Latest Features in Topaz Photo AI (V. 4.0.x)

Topaz Photo AI version 4.0x is a major milestone for the software, bringing deeply meaningful improvements in restoration, sharpening, facial recovery, user experience, and performance.

Introduction

"Topaz’s Photo AI has long been a powerful tool for photographers who want to authentically enhance images using artificial intelligence — sharpening, denoising, upscaling, and correcting common photographic flaws. With the release of version 4.0, Topaz takes a big step forward: introducing a brand-new Dust & Scratch removal model, more intelligent face-recovery controls, a deeply reworked Autopilot experience, performance optimizations, and usability refinements. These changes expand the use cases of Photo AI substantially, making it more effective not just for modern digital images but for restoring older, damaged photographs, film scans, and other archival media. Below is a break-down of the most important new features in v4, how they work, and why they matter.

Dust & Scratch AI Model: Restoring Old Images

Arguably the headline feature of Photo AI 4.0 is the introduction of the Dust & Scratch AI model. This is the first time Topaz offers a dedicated model to intelligently detect and remove dust particles, surface blemishes, and fine scratches — common artifacts in scanned film, old prints, or even poorly maintained negatives. According to the release notes, the model is designed to deliver very clean results right out of the box, meaning you don’t need to fine-tune dozens of parameters to fix basic dust issues. (Topaz Community)

This model can run either locally or in the cloud, giving users flexibility depending on their hardware. For machines that meet the system requirements, local processing means no need for credits; for less powerful devices, using the cloud can offload heavy computation while keeping results fast. (Topaz Community)

Once the dust and scratch pass is complete, there’s an integrated healing brush built into the same tool. If the AI misses some stubborn scratches or marks, you can manually paint over those areas and the software processes them as soon as you lift the brush — no extra clicks required. (Topaz Community)

Topaz recommends using Dust & Scratch early in the editing workflow. Because it’s designed to deal with surface-level damage, removing dust first can provide a much cleaner base before applying sharpening, upscaling, or other enhancements. (Topaz Community)

Users who work a lot with scanned family portraits, film negatives, or vintage prints will likely find this feature transformative: it automates a step that would otherwise take a lot of manual retouching in Photoshop or other restoration tools.

Super Focus v2: Sharpening and Focus Recovery Refined

Another major upgrade in Photo AI 4.0 is Super Focus v2, the second-generation focus-recovery model. This new version significantly boosts speed and performance. According to Topaz’s release notes, the architecture is shared with their Recovery v2 model, and they claim it can run up to 500% faster than the previous Super Focus (generation 1) under certain conditions. (Topaz Community)

Notably, Super Focus v2 supports cloud processing, which means even users without powerful GPUs can run the model very quickly using Topaz’s servers. (iDownloadBlog.com) For users with capable hardware, the local speed boost means faster iterations, and the ability to experiment with different focus-recovery strengths without waiting for each pass is a real workflow advantage.

In terms of control, Super Focus v2 introduces a more flexible workflow: after processing, you can adjust the sharpening or deblurring strength using a slider without having to re-run the entire model. This is a huge improvement, because previously you might have to re-process the entire image if you wanted a slightly subtler or stronger result. (Topaz Community)

This generation of Super Focus is more consistent in preserving natural textures while sharpening, which helps avoid the “artificial” or over-sharpened look that can sometimes plagues AI-based deblurring. For photographers dealing with soft RAW images, out-of-focus shots, or motion blur, this improvement makes Super Focus v2 one of the strongest tools in the Photo AI toolbox.

Revamped Autopilot Experience

Topaz Photo AI 4.0 brings substantial improvements to its Autopilot system, designed to simplify enhancement suggestions and make AI assistance more helpful and less intrusive.

When you open the app with version 4 for the first time, you’ll get a prompt to opt in or out of Autopilot. This is more than cosmetic: it lets you decide whether you want Topaz’s AI to suggest enhancements automatically, or whether you prefer to build your own stack from scratch. (Topaz Community)

Suggestions from Autopilot are now shown at the top of the filter stack, making them more prominent and easier to act on. This reversed stacking order (most recent suggestions on top) aligns better with standard photo-editing workflows, where you want quick access to the last or most important adjustments. (Topaz Community)

The new Autopilot is also smarter about image type. It can now detect whether an image is likely a portrait, a landscape, or something else, and choose more relevant default enhancements accordingly — for Sharpen, Recover Faces, Adjust Lighting, and Balance Color. (Topaz Community) This means for batch editing, Autopilot’s suggestions become more meaningful and less generic, which saves time.

Enhanced Face Recovery: Hair, Neck, and Dynamic Strength

Face recovery (or “Recover Faces”) is an important feature in Photo AI, especially when dealing with low-resolution, blurred, or compressed images. In version 4.0, new controls have been added to let you manage not just the face, but also surrounding areas like hair and neck. (Topaz Community) These options were previously hidden or buried in preferences, but now they’re front and center in the editing interface. This gives you finer control over how the recovered face blends into the rest of the subject, making retouched portraits look more natural and integrated.

Topaz has also split the face selection and strength controls. Instead of managing which faces to recover and how strongly to recover them in one place, you now choose faces separately from adjusting recovery strength. (Topaz Community) That means you can manually select which faces to target (very useful in group shots) and then set how much the AI should reconstruct or enhance each face.

One of the smartest upgrades is dynamic face recovery strength. Autopilot now adapts the recovery strength based on the size of faces detected in the image. Larger faces get more conservative enhancement so they don’t look over-processed, while smaller or lower-quality faces can get stronger reconstruction. (Topaz Community) Autopilot also tends to ignore background faces that are not central to the subject, reducing unnecessary or distracting enhancements.

These changes help achieve more realistic results, especially in group portraits or when restoring old family photos, because they give the AI context about what “matters” in the image.

Adjust Lighting v2: More Control Over Light and Color

Lighting correction is a critical part of making photos look polished, and with Adjust Lighting v2, Topaz brings a more refined, nuanced tool into Photo AI 4.0. This version is trained to understand a wider variety of lighting styles, and it produces outputs that are more natural and visually balanced. (docs.topazlabs.com)

One of the biggest changes: you now get separate sliders for highlights and shadows, giving you fine control over both the brightest and darkest parts of the image. This lets you recover detail in blown-out areas or bring up information in deep shadow regions without flattening the picture. (Topaz Community)

There’s also a color correction toggle. With it off, Adjust Lighting v2 improves exposure or contrast while preserving the original color palette. If you turn it on, the tool further refines color as it adjusts light. (docs.topazlabs.com) This flexibility is especially useful when you want to maintain mood or colour temperature — for instance, in warm sunset shots or cool low-light scenes.

Topaz made Adjust Lighting v2 the default lighting model in Photo AI 4.0, so users immediately benefit from these improvements without needing to manually switch. (Topaz Community)

Sharpening Changes and Minor Denoise Control

While Super Focus v2 handles major focus recovery, Photo AI’s general Sharpen filter also gets refined in version 4.0. The Sharpen tool supports multiple models — Standard, Strong, Lens Blur — but now there’s more streamlined control. (docs.topazlabs.com)

One notable change: the Minor Denoise slider, which helps reduce noise while sharpening, is disabled by default in version 4.0 and above. If you still want to use it, you need to explicitly enable it in preferences. (docs.topazlabs.com) This change reduces clutter for users who do not need that control, while still giving access to advanced users who do.

For cases where the Sharpen filter isn’t sufficient (e.g., very soft or blurred shots), Super Focus v2 remains the recommended model. (docs.topazlabs.com)

Performance and Stability Improvements

Photo AI 4.0 includes a variety of under-the-hood performance and stability upgrades. According to the version 4.0.1 patch, the update adds a toggle to disable personalization (which presumably limits how the app learns from your usage), and a cloud processing on/off switch, giving more control over where your AI enhancements are computed. (Topaz Community)

The patch also improves the healing brush animation, making it clearer when the software is processing spots you paint over. (Topaz Community) Several bug fixes are included too: for example, an earlier issue where face selection could be incorrect before running Super Focus was resolved, and text-color issues in the Dust & Scratch modal were fixed. (Topaz Community)

These changes are important because they indicate Topaz is prioritizing not just feature richness but also usability and robustness.

User Experience & Interface Enhancements

Beyond raw feature improvements, Photo AI 4.0 brings important UX refinements designed to streamline your workflow. The filter stack (i.e., the list of enhancements you apply) now shows the most recent filter at the top. This “reverse chronological” ordering makes it easier to track your editing history and adjust recent changes. (Topaz Community)

The right-hand panel has also been redesigned: Autopilot suggestions, filter options, and controls are now presented in a more intuitive layout. First-time users will immediately see the opt-in toggle for Autopilot, while experienced users can switch between Autopilot-suggested edits and manual control with minimal friction. (Topaz Community)

Login has been improved too: Photo AI now supports Single Sign-On (SSO) using Google or Apple accounts. This makes it easier to access the software, especially for users who dislike managing yet another password. (Topaz Community)

These design tweaks are small in isolation, but together they make the app feel more modern, cleaner, and more accessible — especially for new users or those coming from other editors.

Cloud Processing Flexibility

Photo AI has long supported cloud rendering, but version 4.0 gives users more explicit control over when and how they use it. The enable / disable cloud processing toggle introduced in the 4.0.1 patch lets you choose whether to rely on your local machine or offload to Topaz’s servers. (Topaz Community)

This matters for a few use cases. If your GPU is powerful, you might prefer to process everything locally to save time or avoid credit use. If not, cloud processing can help you run intensive models like Super Focus v2 or Dust & Scratch without burdening your hardware. And because the cloud model supports many of the new features, you don’t need to compromise on quality just because your device is less powerful.

When to Use Which Feature: Practical Scenarios

Given all these new tools, here are some practical examples of when version 4.0 really shines:

• Restoring Old Photos: If you have dusty family prints or scanned film, the Dust & Scratch model is ideal. Remove surface flaws first, then run enhancements like Sharpen or Face Recovery, giving you clean, restored images.

• Sharpening Soft or Blurry Images: For out-of-focus or motion-blurred shots, Super Focus v2 is the go-to model. Use it early in your workflow, adjust the strength after processing, and then layer in other enhancements.

• Portrait Retouching: The enhanced Face Recovery tools make it easier to fix and improve low-quality portraits. Use neck and hair controls to blend recovery more naturally, and let Autopilot decide appropriate strength based on face size.

• Lighting Correction: Use Adjust Lighting v2 when your image has blown-out highlights or blocked-up shadows. The highlight/shadow sliders let you tune exposure precisely, and the colour-correction toggle helps preserve your original tones.

• Workflow Optimization: For batch processing, let Autopilot make initial enhancement suggestions. Then review and tweak manually using the filter stack. Use the new filter ordering to track your workflow more easily.
  
  

Limitations and Considerations

While Photo AI 4.0 brings many major improvements, there are some important caveats:

• Dust & Scratch Limitations: According to user reports, the model doesn’t always perfectly remove very deep or large scratches, and some very stubborn marks may still require manual healing. (Topaz Community) The healing brush helps, but expect to spend time on tricky areas.

• Hardware Constraints: Running Dust & Scratch locally can demand significant VRAM. For users on older hardware, using cloud rendering may be necessary. (Reddit) Also, some GPU-specific bugs remain (Topaz notes fixes for certain NVIDIA series in the 4.0.1 patch). (Topaz Community)

• Autopilot Isn’t Perfect: Although smarter, Autopilot’s suggestions aren’t always ideal. There may be images where its “intelligent selection” misidentifies the key subject or applies enhancements you don’t want.

• Face Recovery Constraints: Face Recovery works best on human faces; the model doesn’t reliably detect or improve animal faces. (docs.topazlabs.com) Also, if the original face is extremely low resolution, detail reconstruction is limited and may look overly smooth or “plastic.”

• Personalization Trade-offs: Version 4.0.1 introduces a toggle to disable personalization. While this may preserve privacy or consistency, it can limit how well the software adapts to your editing style.

• Cloud Credit Usage: Although many powerful models run in the cloud, users with heavy cloud usage should still watch their credit usage, depending on their Topaz plan or subscription.

Minimum Requirements for Topaz Photo AI

Conclusion

Topaz Photo AI version 4.0 is a major milestone for the software, bringing deeply meaningful improvements in restoration, sharpening, facial recovery, user experience, and performance. The addition of a specialized Dust & Scratch AI model opens up restoration workflows that were previously labor-intensive, making it far easier to clean up scanned film or old family photos. Super Focus v2 accelerates focus recovery dramatically, while giving users more control over sharpening strength. The Autopilot system is smarter and more flexible, with contextual suggestions, better filter ordering, and easier opt-in/out control. Face Recovery now handles blending around hair and necks, and adapts strength dynamically based on face size. Adjust Lighting v2 provides refined control over highlights and shadows, with optional color correction to preserve tone. On the UX side, improvements like SSO login, a cleaner right-hand panel, and more intuitive stack ordering enhance usability. Cloud processing remains highly relevant, with new toggles to control where computation happens.

All these updates make Photo AI 4.0 more powerful and versatile—and more appealing both to professional photographers working on high-quality restorations, and to hobbyists who want top-tier AI tools for clean-up, sharpening, and creative enhancement. While it still has limitations (especially when dealing with very damaged originals or limited hardware), version 4 offers a strong step forward and opens up new creative possibilities for many users." (Source: ChatGPT 2025)