Mask AI - I think had some problems with this but then it stopped getting supported which isn’t that good. Ideally it could at least be supported until there were 0 issues with it. Though it could also be made to create auto-masks for video too (if it could be better/cheaper than Adobe’s robobrush).
For Gigapixel - maybe it could be improved with better AI models, where you could be more specific (eg. maybe it’s a font your are uspcaling or want a better way to specify what scenery etc. is in it). Maybe it also shouldn’t limit to max 6x upscale.
For Video AI - like has been said, they removed the preview comparison (and the preview in the last change I got). So previously you could compare 3 ai models vs the original (in Video Enhance AI), now you can’t, which is a bit of a backwards step. And the face upscaling in this and gigapixel is also a bit limited (and can decrease res if used sometimes. I think they were going to train it on higher res faces, but that also would mean it wasn’t the actual face of the person that it would be replacing it with).
I agree (and might have posted similar suggestions before) that they could add better upscaling (as well as the other functions like interpolation). The suggestions of giving it images of things in the scene (such as higher res images of the person’s face, different backgrounds that are in the scene etc. could help). There’s also a new AI generator function that can create a “1 shot” output of a face based on 1 image. Maybe something like that could be used to help for video AI for keeping it consistent. But really it should also be possible for Video AI to use a specified number of frames in the vid to help to upscale the current frame, which it doesn’t really seem to do (or at least the current upscaling models don’t specify how many frames they take into account). eg. if something is close to the camera in some frames, why not use that info for when it’s further away and blurry (or combining info from multiple frames). Like I think I read super resolution does.
The first post says “Natural and photorealistic results, vs fantastical: for your photos and videos where preserving reality matters.”. Though with generative AI, if there could be a separate option for fantastical maybe that could be a good option. eg. if you need to add certain effects to a video - maybe by typing prompts like in stable diffusion, eg. “add an explosion in the background” or “add a dragon here”.
But an important point is the apps are about AI, and generative AI (the first post says “We believe that generative AI should improve your work rather than replace it”). There are also other apps that can do AI (eg. midjourney, stable diffusion) that have come out for images, but the problem with those is the copyright risk (the companies related to those have had lawsuits). Adobe indemnifies Firefly (their AI) users against 3rd party copyright claims - they basically guarantee that their AIs (though they use the cloud which isn’t as good as local) were trained on their stuff where there shouldn’t be any issue. Ideally, especially if Topaz is going to go more into generative AI, Topaz too could offer an indemnity to protect users from any potential copyright claims and declare how their content was trained on definitely 100% okay content. That would give more confidence for users buying their products and more confidence to use them for Pro use.
It seems to me that the answer to your question has already been given here
Reasonable requests and it seems to me that until now they have been largely, and I think deliberately, ignored.
I continue to believe that if the foundations of a building are not strengthened, it is useless and dangerous to build new floors.
I think it’s silly for Topaz to compete with features offered by large companies like Adobe: Topaz Photo AI must return to being a top product with few but excellent features not present with equal results in other applications.
Re: Choice: It would be nice if I can choose whether to use, say, generative AI or not. Sometimes I just want to sharpen or Denise, not generate something instead. And, as pointed out in other posts, if using these tools might violate some rules for images or competitions or whatever, I’ll choose to not use the tool. And that becomes a habit.
Re: Consistency: Topaz Labs has demonstrated a habit of having products, then revamping, recombining, deprecating, etc., them all. To the degree that I hesitate to pay for things since I thought I already did or I don’t know how long it will be there (or, that I have to go back and re-examine things to ensure that I haven’t been ripped off). I think that having some consistency in the product and the handling of things will go help a lot.
First for a grand product (that I’ve used from almost inception), second for inviting ideas so openly. I speak as a photographer primarily, so will leave video stuff to the experts.
My high-level view is that there’s an increasingly important place for a tool dedicated purely to enhancement: mission statement: making the very very best of your image. Topaz is the home of the real image.
Other image manipulation tools have bloated out to add and alter in addition to enhancements. Topaz can continue to offer better and better performance, enhancing powers, with deliberately throttled back controls like levels, highlight/shadow etc. to keep images real. For example:
Adaptive white balance correction
Realistic highlight/shadow correction
Easier localisation of non-denoise, non-sharpen areas
Improved export panel e.g. more naming scripts e.g. serial number
Faster refresh of screen image
Zoom with mouse scroll + modifier key
Align Gigapixel abilities with PhotoAI so there’s no confusion/doubt about which is better
It’s cool to get onto the generative bandwagon, but I see Topaz’s USP is more about smart interpolation
Crop tool: can we create our own presets please: a long list of settings we don’t use is not helpful
Crop tool: can we have a choice of grid: personally I hate rule of thirds; phi more useful. Also thinner guide lines
When turning on a control can we have the slider(s) or options panel drop down automatically?
I was very happy that the Gigapixel AI was renewed. It would be nice if Denoise AI and Sharpen AI were renewed as they are very useful applications. Mainly Denoise AI is what I use on a daily basis. Although Lightroom’s new denoiser is very good, it can only work with raw images.
I don’t use Topaz’s applications as intended, I often use them for retouching old photos, because I occasionally get quite good results or partial results with the exaggerated use of parameters.
I have some of these which were poorly converted by relatives whom now seem to have lost them. I found neat video to be the best wa to clean up the dirt. took a few passes though. Used DaVinci resolve as its host program.
Probably like me yours we interlaced into some low res format. Id sort that out first, then the dirt, then look at any enhancements.
To be honest, what brought me to the Topaz suite a few years ago was the Denoise AI. But at the time I bought into everything available and have very much appreciated the benefits that Gigapixel has brought for taking older photos and increasing resolution. Nothing is perfect but the progress has been good.
The other product I bought for creative use was Topaz Studio. I am sad that it’s been abandoned as I have some marvelous paintings rendered from photos I took.
As for Photo AI, even though I was a beta tester and still have it, I don’t really like it that much. The individual tools still do a better job, in general.
But then, I’d rather have my camera produce the image I want and just use these tools to help resolve issues the camera limitations can’t prevent (ie Denoise for a Milky Way shot).
If someone could introduce creative AI to a Topaz Studio 3, I’d be all over that. That’s not photo editing, it’s art creation with a base image from my photograph.
I don’t care for the Adobe Suite, but for processing of Raw images, I prefer DXO PhotoLab.
If Gigapixel were to follow the path of Denoise, I’ll likely come to the day where I don’t upgrade my Topaz anymore.
???
Sorry but I didn’t understand your statement: most image editors, e.g. like Adobe’s LightroomClassic, operate locally without using the internet. So?
If you look at info on Adobe’s generative AI functions I think that those often use the cloud/internet to work, eg. Adobe Firefly. Adobe Photoshop’s neural filters - a tutorial is saying some of them work online.
Ok, but we can’t compare the functionality of a hill (Topaz PAI) with that of Everest (Adobe). However, if the functionality of Topaz PAI continues to grow using AI, Topaz will also be forced to use the cloud/Internet unless it downloads dozens and dozens of gigabytes on each personal computer: the installation of PAI is already quite heavy.
It should still be more secure (and probably faster - depending on connection speed and the GPUs in the PC) than using adobe products that need to use the internet for certain things (probably especially if it’s video). Maybe if you do work on client’s projects they might not want their content being sent to Adobe or some other company. And the AI models could just be a one -time download (unless updating to a newer version of them). So I think Topaz’s way of having them stored locally is the better idea. It also means people can continue working even without an internet connection (except for the fact that Topaz sometimes requires you go log in and authorize it).
I agree but we are still comparing an ant to an elephant . What Topaz does today, and that’s fine, is only possible because the features are few: if they continue to increase, I believe an approach of this type will no longer be possible.
Possibly but they say in the first post “On-device processing, vs cloud: for a faster workflow, more editing options, and private processing.”. So I assume that’s their priority.
Maybe for some things they might increase the minimum requirements. Maybe GPUs with more VRAM and faster processing might become cheaper and they might make those the minimum or that you need 2x certain GPUs for some things. Or maybe there will be cheap AI chips/“AI accelerators” that might be better/more cost effective for it than GPUs.
As a longtime customer, I generally avoid all the chatter that is written with varying agendas.
I would hope original products maintain their development and progress quarterly. Rather than seeing new products being developed at the expense of main applications Denoise, Sharpen, Gigapixel. Where I have spent my hard earned money. The stink I’m smelling is inexcusable when all efforts are spent developing Photo-AI and Video-AI. Your marketing efforts stringing customers along is shameful. Whatever happened to Topaz Studio 2? Is the business model now one where a product is started, get a sizable group using the latest application(s) only to drop it and start new applications down the road.
Thanks for nothing. This customer is saying goodbye!
All companies have a right to create new products, deprecate products or discontinue support for older products or indeed raise or lower prices. This occurs across the world in all industries and is typically driven by innovation and demand.
Even the big software companies do the same … e.g. Adobe.
You have the right to choose to use or not any discontinued products from any company for as long as you like.
I appreciate your email and your query. There isn’t a lot I have appetite for (other than doing even better what you already do) but there are, of course, some things.
Have a way in the product to do horizontal and vertical perspective changes. I think that I get slightly better results when this is done prior to sharpening/upscaling. If you think about going this route, I suggest taking a look at how the lowly Snapseed product does this, which is to allow you to drag points near each of the four corners of the image to get the result you want. I find that this is much faster and much easier than other methods.
I would love a setting that allows me to automatically open the image in my editor of choice following export. (My editor of choice, oddly enough, is Luminar 4.)
I’d also like to add the it is a crying shame that you have dropped support for Studio 2. The looks in Studio 2 are a lot of fun, but the feature that I really would like in a standalone that could be used as a plugin is the Digital Frame filter. There’s no other product that I am familiar with that matches the functionality of this filter in terms of the comprehensiveness of settings combined with easy to use power and excellent execution.
I really appreciate, btw, the ease of updates and the frequency.
There are millions of discolored old photos, BW photos, textured photos (scans of photos with textured paper) - all those are really difficult to bring back without manual work - removing dust, restoring tear and scratches, colorizing those all while considering that grain in old film shots is different from that in modern cameras - a workflow for restoring old photos would be a fantastic addition to PhotoAI and would bring quite a few people trying to keep old photos to life.
Regarding some of the old software suits from Topaz, if Topaz definitely decided that those are “abandoned” - why not open source them? Community can update Qt4 to Qt 6, make it compile on all platforms and introduce improvements and fixes. Topaz Studio can have 2nd life and since there is no commercial interest it will be no loss for the company. I’ve paid for every single product Topaz created in the last 10+ years - I would love to see Topaz Impression, Adjust, etc be usable again.
Good afternoon Mr. Yang, thank you for providing this update. One suggestion that I do have to open your products up to more potential customers is to offer an “All-Access” subscription at a fair monthly price (perhaps $29.99/mo). In this model, you’d include access to all applications (including all updates/support) for one low monthly price.
This would create a persistent revenue stream for you as well as entice more customers to use the product (so that they aren’t having to pay $299 outright for just Video AI or $199 outright for just Photo AI).
Now, understanding that not everyone would like to use this model, I’d still retain the option to purchase the product outright for those that want it.
. What Topaz does today, and that’s fine, is only possible because the features are few: if they continue to increase, I believe an approach of this type will no longer be possible.