This release for Gigapixel brings a number of reliability improvements and fixes under the hood to the application, paving the way for future improvements to our newer features, including the Recovery model and Comparison view. We are also looking forward to improving the results and performance of some of our new models, so please stay tuned for some exciting updates.
Gigapixel is getting better every single day behind the scenes.
As always, a full change log is below.
Comment below to share any feedback or report any issues you encounter with this release. Weāll be updating Gigapixel regularly to address your feedback and any issue reports. We would also love to see some your great upscaling results!
Hey Dakota, my intention is not to sound ungrateful for the work the team put into the update but I doesnāt resolve the most glaring issue, recovery still uses 100% of the resources and can only be used on small images. I purchased the GP update and just upgraded my system. With the new version of GP I am right back were I started or worse since my GPU temperatures is fairly extreme when using recovery.
I canāt use 100% of the application I purchased or am very limited when I want to use recovery. It would be helpful to understand how this could even be remotely acceptable to your client base. When I purchased the application my expectation is that I will be able to use the functionality without restrictions.
Can you roughly say what your processing times are for the recovery model? What picture size do you use as an input? And what are your system specs after upgrade?
Does your system meet the requirements for the Recovery Model?
The Recovery Model is based on stable diffusion and requires a decent GPU to work effectively. It is the only way to generate extra details in the image. If you donāt have a decent GPU, donāt worry; you can still use other traditional upscale models that require fewer resources. Simply ignore the āRecovery Model.ā
To be honest, the āRecovery Modelā in GAI already uses very little resources compared to other state-of-the-art AI upscale algorithms. The more resources it utilizes, the better the image quality becomes. Topazās developers have made significant efforts to optimize performance for users with subpar hardware.
Personally, I wish Topazās developers prioritized image quality above all else. I donāt mind it use lots more resource or processing time, as far as it give the best image quality.
Hi there. I have the three seperate Topaz apps on my desktop - Denoise, Sharpen and Gigapixel which seem to upgrade automatically. Does this basically have everything that Photo AI 3 has? Or is it worth buying Photo AI as well to get the latest abilities that your software contains??
Windows 11 Home 64-bit
Processor: AMD Ryzen 7 5700G with Radeon Graphics (16 CPUs), ~3.8GHz
Memory: 32768MB RAM
Card name: AMD Radeon RX 6600 XT
AMD Radeon Graphics Processor
A couple points here for you guys.
My system meets and exceeds the requirements
Whether GAI already uses very little resources compared to other state-of-the-art AI upscale algorithms is not germane to the conversation. I upgraded to my current system based on a direct recommendation from Topaz.
I stop processing @ 74Ā° C as the temp continues to raise and processing is projected at hours not minutes. I am not going to stare at my computer for hours watching the process to ensure it doesnāt overheat.
The goal post should not continue to move with regards to the āminimum requirementsā. My system is only a couple months old and purchase with Topazās recommendations. I canāt fully use the application and have limitations. And multi-tasking is not an option when running GP recovery, my system becomes useless.
Hello msh, the Recovery model is currently in Beta - it is our most powerful diffusion model but produces phenomenal results in certain scenarios where our other upscaling models cannot.
However, we are certainly aware that a faster version of this diffusion model with lower system requirements would be helpful, and would also allow for upscaling larger images using this model. Topaz models currently run on the device instead of in the cloud, so thereās a lot of work required in order to get there.
But what I can say is that we have an amazing and talented team tackling some of these challenges, and since youāre already in our Beta program - you may want to keep an eye on your email!
Hey Daniel - Gigapixel is an upscaling tool focused on whole-image processing, and includes some upscaling models and capabilities as a result that arenāt in Photo AI. It does have sharpening and denoise capabilities but they are attached to our upscaling models
In Photo AI you can process RAW files and iteratively denoise and sharpen images with intelligent masks which you cannot do in Gigapixel. You can also upscale in Photo AI, but if that is a heavy part of your day-to-day you may still want to own Gigapixel.
It depends on your workflow. If you often process RAW photos in addition to upscaling you may want to consider Photo AI.
Thanks for replying and explaining about GigaPixel which I have already . My point primarily was that I also own SharpenAI and Denoise AI separately and was wondering if they are as good as Photo AI? Iām wondering if Iām wasting my money by spending $199 on Photo AI when I already have the three other products already? Ty
Hey Daniel, Iād say the same comments go for SharpenAI and DenoiseAI. Photo AI gives you the ability to apply multiple denoise or sharpen effects to different selected areas of the image, among other things. The 4-up comparison view is one feature that is not currently in Photo AI, and there are slight differences to the available models, but Photo AI will also be receiving some new capabilities this year in those domains and will be receiving any new updates for sharpening and denoising.
GAI 713 on an Intel Mini (64GB, 2TB, Ventura) with AMD RX 5700 XT does not engage the 5700. Have checked āPrefer external GPUā in Info window, also have selected the AMD 5700 in āPreferencesā. Activity monitor shows zero eGPU usage.
On a MacBook Air (M2, 24GB, 2TB, Ventura), GAI 713 engages the two graphics cores nearly 100%.
Speed on the Intel system is uselessly slow. Iāve done the usual shutdown/unplug/replug/restart hocus-pocus without effect.
Hey Michael, did you recently upgrade your GPU by chance? You may not have the models downloaded properly. You can try re-installing the app which may fix the issue - also feel free to reach out to our support team who can assist you with this.
No, the Intel/AMD combo has been unchanged for two years except for Ventura updates, the last of which happened a few months ago. Iāll try reinstalling, will report what happens. Thank you.
TPX, first my comment was directed to you and while you may think the higher temps are fine I will manage my system and the expectation for performance. I noticed that you do not address that when running recovery my system doesnāt allow multi-tasking which is a significant challenge for me. Personally your point is moot as what you and I deem appropriate are vastly different.
Dakota, I appreciate the response and acknowledgment that there is room for improvement and that the team are prepared to tackle the challenge. It most appreciated!
