PCM as WAV by default and DSD as DSF. WavPack and FLAC provide additional output options. Metadata, including cover art, is transferred where possible.

The PGGB algorithm.

PGGB has two processing engines. The Precision Optimized Upsampling Engine, introduced in V9, is the default and the one we recommend. The earlier Nyquist-Shannon engine remains available for whole-track processing. Both use PGGB's reconstruction approach, but they differ in processing speed and memory use. Read about the PGGB engine ↗

Nyquist-Shannon engine. This engine considers the track as a whole for upsampling rather than using a fixed-length windowed sinc filter. With high-frequency noise filtering disabled, it preserves the original samples and creates intermediate samples by time-shifting the original data. The approach is non-apodizing and aims to approach the reconstruction limit for the track's length and sample rate. As these increase, it approaches pure sinc interpolation. For the mathematical background and a comparison with windowed Whittaker-Shannon interpolation, see The PGGB Philosophy.

Precision Optimized Upsampling Engine (the default). Introduced in V9, the default engine is an efficient reformulation of the Nyquist-Shannon upsampling engine. It is a true breakthrough in both upsampling quality and processing speed. It builds a processing pipeline tailored to the selected precision and the input and output sample rates. Within the passband (the frequency range it preserves), reconstruction accuracy is limited by the selected precision. It uses much less RAM and is considerably faster than the legacy engine. Listeners describe more natural timbre, clearer instrumental separation and a deeper sense of space. See Precision Optimized Upsampling Engine for its settings. Read about the PGGB engine ↗

The aim of higher reconstruction accuracy is to preserve the recording's timing, detail and tonal character. Listeners describe improvements in depth, layering, transients and timbre; the audible result depends on the recording and playback system.

High precision computing.

The ‘Plus’ in PGGB Plus stands for PCM plus DSD, with processing precision up to 256 bits. High precision calculations are used throughout resampling, noise shaping, DSD modulation, EQ and crossfeed. PCM precision choices are 64, 107, 128 and 256 Bit. Higher precision improves numerical accuracy but takes longer to process. This is the precision of PGGB's internal calculations, separate from the output file's bit depth. For DSD, PGGB chooses the working precision automatically.

See the processing precision menu

Menu preview · expanded to show choices; availability depends on license and settings · select to enlarge

How processing works.

This guide focuses on offline remastering: PGGB writes processed files for you to play later. For processing during playback, use the PGGB-Live tab and follow the PGGB-Live guide.

PGGB processes supported audio files in each input folder and its subfolders. For example, if you select C:\Music, it also processes albums in C:\Music\Album-1 and C:\Music\Album-2. Supported extensions are .wav, .flac, .aif, .aiff, .m4a and .dsf. The output follows the same album-folder structure, and PGGB also copies JPG, PNG and PDF files to the destination folders.

Choose an Input folder and an Output Folder; both are required. You can browse for folders or drag and drop them. Press Start to begin, Pause to pause or Stop to end processing. Pause and Stop can take a few seconds to a few minutes to take effect, depending on the current task. Add more input folders with Add Folder or by dropping them into the Input box. Select folders and press Remove Folders, or drag them out, to remove them. Each folder's checkbox controls whether it is processed. PGGB also discovers new subfolders added during a batch and processes them after the current set.

Interface overview.

Use the Live and Offline controls in the header to choose playback or file processing. Open Settings for processing options. Live playback is covered in the PGGB-Live guide.

For an illustrated introduction to the controls, see Explore PGGB App. Each item links back to the relevant section of this guide.

About: Open Settings > Application > License and About for the application version, license information and activation controls.

Status bar: The bottom of the PGGB window shows the current configuration on the right and activity on the left. During processing, it includes track details, CPU activity and memory use. When a batch finishes, it reports how many tracks completed, were skipped or failed.

Output Folder: The destination for processed files. Browse to a folder or drag and drop it into the field.

Temp Folder: The location for temporary files. If this is empty or does not name a valid folder, PGGB uses the output folder.

Write Mode: Create New creates a new album folder if that album already exists in the output. Skip leaves existing processed tracks in place. Overwrite replaces existing output tracks.

Output File Format: PCM output supports WAV, FLAC and WavPack; DSD output supports DSF and WavPack. The defaults are Wav (4GB) for PCM and DSF for DSD. Choose Wav (RF64) if your player supports it and you want files larger than 4 GB to remain unsplit. Otherwise, use Wav (4GB). FLAC and WavPack offer lossless compression.

File Suffix: Adds a description of the processing to each output filename. Choose None to omit the suffix.

Album Tag: Adds [PGGB] to the album metadata by default so you can identify processed albums in your player. Choose None to keep the original album tag.

Folder Suffix: Output album folders use the original folder names by default. Set a suffix if you want to distinguish folders created with different processing settings.

Advanced.

For larger batches, create pggb_to_process.txt in an input folder and list additional folders to process. PGGB includes their subfolders as well as the contents of the selected input folder. By default, output uses the parent folder's name. Put fqn, meaning fully qualified name, on the first line to reproduce the full input folder structure in the output.

Example pggb_to_process.txt:

fqn

"Z:\Native\PCM\Jazz\Ella-And-Basie"

"Z:\Native\PCM\Jazz\Ella-Fitzgerald"

Windows may ask for firewall permission when PGGB first runs. Allow access for trial licensing, activation and PGGB-Live.

DAC.

• High-rate PCM and DSD: DACs accepting PCM at 705.6/768 kHz or higher, or DSD256 or higher, are good candidates for PGGB remastering. A DAC converts digital audio into the analog signal used by your amplifier or headphones.
• PCM NOS: A non-oversampling (NOS) mode, or an option to disable the DAC's oversampling filter, lets PGGB perform more of the reconstruction work. This is available on some R2R DACs.
• Pure DSD: A DAC with a pure DSD path is a good candidate for PGGB's DSD output. Enable that mode if your DAC provides it.
• A DAC limited to DXD-rate PCM or DSD128 can still benefit. For example, try converting CD-quality tracks to 352.8 kHz PCM or DSD128.
• For a PCM-only DAC, convert DSD files to a PCM rate the DAC supports.
• For a DAC limited to 96 or 192 kHz, downsample higher-rate PCM or DSD files to a supported rate.

System, OS and Memory.

64-bit Windows 10 or 11, or macOS 12 Monterey or later. Apple Silicon and Intel Macs are supported. Intel and AMD processors must support AVX2. For Offline PCM up to 256-bit precision and DSD up to DSD256, start with a 4-core processor and 16 GB RAM. For DSD512, we recommend at least an 8-core processor and 24 GB RAM. DSD1024 and DSD2048 benefit from more cores and 32–64 GB RAM. A fast internal SSD helps with larger libraries. Your chosen settings and desired processing speed determine how much additional performance is useful.

Precision optimized upsampling (the default). With Precision optimized selected, memory use does not grow with track length and paging space is rarely needed. For Offline PCM up to 256-bit precision and DSD up to DSD256, start with a 4-core processor and 16 GB RAM. Read about the PGGB engine ↗

• PCM conversions: PGGB uses available processor cores when memory permits. For precisions from 64 to 256 bits, more cores can shorten processing time. Once sufficient memory is available, a faster processor and faster RAM matter more than adding RAM for large batches.
• DSD conversions: The output rate and modulator choice determine much of the processing load. An 8-core processor with 24 GB of RAM is a useful guideline for conversions up to DSD512. DSD1024 and DSD2048 benefit from more cores and 32–64 GB of RAM.

Selecting Legacy uses the legacy Nyquist-Shannon engine. This engine works with the whole track at the selected precision, so longer tracks and higher DSD rates require more RAM or paging space. Paging uses temporary disk storage for data that cannot fit in RAM.

Internet access: PGGB needs internet access for installation, the trial, license activation and PGGB-Live. Allow the firewall request on first use. PGGB does not use or share your personal information or use idle processor time.

The memory behavior below applies to the Nyquist-Shannon engine. With the default Precision Optimized Upsampling Engine, memory use remains independent of track length and paging is rarely needed. Read about the PGGB engine ↗

PGGB Plus improves the legacy engine's memory handling and paging, allowing more work to fit within a given amount of RAM than earlier versions.

The legacy engine adapts its processing to the available RAM and uses paging for data that does not fit. Its memory-management approach preserves reconstruction accuracy at the selected precision, but paging can increase processing time. The optional DSD block-processing mode described later has a separate quality trade-off.

The legacy engine still needs enough RAM to hold the input data. Its memory requirements depend on the following:

• Track length.
• Output format and sample rate.
• Working precision: 64, 107, 128 or 256 bits.
• Input sample rate. Higher PCM input rates increase memory use slightly. For DSD input, memory use depends on track length, output rate and working precision rather than the source DSD rate. For example, 10-minute DSD64 and DSD256 tracks need the same memory when processed to the same output rate and precision.

PGGB-Live uses calibration to measure the rates your computer can sustain in real time. The legacy-engine paging guidance does not apply to live playback. See the PGGB-Live guide.

Quick Start: DSD Configuration.

1. Choose a DSD rate

Start with the highest DSD rate supported by both your DAC and player, then compare lower rates if processing takes too long or you prefer their sound. PGGB supports output up to DSD2048. With the Legacy engine, single-stage processing offers the highest reconstruction accuracy and takes longer. Two-stage processing first upsamples to an intermediate rate, then completes the conversion to DSD. A higher first-stage rate improves reconstruction accuracy but takes more time. The final DSD rate determines file size; the number of processing stages does not.

Rate labels use fS for a base rate of 44.1 or 48 kHz. For example, 16fS is 705.6 or 768 kHz. As a rough speed guide, if your system handles 16fS PCM at 256-bit precision comfortably, under the Legacy engine, try DSD512: 32fS X 16 for two-stage DSD512. Its first stage runs at 32fS and the second multiplies that rate by 16. Conversion speed is expected to be similar, though it depends on your settings and computer.

2. Choose a DSD modulator

Choose from Low Noise, Very Low Noise and Ultra Low Noise modulators. We recommend starting with the default Very Low Noise option. Higher-order options can reduce quantization noise further but require more processing; the Ultra Low Noise 25th-order modulator also uses higher precision and more memory. The modulators adapt to the output rate. For Very Low Noise and Ultra Low Noise, the order can also adapt, so the menu shows the highest possible order. Compare lower orders if you prefer their sound or processing speed. See DSD Modulator Performance for details.

3. Choose the DSD rate family

DSD output can use the 44.1 kHz or 48 kHz base-rate family. Choose Auto if both your DAC and player support 48 kHz-family DSD rates. Otherwise, keep the default, Always output 44.1kHz rates. PGGB can convert between the two families.

Leave the remaining settings at their defaults. Choose Offline in the header. Under Settings → Offline, choose Input folders and Output files to set the folders, then close Settings and press Start.

4. Set up paging (Mac, legacy engine only)

Skip this step if Precision optimized is selected, as it is by default. The Precision Optimized Upsampling Engine disables the Optimize Paging and Memory controls and does not require this paging setup. Read about the PGGB engine ↗

For DSD processing with the legacy Nyquist-Shannon engine on Mac, enable paging and select a folder for the page files. We recommend at least 512 GB of free space. Paging lets the engine process data that exceeds available RAM. See Page File (Mac) for setup details.

Paging is used only when enabled. The folder browser opens at HomeFolder/PGGB/cache by default.

---

Between PGGB and Your DAC.

For the best results, send PGGB-remastered files to your DAC without additional digital processing. This is often called bit-perfect playback: the player passes the audio data through unchanged. Where your DAC allows it, use a mode that minimizes its own processing.

This section covers playback of files you have already remastered. You can use your usual player or PGGB-Live in Bypass mode. PGGB-Live can also process music as it plays and send the result through WASAPI, ASIO, Core Audio or a network renderer. See the PGGB-Live guide for that setup.

Setting up your DAC.

• Connect the playback device directly to the DAC where possible. Re-clockers or USB regenerators are compatible with this approach if they leave the audio data unchanged. Avoid adding another upsampling stage.
• Set PGGB's output rate to the highest PCM or DSD rate supported by your DAC and playback path.
• For PCM, choose the bit depth appropriate to your DAC, rather than relying only on the audio driver's maximum. Some R2R DACs have different preferred bit depths at different sample rates.
• If available, enable NOS mode for PCM or pure DSD mode for DSD.

Your playback software:

Disable processing in the player when you want bit-perfect playback of remastered files. This includes volume adjustment, dither, noise shaping, resampling and bit depth reduction. Padding to a larger bit depth is acceptable because it preserves the audio values. If you need software volume control, use one of the approaches below.

Software volume control

Software volume control changes the audio values after PGGB has processed them. For PCM, returning the adjusted signal to the output bit depth introduces quantization error, the rounding error from representing values with a limited number of bits. DSD volume adjustment generally requires remodulation. This can change the very low noise floor produced by PGGB, so the choice of playback processing matters.

If you need software volume control, choose an approach suited to your output format and whether you adjust the level regularly.

• For PCM:
  • 64-bit PCM for further processing: Save PGGB output as 64-bit doubles, then apply volume control in your player and output at the required bit depth. Enable the player's noise shaping or dither if available. This preserves more precision for later adjustment but creates larger files.
  • 32-bit PCM alternative: Save 32-bit output with PGGB noise shaping. Apply volume control in your player, then use its noise shaping or dither when reducing to the final output bit depth, if supported.
  • Fixed gain in PGGB: If you normally use one digital volume setting, apply that reduction during remastering instead. For example, set a fixed gain of −3 dB in PGGB and keep noise shaping enabled. During playback, disable the player's dither and noise shaping. Begin with the player's volume reduced, then gradually bring it to 0 dB while checking the listening level. The attenuation is already in the processed file.
• For DSD:
  • Fixed gain in PGGB: If you normally use a fixed digital level, apply that attenuation when creating the DSD files. For example, set PGGB's fixed gain to −3 dB. Begin playback at a reduced level and gradually bring the player to 0 dB while checking the listening level. Then disable further processing or enable direct SDM playback, as supported by your player.
  • Adjustable volume in the player: Enable the player's volume control and choose a suitable DSD modulator for the output rate. This allows ongoing level adjustment but adds another modulation stage.

PGGB-EQ.

Open Settings → Personalize Sound → Equalization to apply equalization. EQ adjusts the balance of frequencies in the recording. You can import a convolution filter as a .wav file or parametric filter settings as a .txt file. PGGB also accepts .zip packages of convolution filters for use without modification.

Bring your own EQ.

PGGB applies filters you supply. Choose or create the correction for your room or headphones, then import it into PGGB.

• Digital room correction: Use measurement software to create a correction filter for your listening room, or have a specialist prepare one. Import the stereo filter in .wav format. PGGB applies the filter; it does not perform the room measurements.
• Headphones: Choose one of these sources for headphone EQ:
  • Import your existing headphone EQ filters if you already have a set you prefer.
  • To use the AutoEQ project, select your headphones and choose Equalizer APO Parametric EQ as the export format. Download the .txt file and import it into PGGB.
  • Alternatively, find your headphones in the AutoEQ results and import the 44.1 or 48 kHz minimum-phase .wav filter.
  • Choose either a parametric text file or a convolution file for the correction you want to apply.

Convolution EQ: import and processing options.

A convolution filter stores its response as an audio file. A parametric filter describes the correction using frequency, gain and bandwidth settings. In either case, the filter determines the tonal change, so start with a correction appropriate to your room or headphones.

Convolution EQ from a .wav file: Press Browse beside Convolution filter and select a mono or stereo filter. PGGB derives new filters from the imported response for the input rates it processes, rather than simply resampling the filter file. A mono filter applies the same correction to both channels. Enable Linear phase for linear-phase filters, or leave it unchecked for minimum-phase filters.

Convolution EQ used as supplied: To preserve the filters without modification, import a .zip file containing .wav filters at all required sample rates. Use this route when the supplied filter's timing or phase correction needs to be retained.

Import Convolution EQ directly

Select Frequency Range.

The frequency-range controls apply when importing a single .wav filter. They let you limit EQ to a selected range by setting its start and end frequencies. Leave the defaults if the filter already covers the range you want. These controls do not apply to .zip packages used without modification.

Import Parametric EQ Directly.

For parametric EQ, export filters from REW as unformatted text, use an Equalizer APO filter file, or export Equalizer APO Parametric EQ from AutoEQ. You can also use PGGB's simple text format: start with #pggb, then list Fc Gain Q values separated by spaces. Fc is the center frequency, Gain is the level adjustment and Q controls bandwidth. This simple format supports peak filters only. Enable Linear phase if you want PGGB to linearize the filters' phase response.

Use Browse beside Left or shared parametric filter to select the file. A left file on its own applies to both channels. For a separate right-channel correction, use Browse beside Right parametric filter. Choosing the file imports it; there is no separate import step. With EQ enabled, the file is read immediately and the EQ status shows any problem. Offline uses the change at the next Start. In Live, enable Apply live changes to restart playback from the same point with the new EQ. The supported filter types are shown below.

AutoEQ filter names: Exports may label low- and high-shelf filters LSC and HSC. These correspond to LSQ and HSQ. If your PGGB version does not recognize the exported names, use the corresponding supported labels.

Enabling EQ and checking its status.

After importing filters, turn on Equalization to apply them. The Equalization icon in the footer shows whether EQ is enabled. Purple indicates an EQ issue; hover for details. Click the icon to open the EQ settings.

EQ applies during both offline remastering and live playback. In PGGB-Live, enable Apply live changes to apply changes to the current stream. Playback pauses briefly while the settings are updated, then resumes from the same position.

---

PGGB-ARC.

PGGB-ARC is designed to make headphone listening more natural and comfortable. ARC stands for Anthropometric Room and Crossfeed. Open Settings → Personalize Sound → PGGB-ARC to use it.

When you listen to speakers, each ear hears both channels with small differences in timing and level. Headphones largely separate the channels, which can make the sound seem to sit inside your head. ARC adds selected spatial cues to soften that effect; it does not fully simulate loudspeakers.

ARC provides two independent effects. Crossfeed adds a quieter, filtered and delayed contribution from each channel to the other ear. Room adds reflections and reverberation to create a sense of space. Use either effect or both, according to the recording and your preference.

Turn on PGGB-ARC to activate the controls. ARC is disabled by default. The panel has three panes:

• Crossfeed: Choose Off, Mild, Medium or Strong; Medium is the default. Speaker Angle offers 30, 45, 60, 90 or 120 degrees, and Brightness Compensation ranges from −6 to +6 dB. Frequency Dependent Delay is on by default. Torso Bounce is off, with Torso Absorption available when it is enabled.
• Head measurements: Adjust the head and shoulder measurements used by crossfeed. Select centimeters or inches, choose Small, Medium, Large or Custom Head Size, and enter Head Circumference and Ear-to-shoulder distance if desired. The defaults are a useful starting point. This pane is grayed out when Crossfeed Preset is Off.
• Room: Choose Off, Studio, Living Room or Concert Hall. Each room sets a reflection level and decay time, which you can adjust. Reset room settings restores the selected room's defaults. Start with Room off, then try Living Room if you want more ambience.

We recommend starting with Medium crossfeed and Room off. Compare a familiar track with and without ARC, keeping the other processing settings the same. Once you have chosen a crossfeed setting, try Living Room on a dry recording. Recordings with strong existing reverberation may sound better without an added room effect.

Use PGGB-Live with Apply live changes enabled to compare crossfeed settings during playback. Room is available only for offline processing, so its pane is grayed out while PGGB-Live is active.

The PGGB-ARC guide explains every control, including how to take your head and shoulder measurements.

---

Peak Normalize, Trim and Add Silence.

These optional controls adjust the final peak level and the silence at the beginning and end of tracks. Use them when you want consistent peak headroom or different track spacing. Leave them unchanged when you want to preserve the recording's existing levels and transitions.

Peak normalization.

Choose Peak Normalize under Settings → Offline → Gain to set each track's highest peak to a chosen level after resampling and EQ. We recommend starting at −3 dBFS. The unit dBFS expresses level relative to digital full scale, so a negative value leaves headroom below the maximum. Peak normalization aligns peak levels; it does not make recordings equally loud or give them the same average level. Tracks with different dynamics can still sound different in volume.

Trim silence (PCM Input Only).

Open Settings → Offline → Silence to remove silence or very quiet material at the beginning and end of PCM source tracks. Start with a threshold between −60 and −70 dBFS, then check the result. A higher threshold can remove more material, including quiet musical fades. Choose a setting that preserves the beginning and end of the performance.

Add silence (PCM Output Only).

Open Settings → Offline → Silence to add a fixed amount of silence at the start or end of PCM output tracks. A starting value of 1 second can provide a clear pause and may help players that cut off the beginning or end of a track. Avoid adding gaps between movements or tracks intended to play continuously.

---

Combining Tracks.

Use Combine to treat consecutive tracks as one continuous recording, such as movements from a classical performance. This can improve reconstruction accuracy with the legacy whole-track engine. PGGB writes each combined group as one output track and does not split it back into the original track numbers. The following requirements apply:

1. Track ranges must not overlap.
2. The source files must contain track numbers in their metadata.

Simple Combining of Tracks.

Create combine.json in the album folder. Enter a JSON array of quoted track ranges, separated by commas. In the example below, tracks 1–6 become one output track and tracks 7–16 become another. List only the ranges you want to combine.

Combine Tracks in Multi-Disc Album.

For multiple discs stored in one album folder, use the disc-specific combine.json format shown below. This example combines ranges from discs 2 and 3 while leaving disc 1 unchanged. The files must contain disc-number metadata as well as track numbers. If no disc number is present, PGGB treats the files as belonging to one disc.

---

Load or Save Settings.

Use Settings → Profiles to save settings for different DACs or to compare processing choices. Enter a New profile name and click Save as profile to create one. Leave the name blank to save into the selected profile. If a name already exists, PGGB asks before updating it. Ctrl+S saves the same way. Remove profile asks for confirmation before permanently deleting the selected profile; Default cannot be removed. Export profiles for backup or sharing, and import them to restore or reuse settings. Profiles do not include Live’s own settings, the Live/Offline mode or the theme.

Default profile: You can save changes to its settings, but you cannot remove it.

---

Processing options.

In Settings → Offline → Workers, the Processing speed dial trades processing speed for capacity left for other apps. Its percentages are approximate. Changes apply to the next track; the current track keeps its worker count. Live has separate Workers and Memory settings.

Start, Pause and Stop are in the footer. Open Settings → Offline → Workers for the operation and worker count, or Processing logs for logging.

1. Mode: Choose Processing or Analysis.
• Processing: Converts the selected files. This is the default mode.
• Analysis: Reports how tracks would be processed without writing remastered audio. Use it to identify potential problems and, with the legacy engine, check paging requirements.
2. Workers: Worker threads let PGGB divide processing across CPU cores. Leave this at Auto to use the available logical cores as the upper limit; PGGB selects a count for each track based on memory and processor resources. Set a lower maximum if you need to reduce heat or fan noise. Manual choices are even numbers from 2 up to the available logical core count.
3. Logging: We recommend Summary or Simple for normal use. None disables logging. Verbose and Debug provide more detail when investigating a problem or sending logs to support.

PGGB restores your settings when it next starts, except for Mode, which returns to Processing. This prevents an earlier Analysis session from leaving file conversion disabled.

---

Status.

Status bar: The current configuration appears on the right. Activity appears on the left, including the current track, CPU activity and memory use. After processing, it reports the number of completed, skipped and failed tracks.

Progress bars: The Offline view shows batch, album and track progress, including the album and track currently being processed.

---

Log files.

Enable logging under Settings → Offline → Processing logs. Album logs are saved in the output folders. Use Summary or Simple for normal work; select Verbose or Debug when troubleshooting.

V9 log files are pggb_debug_v9.log and pggb_album_debug_v9.log. Analysis reports are pggb_album_analysis_v9.csv and pggb_batch_analysis_v9.csv. The version suffix keeps them separate from V8 files. For live playback, enable Log diagnostics under Settings → Live → Logs and statistics.

DAC and System.

The defaults are a good starting point. Open Settings → DAC to select your DAC, output format and rate, and Settings → Processing for the engine and processing options. Begin with a matching DAC preset if one is available, then adjust only the settings you need.

The controls on the left depend on your DAC's capabilities and your computer's processing resources.

1. DAC Selector: Select a matching preset to load our suggested settings, or choose Other DACs and select a DAC type. You can adjust a preset after loading it. Use Restore recommendations to restore its suggested settings.
2. Output Mode: Choose PCM or DSD. A DAC preset supplies a recommended mode, which you can change.
3. PCM / DSD output settings: Use Output format under Settings → DAC → Output settings to select PCM or DSD.
4. Output Sample Rate: Choose a PCM or DSD output rate supported by your DAC and player. Use Output Rate Family below to choose between the 44.1 and 48 kHz families.
   See the PCM output rate menu

   

   Menu preview · expanded to show all choices · select to enlarge
   • PCM: Select the highest PCM rate supported by your DAC and playback path.
   • DSD: Select the highest supported DSD rate, from DSD64 through DSD2048, then choose single-stage or two-stage processing.
     • Single-stage processing: Converts directly to the final DSD rate. It takes longer but provides the highest reconstruction accuracy.
     • Two-stage processing (Legacy engine only): Upsamples first to 16fS or higher, then uses a simpler second stage to reach the final DSD rate. It is faster than single-stage processing. A higher first-stage rate improves reconstruction accuracy but increases processing time. Several first-stage rates are available.
     • With the legacy Nyquist-Shannon engine, single-stage conversion also requires substantially more memory than two-stage conversion.
5. Output Rate Family: Both PCM and DSD support conversion between the 44.1 and 48 kHz base-rate families.
   • Auto: Keeps the source track's base-rate family. For example, a 44.1 kHz source produces a multiple of 44.1 kHz.
   • Always output 44.1kHz rates: Produces a multiple of 44.1 kHz regardless of the source family. Use it when your DAC or player requires this family, including DSD playback through Roon. It can also be useful when the DAC's highest supported rate is in this family. Conversion between families takes additional processing time, with little change in memory requirements.
   • Always output 48kHz rates: Produces a multiple of 48 kHz regardless of the source family. Consider it if your DAC internally converts to this family or performs best at one of its rates. Conversion between families takes additional processing time, with little change in memory requirements.
6. Output Bit Depth: Choose the output bit depth for your DAC or for any processing you plan to apply later. It is separate from PGGB's internal processing precision.
   See the PCM bit depth menu

   

   Menu preview · expanded to show choices; availability depends on license and settings · select to enlarge
   • PCM direct to DAC: Choose a bit depth suited to your DAC. Use 32-bit output if the DAC supports it appropriately; PGGB offers fixed bit depths from 12 to 32 bits. Some R2R DACs may perform better at a lower bit depth at very high sample rates, such as 1 MHz or above, because of settling-time limits.
   • PCM for further processing: Choose 64-bit doubles when you plan to apply further EQ or volume adjustment. This uses twice the uncompressed storage of 32-bit output. Alternatively, use 32-bit output with noise shaping.
   • DSD: DSD is a one-bit stream, so the PCM bit depth controls are unavailable.
7. Noise shaping/Modulators: Noise shaping reduces quantization noise in the audible range by moving more of it to higher frequencies. Quantization noise is the error introduced when audio values are represented at the output bit depth. PGGB uses noise shapers for PCM and sigma-delta modulators for DSD.
   • Adaptive DSD Modulator: Choose from three groups. Low Noise offers 5th-, 7th- and 9th-order modulators, with audible-band quantization-noise figures of roughly −200 to −400 dB. Very Low Noise offers 13th- and 19th-order modulators, roughly −400 to −600 dB. Ultra Low Noise extends to roughly −600 to −800 dB. These are numerical processing figures, not the physical noise floor of a DAC. The modulators adapt to the DSD rate to reduce audible-band noise while keeping the rise in high-frequency noise gradual. Very Low Noise and Ultra Low Noise also adapt their order; the menu shows the maximum order.
     If you hear hiss: Higher-order modulators, from 13th order upward, can occasionally interact with DAC filtering or downstream equipment. Try a lower-order modulator if this occurs.
   • Adaptive Noise Shaping (PCM): PGGB tailors the noise-shaping profile to your output bit depth, sample rate and filter settings.
     • NS: Optimal: The recommended choice. PGGB selects a noise-shaping profile for your settings to minimize quantization noise in the audible range.
     • NS: Alternate: Available at 256-bit precision. Uses a different noise-shaping profile that can change the perceived presentation of transients and high frequencies.
     • NS: Extended: Try this for recordings with high background noise or older analog transfers that sound harsh. It can be combined with HF: Aggressive when that filter is available.
     • Dither only: Applies dither without adaptive noise shaping. Dither is a small amount of noise used to reduce distortion from quantization. We do not recommend this option for rates of 8fS or higher.
8. Precision: For PCM, select the numerical precision used throughout PGGB processing and EQ. The choices are 64, 107, 128 and 256 Bit. A fresh install uses 107 Bit with the Your DAC preset; named DAC presets select 256 Bit. The 64 Bit option is the fastest. The 107-bit and 128-bit modes have similar memory requirements, but 107-bit processing is about 2.2 times faster with a small reduction in accuracy. Higher precision mainly increases processing time on the default engine; it also increases memory requirements on the legacy Nyquist-Shannon engine. For DSD, precision is selected automatically from the modulator and output rate, so the precision selector is disabled. See System requirements.
   
   
9. Gain: Open Settings → Offline → Gain to choose how PGGB adjusts the final level. Auto (avoids clipping) is the default; it leaves gain unchanged unless an inter-sample over requires attenuation. An inter-sample over occurs when reconstruction produces a peak above digital full scale between the original samples. Fixed (software pre-amp) applies your chosen negative gain after processing. Peak Normalize sets each track's peak to your chosen level. None (allow clipping) leaves gain unchanged even if clipping occurs. ReplayGain: Album Peak uses existing album-peak metadata to apply a consistent peak-based adjustment across the album; analyze the tracks with a tool such as foobar2000 first to add that metadata.
   
   
   

   Current interface preview · select to enlarge
   
   
   

10. Optimize Paging and Memory:

    Find these controls on the right of Configure, above Manage Configurations. On Mac, you can select up to two paging locations. See Page File (Mac) for details.

    The Optimize Paging and Memory controls apply only to the legacy Nyquist-Shannon engine. They help manage whole-track data that exceeds available RAM. The default engine disables them. They become available when you select Legacy under Settings → Processing.

    
    
    
    

    Legacy-engine controls · paging locations apply on Mac · select to enlarge

    Optimize paging for PCM

    Enable Optimize PCM Paging to help process long tracks with the legacy engine. If a track does not fit in RAM, PGGB processes one channel at a time and reduces gain by 1 dB to provide clipping headroom. For a continuous performance split across tracks, consider a fixed gain of −1 dB or lower to keep the level consistent. You can also force one-channel-at-a-time processing when high CPU use causes thermal instability.

    Process in Blocks (>DSD256)

    Process in Blocks allows the legacy engine to process DSD tracks that would not otherwise fit in available memory. It trades a small reduction in reconstruction quality for lower memory requirements. Leave it off when the track can be processed without it.

    Use only when needed: Enable block processing only if a track cannot otherwise be processed with your available memory. PGGB asks ‘Continue with block processing enabled?’ before applying the option.

    Before using block processing, also compare a lower DSD rate without blocks. If you selected the 25th-order modulator, try the 19th-order option. Choose the result you prefer.

    A kangaroo (🦘) icon in the status bar indicates block processing. The number of ‘hops’ is the number of blocks used for the track.

    This option is available only when at least one block fits in RAM at the selected DSD rate and modulator setting.

    Reduce contention

    Enable Reduce NVME Contention to limit competing reads and writes during paging, particularly when a single NVMe SSD is used. This can improve speed when disk activity is the bottleneck.

---

Precision Optimized Upsampling Engine.

Introduced in V9, the default engine is an efficient reformulation of the Nyquist-Shannon upsampling engine. It is a true breakthrough in both upsampling quality and processing speed. It builds a processing pipeline tailored to the selected precision and the input and output sample rates. Within the passband (the frequency range it preserves), reconstruction accuracy is limited by the selected precision. It uses much less RAM and is considerably faster than the legacy engine. Listeners describe more natural timbre, clearer instrumental separation and a deeper sense of space. Read about the PGGB engine ↗

In Settings > Processing, choose Precision optimized for precision optimized upsampling, or Legacy for the Nyquist-Shannon engine, which requires more memory. Read about the PGGB engine ↗

The Optimize Paging and Memory controls are disabled while Precision optimized is selected. Those controls manage whole-track storage for the legacy engine.

V8 configurations still load. PGGB converts the previous setting and displays a message; no manual correction is needed.

Open Settings > Processing > Precision optimized to choose the processing precision and an Optimal, Alternate or Minimal filter card.

Filter Choice

Changing Filter Choice during Live playback restarts the stream from the current position. The three choices are available only while Precision optimized is selected.

Open Settings > Processing to choose the engine and filtering. The new engine’s memory use does not grow with track length, and paging is rarely needed. Calculation precision is not reduced to save memory. Choose Legacy to use the older engine and its filtering and memory controls. PCM and DSD retain their own choices.

With Precision optimized selected, choose Optimal, Alternate or Minimal. Optimal is the default and the recommended starting point. Alternate keeps a slightly wider frequency range before rolling off. Both match reconstruction accuracy to PGGB’s internal calculation precision within the frequency range they preserve. Minimal applies the least high-frequency filtering and may allow more high-frequency distortion. Compare on familiar music and choose what you prefer; a wider bandwidth is not necessarily better.

Filter Choice is available only when Precision optimized is selected.

Live PCM and DSD processing requires this engine, so PGGB-Live keeps Precision optimized selected and locked.

HF Noise Filters are separate controls for high-frequency noise already present in the recording. Select the reconstruction engine separately under Settings → Processing.

The Precision Optimized Upsampling Engine handles source-noise filtering as part of its processing, so the separate HF Noise Filters controls are disabled. They become available with the legacy engine when Legacy is selected. Read about the PGGB engine ↗

---

Filter Preferences.

Under Settings → Processing → Legacy → Legacy processing, Filtering offers Automatic and Manual control of the legacy engine's HF noise filters. These settings apply when Legacy is selected.

1. Automatic: The recommended setting for the legacy engine. PGGB selects the HF noise filter for each track, so the corresponding manual controls are disabled.
   
   
2. Manual: Choose Manual to set the legacy engine's HF noise filters yourself. The appropriate choice depends on the source recording, so PGGB groups the controls by input format and rate:
   

   Manual controls with Legacy selected · select to enlarge
   • 2fS PCM: 88.2 and 96 kHz sources. These may contain music up to about 40 kHz and generally less noise above that range than higher-rate sources.
   • ≥ 4fS PCM: 176.4 kHz, 192 kHz, DXD (352.8 kHz) and higher. These recordings may contain significant noise above 40 kHz as well as ultrasonic musical content.
   • DSD: DSD64, DSD128, DSD256, DSD512, DSD1024 and DSD2048 sources. DSD contains substantial high-frequency quantization noise.
   
   
   The filter choices are Aggressive, Minimal and Off.

---

Noise Shapers.

The noise figures quoted on the landing page describe numerical processing, not the physical noise floor of a DAC. The landing-page illustration shows PCM noise-shaper profiles; the quoted figure of up to −800 dB refers to documented DSD results.

The displayed noise-shaper type follows the output mode and your Adaptive Noise Shaping selection:

1. SDM: A sigma-delta modulator used for DSD output, optimized for the selected DSD rate.
2. Optimal: Minimizes audible-band quantization noise, with the low-noise region extending to at least 40 kHz. At 256-bit precision, the numerical noise floor can reach about −740 dB.
3. Extended: Keeps quantization noise low over a wider frequency range, with slightly more noise in the audible band than Optimal. Reported numerical floors reach about −540 dB at 128-bit precision, −500 dB at 107-bit and −360 dB at 64-bit. Some listeners describe a more open or airy presentation.
4. Dither: Applies Gaussian dither without noise shaping when Dither Only is selected. We recommend noise shaping for high-rate PCM output; Dither Only is intended for lower-rate use.

The annotated plot below compares the noise-shaper profiles. These numerical noise floors describe PGGB's processing, not the physical noise floor of your playback hardware.

---

HF Noise Filters.

HF noise filters reduce high-frequency noise already present in a source recording, including noise from analog-to-digital conversion. They are separate from the processing engine selection under Settings → Processing.

These controls are disabled by default. The Precision Optimized Upsampling Engine handles source-noise filtering internally. The options below apply only after you turn off Optimal HF Filtering and select the legacy Nyquist-Shannon engine.

The legacy engine offers three filter choices. We recommend Automatic selection unless you want to compare their effects on a particular recording.

1. Aggressive: Applies stronger filtering above the audible range. This removes more noise but also more ultrasonic musical content. Some listeners describe a smoother presentation.
2. Minimal: Applies less filtering, retaining more ultrasonic content while also leaving more noise. Some listeners describe a more open or lively presentation.
3. Off: Disables source-noise filtering for PCM. On higher-rate material, this can leave noise that contributes to harshness or fatigue. Filtering remains enabled for DSD because its high-frequency noise can affect downstream equipment.

High-rate recordings can contain substantial noise outside the audible range, introduced by the recording's analog-to-digital converter (ADC). This is particularly relevant to DXD and DSD, but some 88.2 kHz and higher PCM recordings can also benefit from filtering. At ultrasonic frequencies, noise and musical content overlap, so stronger filtering trades retained content for noise reduction. Start with Automatic, then compare Aggressive and Minimal if you want to fine-tune a recording. The audible difference is usually subtle and depends on your system and preferences.

The annotated example below compares the HF noise-filter options.

---

Folder Paths and Permissions.

PGGB checks free space before each track. For FLAC and WavPack, it allows for the uncompressed size, so the message says “needs up to”. A track may be stopped even if its eventual compressed file would have fitted. WAV, AIFF and DSF use an approximate size estimate.

PGGB needs read access to input folders and read/write access to output and temporary folders.

For the reconstruction approach, supporting measurements and legacy-engine performance details, see Understanding PGGB.

V9 uses your email address and an activation code for normal license activation. You do not need to locate a Hardware ID or install a license file unless you are using the offline activation method described below.

Upgrading from V8: Your existing license works in V9. On the first start with internet access, PGGB registers it to the original email address. The Register button then changes to Activation Code, where you can view your code.

Activating a new purchase: Use the activation code sent by email. Install PGGB, open Settings → Application → License and About, press Register, enter the purchase email and activation code, then restart PGGB.

Moving to another computer: Install PGGB there and open Settings → Application → License and About → Register. Enter the same email and activation code, then select your license. If all permitted installations are already in use, PGGB asks you to confirm the move. The previous computer changes to Lite the next time PGGB starts there.

Without an activated license, PGGB runs as a 30-day trial and then becomes Lite. Licenses are personal and can move with you. Commercial use requires a commercial license.

Offline activation: If the computer cannot reach the license service, email us its Hardware ID from Settings → Application → License and About. We will send a license file to install through Settings → Application → License and About → Activate → License File….

For deactivation, a list of licensed computers or help recovering an activation code, see Activating your license.

Tiers.

The five tiers are Trial, Lite, Plus, Pro and Max. They differ in the processing precisions and DSD rates available without limits, and in live-playback time limits. Pro and Max also include a second Live license for a computer used only for playback.

See Compare licenses for the complete limits, or Buy a PGGB license to purchase or upgrade.

---

So long and thanks for all the fish

.·´¯·.´¯·.¸.ZB.´¯·.¸¸.·´¯·.¸><(((º>