Awesome demo, Rens! I can definitely see how it will be a time saver even in it's current state. I'm sure that as the models get updated, it will get even better.
Thanks Trent! That's exactly what I thought. Right now it will already save us time, but it can only get better over time. Exciting times for the Affinity Photo community. If you're interested, feel free to join the Affinity Photo Creatives community. You're welcome to share your videos and wisdom there, I'm sure the members would love it 🔥
You're welcome man! Just checked out your channel, you're crushing it! If I may give a thumbnail tip: Sort your videos by popular, and check the 2nd and 3rd vid. You uploaded both of them very recently and they outperform the rest. I'd give it a shot and double down on that thumbnail style!
A small step in the right direction, at least they are moving forward in line with other programs. Hope they offer an improved version in the official release. Thanks Rens for showing it.
Now that’s a really nice surprise. Of course it’s not perfect, it’s a beta. But so far I like it and in time it might grow better and better. It was also nice to see you being really enthusiastic!
Haha, in fact I was so happy to find out about it that I immediately hit record. So it was basically a Live demo just trying out the new tools. I'm excited for this one!
Great. What i would love to see Affinity do is get some 3D tools in there similar to illustrators (3d bevel extreu inflate tool) or the vanishing point with mulitple planes and being able to copy content on x,y,z axis. And move and place along an individual axis. But i love what Affinity is doing. :) Thanks for covering this.
I find this new tool absolutely interesting! I hope it will be implemented and made available in future AP upgrades! I didn't know they were experimenting with something like this! Thanks for sharing!!😎👌🏻
I'm using both software and Adobe is still some years ahead, BUT with this machine learning models being improved maybe we can have something to beat adobe
That's really awesome! I installed it immediately and it crashed on my first try.. Second try went better though. It works pretty good! And while you cannot open the .afphoto file in the older (2.5) version of Affinity Photo, you CAN copy layers across.. So, make your selection in the beta and copy the masked layer to the more stable non-beta version to continue your work :)
Finally, this is huge for many photographers! I've been wanting this a long for time since selections can take a long time, I just hope it's good at complicated things like hair and fur. Also, interesting that they call it machine learning and don't say "AI".
My reading of that is that they really don't want the baggage (and hype) with calling it AI. Since up until the last couple of years, ML and AI were often seen as synonyms, but lately "AI" seem to only mean a very limited selection (/models) of machine learning. So I like that they call it that =) Mind, I suspect that while I will use the tool, I'll probably not install the saliency model to get the option--it's a nice convenience and I'm sure for some workflows it'll be a game changer, but for me personally I rather use the select tool to "feel" like I have more control. After all, *I* know which one's my subject =) (again, not poo-pooing the option or people for whom the "select subject" is important, just ... not needed for my workflow).
@@Melindrea I think Affinity also is trying to stick with efficient algorithms that run locally on your own machine because if they had to run a service that the app connected to, they'd have to charge a subscription. And the lack of subscription is their main selling point. But locally run stuff is a lot more limited since you can't rely on people having a datacenter's worth of gpu's to run the model lol
@@omgbutterbee7978 As someone that has messed around a fair bit with local language models for coding, you don't really need a datacenter of GPU, especially not for something this focused. What it will do is make your local GPU speed or a CPU with AI acceleration somewhat more important. Not sure about AMD GPU's, but any fairly recent, mid-ish tier Nvidia RTX card should be more then fast enough.
I've been having such a hard time getting used to affinity. I've used photoshop for 15yrs and the way masking and clipping masks work throws me off. I'm also trying to used to the move tool and transform tool being the same thing. Kind of wish I could separate them but I'll keep trying to get used to it.
I can imagine after 15 years. However, I've got a whole lot of members inside of my community who have switched from Ps. It's just that you're so used to it, that any change throws you off. But if you take the time to get used to it, it's actually a very logical thing that for instance the move tool can also transform layers!
I like Affinity's low price (and they have a deal right now that if you own the v1 software you can update all 3 programs for $30), but good lord they don't make their workflow easy. It is 20x harder to be productive with their software versus using Adobe's famous trio of Illustrator, Photoshop and InDesign. If Affinity wants to really steal away Adobe users, they absolutely need to work on workflow. If they just offered the same key commands and toolset as Adobe, that would be a massive improvement.
They have got a long way to go, especially with 'Select Subject' and hair/fur. But I'm super excited to see where it's going and also good to know that they won't steal our work for machine learning
Hello IAMRENSI, these new features are beautiful, if they improve them a little more they will save us a lot of time instead of manual selection. 🔝🔝🔝👏👏✨👍We hope they arrive soon. 💥👍A big hello✨👍
Anyone got experience with the affinity suite and corel draw? I wonder how these stack up against each other, especially with the humble bundle still up
For a second I thought I you meant rasterize before using the 'Select Subject'-feature. That would've been a fail from my side. But I just checked, and for now it seems pretty random when it works or doesn't. So what exactly do you mean?
It's got a very long way to go yet, I made a selection much faster and more accurately with the selection brush, refinement and lasso tool compared to the AI selection Pleased to see it in beta now though Too much refining and too much tweaking to be of any real time saving yet but I know it will get there
It probably depends on what you're selecting. For simple selections it does a good job. For more difficult selections there's quite some tweaking necessary indeed. For a beta version it looks promising to me!
I hope this tool will be kept away from future releases of the so far clean Affinity Photo until it is perfect and not just 'almost right' without exceptions. Look what has happened to Photoshop - the oversold AI is an idiot which most of the time destroys critical photography.
I WANT it to learn from my data. People are too sensitive about this stuff. Anything that helps improve the feature should be fair game. They should offer an opt in mechanism in case you want it to learn based on your specific decisions/data.
@@IAMRENSIThats because some of adobe users are artists that paint and draw.. so they make that art from scratch. Mostly the one who are angry are them not photographers etc... adobe should just ask. Not force everyone. An option is nice.
so far I am not exited at all. And maybe never will be exited because all images never work - not in your cases not in my cases. Wether ON1 or PS - it`s all promising but the final results always have to be fixed and so … even the final image of your eyamples - a white dress in front of a dark background? I will never get, why this is so difficult and I am glad that it is difficult - AI is just one tool and never the solution.
It's always a good idea to have a good look and fix your final result where necessary. I don't mind that it's not perfect. It's here to speed up our workflow, not to steal our work 😅
Awesome demo, Rens! I can definitely see how it will be a time saver even in it's current state. I'm sure that as the models get updated, it will get even better.
Thanks Trent! That's exactly what I thought. Right now it will already save us time, but it can only get better over time. Exciting times for the Affinity Photo community. If you're interested, feel free to join the Affinity Photo Creatives community. You're welcome to share your videos and wisdom there, I'm sure the members would love it 🔥
I have been waiting for this!! This is getting so much closer to me completely moving on from Photoshop
Same here! Affinity is making moves 🚀
Even though not perfect this is going to be a time saver for many.
Yea, and as you said in the community it can only get better 🙌🏼
This is HUGE!!! As long as Affinity respects peoples privacy unlike others this is a great
As far as I've read they seem to be very careful with their customer's privacy. I'm sure they've learned from Adobe ;)
I'm super excited about these selection features!
Same, this is gonna be a massive time saver 🚀
OMG YAY, THE tool it was missing !
Great video, good explanation. I'm looking forward to seeing how this selection tool develops.
Thank you! If they would only get it decently good at launch it would already save me hours in the selection process :)
FINALLY! Thanks for your video on that!
You're welcome man! Just checked out your channel, you're crushing it! If I may give a thumbnail tip: Sort your videos by popular, and check the 2nd and 3rd vid. You uploaded both of them very recently and they outperform the rest. I'd give it a shot and double down on that thumbnail style!
The new features look promising, Rens!
Definitely, pretty excited for this one Bill!
A very safe implementation from a cybersecurity point of view. For webshops this is a supertime saver, since photos start with a ‘white’ background.
Yes for product photos it's gonna be perfect. Only-click selections 🙌🏼
Thanks for showing this - it is interesting to see how far behind current SOTA these apps are currently.
This will save so much time. I hope it will be available for the iPad as well.
Yes it will! Now it's just the question when it will actually be decently good :)
A small step in the right direction, at least they are moving forward in line with other programs. Hope they offer an improved version in the official release. Thanks Rens for showing it.
I'm pretty sure they will. This is probably the worst version we get to see (I hope).
now this is interesting! it's a life improvement! and just joined the Skool Community! Hope to learn a lot from fellow Affinity Creatives!
Awesome, welcome to the APC fam! 🙌🏼
Very nice demo! Thanks for sharing.
Thanks for watching 👊🏼
Now that’s a really nice surprise. Of course it’s not perfect, it’s a beta. But so far I like it and in time it might grow better and better. It was also nice to see you being really enthusiastic!
Haha, in fact I was so happy to find out about it that I immediately hit record. So it was basically a Live demo just trying out the new tools. I'm excited for this one!
Great. What i would love to see Affinity do is get some 3D tools in there similar to illustrators (3d bevel extreu inflate tool) or the vanishing point with mulitple planes and being able to copy content on x,y,z axis. And move and place along an individual axis. But i love what Affinity is doing. :) Thanks for covering this.
There's definitely some cool features they could add in the future, but this one is probably the most requested one!
This is awesome. Thanks for the heads up
You're welcome! Looking forward to what's coming!
Awesomeness! Ok, so it's not perfect but it's a great step forward.
It's far from perfect, but like this it will already save us a whole bunch of time and it can only get better 😊
man that new icon looks just like the indian coffee company logo thing you did. the upscale low res photos
I find this new tool absolutely interesting! I hope it will be implemented and made available in future AP upgrades! I didn't know they were experimenting with something like this! Thanks for sharing!!😎👌🏻
I also didn't know and I have happily surprised when I found out about it! Pretty exciting 🔥👊🏼
I feel it will be a lot better upon its official release. This is amazing! It would have saved me HOURS on this project i just finished working on. 😂🤣
Yes I think so too! Haha well in the future you will save lots of time 😅 Pretty exciting stuff if you ask me!
Great, now selecting will be much easier in the future!
Great info. Thanks. :- )
You're welcome. Exciting times ahead :)
very cool update
Yea stoked about this one too 👊🏼
Finally! 🙌 it takes me ages to make a clean mask manually 😅
Haha, well I noticed we still have to clean up the mask. But the most work is done by the tool!
I'm using both software and Adobe is still some years ahead, BUT with this machine learning models being improved maybe we can have something to beat adobe
That's really awesome! I installed it immediately and it crashed on my first try..
Second try went better though. It works pretty good!
And while you cannot open the .afphoto file in the older (2.5) version of Affinity Photo, you CAN copy layers across.. So, make your selection in the beta and copy the masked layer to the more stable non-beta version to continue your work :)
Xzenor always finding loopholes in the matrix haha. Awesome, didn't even think about that. Nice workaround!
Finally object selection come to affinity also
Yea, exciting times ahead! 🔥
Finally, this is huge for many photographers! I've been wanting this a long for time since selections can take a long time, I just hope it's good at complicated things like hair and fur. Also, interesting that they call it machine learning and don't say "AI".
I know! We all have been waiting for this for so long. So far it's quite shitty with hair/fur but I'm sure it'll improve it over time
My reading of that is that they really don't want the baggage (and hype) with calling it AI. Since up until the last couple of years, ML and AI were often seen as synonyms, but lately "AI" seem to only mean a very limited selection (/models) of machine learning. So I like that they call it that =)
Mind, I suspect that while I will use the tool, I'll probably not install the saliency model to get the option--it's a nice convenience and I'm sure for some workflows it'll be a game changer, but for me personally I rather use the select tool to "feel" like I have more control. After all, *I* know which one's my subject =) (again, not poo-pooing the option or people for whom the "select subject" is important, just ... not needed for my workflow).
@@Melindrea I think Affinity also is trying to stick with efficient algorithms that run locally on your own machine because if they had to run a service that the app connected to, they'd have to charge a subscription. And the lack of subscription is their main selling point. But locally run stuff is a lot more limited since you can't rely on people having a datacenter's worth of gpu's to run the model lol
@@omgbutterbee7978 Yeah, that too =)
@@omgbutterbee7978 As someone that has messed around a fair bit with local language models for coding, you don't really need a datacenter of GPU, especially not for something this focused. What it will do is make your local GPU speed or a CPU with AI acceleration somewhat more important. Not sure about AMD GPU's, but any fairly recent, mid-ish tier Nvidia RTX card should be more then fast enough.
I've been having such a hard time getting used to affinity. I've used photoshop for 15yrs and the way masking and clipping masks work throws me off. I'm also trying to used to the move tool and transform tool being the same thing. Kind of wish I could separate them but I'll keep trying to get used to it.
I can imagine after 15 years. However, I've got a whole lot of members inside of my community who have switched from Ps. It's just that you're so used to it, that any change throws you off. But if you take the time to get used to it, it's actually a very logical thing that for instance the move tool can also transform layers!
Not bad, but Photopea seems to do a better job with its "remove bg" function for hair and edges.
Keep in mind this is an early beta version, I assume they'll train their model and improve it over time :)
Thats crazy, we get AI tools?
Way to go Affinity!
Yes, as long as we don't go generative AI I'm happy :)
I like Affinity's low price (and they have a deal right now that if you own the v1 software you can update all 3 programs for $30), but good lord they don't make their workflow easy. It is 20x harder to be productive with their software versus using Adobe's famous trio of Illustrator, Photoshop and InDesign. If Affinity wants to really steal away Adobe users, they absolutely need to work on workflow. If they just offered the same key commands and toolset as Adobe, that would be a massive improvement.
Wonder if a processor has a neronetwork chip will make it work faster and better
At last they caught up with the rest of the software.Hopefully it works as good as in Capture One
They have got a long way to go, especially with 'Select Subject' and hair/fur. But I'm super excited to see where it's going and also good to know that they won't steal our work for machine learning
Hello IAMRENSI, these new features are beautiful, if they improve them a little more they will save us a lot of time instead of manual selection. 🔝🔝🔝👏👏✨👍We hope they arrive soon. 💥👍A big hello✨👍
Yea I totally agree! Not sure when it will come out, but it def looks promising! 🫡
Do we know if this will be available on the ipad version?
Yes, it's coming to iPad as well! 🔥
Anyone got experience with the affinity suite and corel draw? I wonder how these stack up against each other, especially with the humble bundle still up
I see you’re using Mac also.
Which download should I choose?
I can’t see a dmg-package there.
There is actually a DMG-package. It took a bit before I found it too, but it's in the list
I mostly edit on my iPad Pro will this beta with subject selection work?
I’ve heard the beta is still pretty shitty, but the update will come for iOS as well 🫡
@@IAMRENSI good to know, I feel like betas for IOS usually suck more then the actual beta for desktop version. I’ll wait this one out..
@@IAMRENSI Also do you think in time they’ll come out with generative fill, somehow?
awesomeee
Finallyyyyy! 🙌🏼
I hope it works on iOS too
Yea, they'll get it over to iPad too. Read it in the forum! 🙌🏼
@@IAMRENSI great 👍
Rasterise first !!
For a second I thought I you meant rasterize before using the 'Select Subject'-feature. That would've been a fail from my side. But I just checked, and for now it seems pretty random when it works or doesn't. So what exactly do you mean?
It's got a very long way to go yet, I made a selection much faster and more accurately with the selection brush, refinement and lasso tool compared to the AI selection
Pleased to see it in beta now though
Too much refining and too much tweaking to be of any real time saving yet but I know it will get there
It probably depends on what you're selecting. For simple selections it does a good job. For more difficult selections there's quite some tweaking necessary indeed. For a beta version it looks promising to me!
the new updates really suck!
Haha okay, why do you think so?
I hope this tool will be kept away from future releases of the so far clean Affinity Photo until it is perfect and not just 'almost right' without exceptions. Look what has happened to Photoshop - the oversold AI is an idiot which most of the time destroys critical photography.
I WANT it to learn from my data. People are too sensitive about this stuff. Anything that helps improve the feature should be fair game. They should offer an opt in mechanism in case you want it to learn based on your specific decisions/data.
Same here, I don't mind them using my data in order to improve their software!
@@IAMRENSIThats because some of adobe users are artists that paint and draw.. so they make that art from scratch. Mostly the one who are angry are them not photographers etc... adobe should just ask. Not force everyone. An option is nice.
so far I am not exited at all. And maybe never will be exited because all images never work - not in your cases not in my cases. Wether ON1 or PS - it`s all promising but the final results always have to be fixed and so … even the final image of your eyamples - a white dress in front of a dark background? I will never get, why this is so difficult and I am glad that it is difficult - AI is just one tool and never the solution.
It's always a good idea to have a good look and fix your final result where necessary. I don't mind that it's not perfect. It's here to speed up our workflow, not to steal our work 😅
If tool isn't there check 'Customize Tootls'. I had to add it from there. Thanks.
Thanks for the tip! 🔥