September 10, 2022
Sharpen AI is the third of our four core apps to gain native Apple M1 support, joining DeNoise AI and Video Enhance AI. Building Sharpen AI to natively support Apple’s integrated “System on a Chip” architecture allows us to provide Apple M1, M1 Pro, and M1 Max users with even faster performance. Everything from app load times, single image editing, and batch processing will be noticeable faster.
In our first benchmark test, I compared the time it took to load Sharpen AI v4.0 from a cold boot using both the native M1 and Rosetta versions. As you can see, the native M1 build provides a 75% faster load time when compared to the Rosetta emulation version.
Next, I wanted to compare processing speeds for resource-intensive tasks such as batch processing. I opened a mix of 10 RAW files from the Sony a7R IV, Sony a7 III, and Canon EOS 5D Mark III cameras, and configured a batch processing job that would auto-select the AI model and parameter settings.
When comparing the batch job performance between the native M1 build and the Rosetta emulation build, the processing time is almost 20% faster. Even more impressively, when comparing the native M1 build to a comparable Macbook Pro with the Intel chipset, the batch process job was almost 70% faster.
One of the challenges Sharpen AI users sometimes face is determining which of our nine models would provide the best results for their photos. Even with our existing Auto Model selector, Sharpen AI wouldn’t always choose the optimal model based on the qualities and needs of the photo. Each of Sharpen AI’s nine models (the three primary models—Motion Blur, Out of Focus, and Too Soft—and their respective sub-models—Normal, Very Noisy, and Very Blurry) were trained for very specific use cases and it often required users to go through every option before determining the best one. And beyond determining which model is ideal, you also have to identify the optimal parameters, either manually or by enabling the Auto Parameter option.
That is why we’ve built the new Standard model. The immediate benefit of using the Standard model with the Auto Parameters option selected is that it will cover all of our existing models, and provide better results when compared to the existing Auto Model/Parameter options. Our goal is that this model will provide “one-click” best results regardless of the photo being worked on.
Another benefit that the new Standard model provides is improved distinction between—and detection of—motion blur and lens (or focus) blur. To illustrate this more clearly, check out the following two photos. Both are using the new Standard model with the Auto Parameters option enabled. In the first photo, the model recommends a reduction in motion blur, and in the second photo, the same model recommends a reduction in lens blur.
Additionally, if you want to further refine the results of the Standard model, you can adjust the four sliders to dial in the specific sharpening output.
Along with the recent update to DeNoise AI, you will now benefit from a refresh to Sharpen AI’s user interface. The improved colors and increased contrast will make using Sharpen AI clearer and easier. We’ve also applied a number of stability improvements, specifically for Adobe Photoshop users who prefer using Sharpen AI as a plug-in. We’ve added some cool features for Wacom pen tablet users, too. In additional to improving overall tablet support, Wacom users can use their pen’s eraser to erase strokes made in Sharpen AI’s in-app masking mode. Finally, we’ve further improved stability by squashing a bunch of bugs from previous versions.
© 2005 - 2024 Topaz Labs. All rights reserved.
