PFClean Disk Caching: The Complete Technical Guide
- May 15
- 14 min read
Complete Technical Guide to Disk Caching in PFClean for Restoration Artists and Archivists — Standards-Based Caching, Workflow Optimization, and Performance Strategies Explained in Depth
The restoration workflow bottleneck
Processing uncached high bit depth 4K scans with complex effect chains can often feel unresponsive, making manual restoration work tricky at best. Export times balloon to hours for feature-length projects, and every settings adjustment forces you to re-render from scratch. Without strategic caching, even powerful hardware can't overcome the computational demands of multi-layered temporal effects.
Caching is one of the most powerful yet underutilized features in professional restoration workflows. When configured correctly, it transforms PFClean from a real-time processing engine into an interactive restoration environment capable of handling 4K footage and complex effect chains with fluid playback.
Most restoration platforms require expensive RAID systems to sustain real-time playback. PFClean avoids this by using intelligent local caching, allowing high-performance restoration on comparatively modest hardware.
This guide covers everything from basic setup to advanced workflow strategies, helping you leverage disk caching to maximize both performance and efficiency, cutting export times by up to 90% and enabling real-time manual restoration work even on complex projects.
PFClean, by The Pixel Farm, is a professional restoration and remastering suite for film and video. It offers automated and manual tools to remove defects and preserve image quality for modern delivery.
Quick Navigation
What is Disk Caching in PFClean
Caching refers to the process of storing processed data in a temporary storage location to improve efficiency and performance. In the context of PFClean, this means processing your effects and toolsets once, then storing the results on disk to be instantly played back when needed, without reprocessing.
Why disk caching matters
When caching is used strategically and configured with appropriate hardware, it enables several critical workflow improvements:
Real-time playback at any resolution: View the results of multiple effects and toolsets in real time, even in 4K and beyond. Hardware processing limits disappear once frames are cached.
Fluid manual restoration: Working with manual tools becomes dramatically more responsive when you've selectively cached complex automatic effects applied to the same clip. Paint, clone, and repair without waiting for upstream effects to recalculate.
Significantly reduced export times: When clips or toolsets have been cached in your project, export times can be reduced by 90% or more. The system reads pre-calculated frames from disk rather than reprocessing the entire pipeline.
Iterative workflow efficiency: Test different approaches, compare results, and refine your work without the computational cost of reprocessing heavy effects every time you scrub through the timeline.
The strategic advantage
Caching isn't just about speed—it's about changing how you work. Without caching, every frame advance triggers a recalculation through your entire effects stack. With caching, you decide which heavy processes run once and which remain interactive. This transforms restoration from a batch-oriented workflow into an interactive, creative process.
Using Standards-Based Disk Caching
Unlike traditional proprietary systems that tether performance to expensive, localized hardware, PFClean’s standards-based caching offers a scalable architectural advantage for studio-scale operations. By moving away from monolithic,single-location disk writing, our metadata-driven approach decouples processing power from storage speed.
This ensures high-performance restoration workflows across standard NVMe or shared network environments, eliminating I/O bottlenecks and "cache wipeouts." For large-scale studios, this architectural shift transforms caching into a collaborative asset, providing the stability and speed required for high-throughput, multi-user pipelines without the need for hardware-heavy infrastructure.
What standards-based caching means
Standards-based caching allows you to assign different storage locations based on the resolution standard of your footage. This granular control enables performance and storage optimization strategies that aren't possible with one-size-fits-all caching.
Strategic advantages:
Performance tiering: Cache large 4K files on high-performance NVMe or SSD arrays to meet demanding throughput requirements, while smaller HD or SD files can be cached on slower disks or even network storage where performance isn't critical.
Storage utilization: Allocate expensive high-speed storage only where it delivers measurable performance gains. Route lower-resolution material to more economical storage tiers.
Workflow flexibility: Different projects with different resolution requirements can automatically route to appropriate storage without manual intervention once configured.
By strategically allocating caching locations based on file sizes and performance needs, you optimize both storage utilization and overall efficiency of your restoration workflow.
Setting up a cache location
A cache location must be configured before you can use caching within your projects. The setup process is straightforward:
Step-by-step setup:
Open the cache configuration: Click the disk icon in the lower left of the Project Manager. This opens a file browser.
Select your cache location: Navigate to the drive or folder where you want cache data stored. For 4K work, choose a fast external SSD or NVMe array. For HD/SD work, standard hard drives or network storage may suffice.
Name the location: Enter a descriptive name that indicates the storage type and resolution standard (e.g., "4K_NVMe_Cache" or "HD_Network_Cache").
Assign a standard: Select the resolution standard this location will handle from the dropdown menu (4K, HD, SD, 2K, etc.). If you want all footage to cache to the same location regardless of standard, select "Any" from the dropdown.
Confirm: Click Confirm to save the configuration.
Once configured, PFClean will automatically cache footage matching that standard to the designated location in all newly created and existing projects. The information bar displays available storage space on the drive, helping you monitor capacity.
Cache location strategy
For single-storage setups: If you're working with one storage device, set the standard to "Any" to cache all footage to the same location. This also handles footage that doesn't have an assigned standard.
For multi-tier setups: Configure multiple cache locations, each targeting a specific standard and storage tier:
4K footage → Fast NVMe array
2K footage → SSD
HD footage → Standard HDD or network storage
SD footage → Network storage
This tiered approach ensures optimal performance where it matters most while avoiding unnecessary hardware costs.
Disk Caching in the Workflow Manager
Within the Workflow Manager, two primary toolsets leverage disk caching to accelerate your restoration pipeline:
Digital Wet Gate: Pre-processes scanned film to remove dust, dirt, and other particulate matter using infrared channel data or advanced algorithmic detection.
Standards: Handles format conversion, scaling, and standards compliance operations.
Both toolsets can process multiple clips in batch mode, making caching especially valuable for high-volume restoration projects.
Initiating the caching process
To begin caching a toolset:
Start the cache: Click the button in the middle of the disk icon to the right of the toolset you want to cache.
Monitor progress: A percentage indicator shows caching progress. The white number represents incoming clips ready to be cached; the blue number indicates clips that have been successfully processed.
Interaction limitation: While caching is running, you cannot interact with the toolset settings. This ensures data consistency during the batch process.
Managing active cache operations
If you need to pause or stop caching at any time:
Access batch processing controls: Click the green batch processing button in the middle of the disk icon. This serves as a convenient shortcut to the Batch Processing panel.
Manage tasks: From the Batch Processing panel, you can:
Pause or resume active caching tasks
Cancel caching operations
View estimated completion time
Monitor multiple simultaneous caching tasks
Prioritize or reorder queued operations
This centralized control panel is essential when running multiple caching operations across different toolsets or projects, allowing efficient resource management and workflow prioritization.
When to cache Workflow Manager toolsets
Digital Wet Gate caching is essential when:
The toolset feeds downstream effects (especially a Workbench with additional processing)
You're performing multiple review passes and need immediate playback
Export timelines are critical and you can't afford reprocessing delays
Standards toolset caching is valuable when:
Format conversion is computationally expensive (upscaling or complex aspect ratio adjustments)
The output feeds multiple downstream nodes that would otherwise trigger redundant reprocessing
You're delivering in multiple formats and want to cache the intermediate standard conversion
Disk Caching in the Workbench
The Workbench offers the most sophisticated caching implementation in PFClean, with granular control at both the clip and individual effect levels. This flexibility is particularly powerful when handling multiple automatic temporal effects that would otherwise bottleneck real-time playback.
Understanding the effects stack
In the Workbench, effects are organized in a stack where each effect processes the output of the effect above it. Effects are designated as either:
Automatic (A): Temporal effects that analyze multiple frames, such as deflicker, dust removal, or stabilization. These are computationally expensive.
Manual (M): Interactive tools like paint, clone, or color correction that respond to user input in real time.
The key to efficient caching lies in understanding this stack relationship and strategically caching automatic effects to accelerate manual work.
Clip-level caching
The simplest caching approach is to cache the entire effects stack for a clip.
To cache a complete effects stack:
Initiate caching: Click the disk icon at the clip level in the Workbench.
Play through the clip: Press play. As the clip plays back, the percentage bar increases.
Complete the cache: When playback reaches the end and the progress bar hits 100%, the cache is complete. The clip now plays back in real time.
Important limitation: If you make any changes to settings in any effect in the stack, the entire cache is lost and must be rebuilt. This approach works well for final-pass restoration where settings are locked, but it's inefficient during active restoration work where you're still iterating on effect parameters.
Effect-level caching: The power of granular control
The Workbench's ability to cache individual effects within the stack is where its caching system truly excels. Each effect in the stack has its own cache button, allowing you to cache specific processing stages while keeping others interactive.
Strategic advantages of effect-level caching:
Downstream efficiency: Effects added below a cached effect work on pre-calculated results, speeding up processing of subsequent effects dramatically.
Interactive manual work over cached automatics: By placing manual effects below cached automatic effects, you can interact with them without impacting playback performance. This is imperative for large-scale manual restoration work where speed and fluidity are critical.
Selective processing: Not every automatic effect needs to be cached. Group automatic effects at the top of your stack, and when you're satisfied with their combined results, cache only the bottommost automatic effect. This saves disk space by not caching intermediate results unnecessarily while still benefiting from smooth playback when using manual effects below.
Example workflow:
Consider a clip with the following stack from top to bottom:
Deflicker (Automatic)
Stabilize (Automatic)
Dust Removal (Automatic)
Paint (Manual)
Clone (Manual)
Efficient caching strategy:
Do not cache Deflicker
Do not cache Stabilize
Cache Dust Removal (the bottommost automatic effect)
Leave Paint and Clone uncached (they're manual and interactive)
Result: The three automatic effects at the top process once and are stored to disk via the Dust Removal cache. The manual Paint and Clone tools work on these cached results, enabling real-time interaction regardless of how complex the automatic processing is.
Preserving caches during iteration
Unlike clip-level caching, effect-level caches remain intact when you modify effects below them in the stack. This allows continuous refinement of manual work without invalidating upstream automatic processing.
However, modifying any effect that has been cached, or any effect above it in the stack, will invalidate that cache and all downstream caches, requiring rebuilding.
Not sure what storage and machine tier to spec? See our Recommended Hardware Guide.
Workflow Tips & Best Practices
Working with large restoration projects can quickly become frustrating when exports take too long, manual tools feel sluggish, or you find yourself repeating the same caching steps across multiple clips. PFClean’s disk caching workflow is designed to remove that friction.
By caching effects strategically, you can dramatically improve playback responsiveness, speed up exports, and prepare complex effect stacks in just a few clicks, without wasting time on unnecessary reprocessing.
Stack Management & Caching
Principle: Structure your effects stack to minimize cache invalidation and maximize reuse.
Best practices:
Keep automatic effects grouped at the top: This allows you to cache once at the boundary between automatic and manual effects, rather than caching multiple automatic effects separately.
Cache the transition point: If your stack contains both automatic and manual effects, cache only the bottommost automatic effect. This creates a "cache boundary" where everything above is processed once, and everything below remains interactive.
Separate iterative work from finalized work: If you're still experimenting with automatic effect settings, don't cache them yet. Once you're confident in the automatic processing, cache it and move on to manual refinement work below.
Presets & Caching
Principle: Pre-configure caching settings in Workbench presets to accelerate batch application across multiple clips.
Workflow:
Build your effects stack: Create the combination of automatic and manual effects you want to deploy across multiple clips.
Set cache flags: Click the cache button on the effects you want cached before saving the preset.
Save as Workbench preset: When you save the stack as a preset, the cache configuration is stored with it.
Apply to multiple clips: When you use "Apply All" to deploy the preset across clips in your Workbench or Worklist, the effects will be pre-configured to cache automatically without needing to manually click the cache button on each clip.
Advanced technique: Add a Remaster node underneath your Workbench and set it to play through once. When playback finishes, delete the Remaster node and return to the Workbench. All clips will now be cached and ready for final review or export.
Caching Toolsets
Principle: When toolsets feed downstream effects, caching the upstream toolset eliminates redundant reprocessing.
Scenario: A Digital Wet Gate passes its results downstream into a connected Workbench where further restoration work is undertaken.
Problem: Whenever a frame is advanced within the Workbench, it must first be processed by the Digital Wet Gate upstream. This means every frame advance triggers reprocessing of the wet gate logic, even though those results haven't changed.
Solution: Disk cache the Digital Wet Gate. The Workbench can then utilize this cached data to process its own effects, significantly reducing overall processing time. Instead of recalculating the wet gate for every frame, the Workbench reads pre-calculated frames from disk and applies only its own effects.
Performance gain: This approach can reduce frame-to-frame latency by 60-80% in typical restoration pipelines, making manual work in the Workbench dramatically more responsive.
Caching & Exports
Principle: Cached clips export exponentially faster than uncached clips because the system reads pre-calculated frames from disk rather than reprocessing the entire pipeline.
Real-world performance comparison:
Consider a 4000-frame clip processed with:
Digital Wet Gate with default settings
Workbench containing 2 automatic effects and 1 manual effect
With caching (Digital Wet Gate and automatic effects cached):
Export time: Approximately 1 minute 30 seconds
Without caching (all caches deleted before export):
Export time: Approximately 14 minutes
Performance difference: Caching reduces export time by over 90% in this example, from 14 minutes to 1.5 minutes. Our PFClean Performnace Benchmark article shows the direct benefits of caching and optimisation showing significant performance improvements across HD and 4K restoration projects.
Strategic implication: For projects with multiple deliverables or iterative client review cycles, maintaining caches between exports can save substantial time. However, this must be balanced against storage costs—cached data can consume significant disk space on large projects.
Best practice for export workflows:
During active restoration: Cache liberally to accelerate review and iteration.
Before final export: Verify all caches are current and valid. Stale caches from earlier iterations can produce incorrect output.
After final delivery: Consider purging caches to reclaim storage, especially for archived projects unlikely to require re-export.
Curious how caching can improve PFCleans exports? See our Performance Benchmark Guide.
Troubleshooting Common Issues
Cache not initiating
Cause: No cache location has been configured for the footage standard.
Solution: Open the Project Manager and configure a cache location for the relevant standard (or "Any" for all standards). See Using Standards-Based Disk Caching for setup instructions. If you have a custom standard and have not set the cache type we recommend using the 'Any' option for your cache disk location.
Slow playback despite caching
Cause: The cache drive is too slow to deliver real-time throughput at the resolution being played.
Solution:
For 4K footage: Use NVMe or high-performance SSD arrays capable of sustained read speeds above 1GB/s.
Check if the cache location is on network storage—network latency can bottleneck playback even if bandwidth is theoretically sufficient.
Monitor drive health—failing or fragmented drives can dramatically reduce read performance.
Cache invalidated after minor changes
Cause: Modifying any effect in the stack at or above a cached effect invalidates that cache and all downstream caches.
Solution: This is expected behaviour with Automatic effects that modify every frame in the clip. To minimize cache invalidation:
Finalize automatic effect settings before caching
Use effect-level caching instead of clip-level caching to limit invalidation scope
Structure your stack so effects you're still adjusting are below cached effects
Running out of cache storage
Cause: Cached frames consume significant disk space, especially at 4K and higher resolutions. A single 4K clip can generate tens of gigabytes of cache data.
Solution:
Monitor the storage capacity indicator in the Project Manager
Strategically cache only essential toolsets and effects rather than caching everything
Purge old caches from completed projects
Use standards-based caching to route high-volume 4K caches to larger storage while keeping HD/SD on smaller drives
Consider adding dedicated cache storage if you regularly work with high-resolution material
Workbench Clip at stuck at 0% cached
Cause: The effect cache is selected but has not been initialized. In PFClean, the Workbench cache operates during playback. This allows you to make small adjustments to individual frames and recache only those frames instead of the entire clip.
Solution:
Enable caching for the clip you want to cache.
Press Play to start caching the effect.
FAQ
How much disk cache do I need for 4K film restoration?
For standard 4K workflows, a minimum of 2TB of dedicated NVMe storage is recommended. This allows PFClean to cache several reels of high-bitrate data without bottlenecking the CPU/GPU. For 8K or high-volume restoration, aim for 4TB to 8TB to minimize cache clearing during a project.
Can I use a standard SATA SSD for caching?
While a SATA SSD will function, it is not recommended for modern restoration tasks. SATA speeds (approx. 550MB/s) are often slower than the data throughput required for real-time 4K playback in PFClean. To avoid dropped frames, use an NVMe M.2 drive with speeds of at least 3,500MB/s.
What is the difference between Disk Cache and RAM Cache?
RAM Cache: Provides the fastest possible access for immediate frame-to-frame processing but is limited by your total system memory (e.g., 64GB or 128GB).
Disk Cache: Stores much larger amounts of rendered data on your drive, allowing you to scrub through entire sequences smoothly without re-rendering, even after a system restart.
How do I clear the PFClean cache to free up space?
In the Workbench, click the cached disk icon next to an effect—or at the bottom of the Effects panel—to remove the disk cache. To delete both the cache and its location, click the information (i) button next to the cache location in the Project Manager.
Do I need a RAID setup for film restoration in PFClean?
A RAID 0 (striped) array of NVMe drives is beneficial if you are working with a large amount of uncompressed 16-bit DPX or EXR files at high resolutions. However, for most users, a single high-performance Gen4 or Gen5 NVMe drive provides more than enough bandwidth for PFClean's caching engine.
Transform Your Restoration Workflow with Strategic Disk Caching
Experiencing slow exports, sluggish playback, or workflow bottlenecks in your restoration projects? See how PFClean's disk caching can cut your export times by 90% and enable real-time 4K playback—even with complex effect chains.
Ready to Experience the Difference?
Work with our restoration specialists to optimize disk caching for your specific footage and hardware setup. We'll analyze your workflow and demonstrate performance gains in real-time on your own material.
About the Author
Adam Hawkes is a PFClean Product Specialist and restoration expert with over 20 years of hands-on experience in film and video restoration. Trained in film handling and film camera operation, Adam has contributed to more than 100 productions, including some of cinema's most celebrated titles. His expertise combines deep technical knowledge of restoration workflows with practical understanding of the physical and optical characteristics of film.
