I have an image that was initially generated by Midjourney and then upscaled using Topaz Gigapixel. Here are the details of the workflow and the issue I'm encountering: Original Image Details: Dimensions: 10752 px x 7168 px Size: 441.0 MB Resolution: 300 pixels/inch Print Size: 91.03 cm x 60.69 cm I resized the image to 50 cm x 40 cm. The new pixel dimensions are 5906 px x 4724 px. The resolution remained at 300 pixels/inch. The file size changed to 159.6 MB. Export Settings: I exported the image as a JPG with the following settings: Format: JPG Quality: High Image Size: 5906 px x 4724 px Scale: 100% Resample: Bicubic Automatic Color Space: Convert to sRGB Issue After Exporting: Upon reopening the exported JPG file in Photoshop, the image size changed to: Width: 208.35 cm Height: 166.65 cm Resolution: 72 pixels/inch Dimensions: 5906 px x 4724 px File Size: 79.8 MB My question is : Why does the print size of the image change to 208.35 cm x 166.65 cm with a resolution of 72 pixels/inch when reopening the file, even though it was set to 50 cm x 40 cm at 300 pixels/inch before export? Could the upscaling process in Topaz Gigapixel affect the resolution and dimensions in a way that leads to this discrepancy? How can I ensure that the image maintains the correct print size of 50 cm x 40 cm with a resolution of 300 pixels/inch throughout the entire workflow, including after exporting and reopening the file?
This is a massive white pill for digital artists. AI will never be able to truly replicate the human eye for detail. It may surpass us in many ways, but it will always fall flat creatively. Thanks Unmesh. Great video, as always!
I'm 36 years old now and throughout my life I've heard again and again "technology will never do this or that" and then it did, and it did, again and again
This video only shows "Chatgpt" is not capable of editing a photo "yet". And it is not its priority. There are many trained or training AI's out there and they can easily understand what red blemishes are.
@@phozelno they dont, ai image generation engines are extremely dumb and inefficient when it comes to understanding the physical world because they are not physics engines, they have no clue about anything that is physical. All that ai is capable of doing is remember what it was told about a specific number of pixels in a specific situation. Maybe in the future when they will teach AI about physics, things will change but it doesnt seem like we are going in that direction, all they money are currently dumped in the diffusion technology
Wow, blown away by how much better your edits are than the ChatGPT edits. Haven't used Chat GPT, and after watching your video, I'm not sure I will be trying it any time soon.
Absolute fantastic - you made my day with "Photoshop is simple if you just understand the concept"... I'm not sure if I understand the concept, but I follow all your videos, hoping that one day I'll understand the concept! Thanks a lot for your work and videos
In my experience, ChatGPT calls on DALL-E (or maybe some other AI art program) to make the requested changes to the photo. But ChatGPT never looks at the resulting edited photo to see if it's what you asked for. And DALL-E prompt engineering is an art all its own, which ChatGPT isn't very good at. For example, I give ChatGPT a photo and ask it "Add a second car in the background." Even though DALL-E failed to do this satisfactorily, ChatGPT unknowingly replies "I have added a second car to the background."
Love your tutorials. Thanks to you, Ai will be bettered (if that is a word) by tomorrow'ers (if that is a concept) algorithm. In all seriousness, your work is superb!
I asked AI to delete a tree stump that was next to a moose. It did delete the stump, but it put a African wild dog in its place. So now I have a picture of a moose licking an African wild dog.😆
Fantastic, but if you want a real challange try it against someone using a local ai, like stable diffusion with all its built in plug ins. :) (It will be a lot more even, though ps will still win some of the challenges or will be needed for slight improvements)
Hey Umnesh, I think your prompt looked a bit too simple compared to what I usually see. To get better results you gotta give it a scenario, the task, and the goal clearly. I think that would yield a better result and help decide whether we'll all be jobless or not. 😅😅
All that AI did for me as a graphic designer was flood the vector websites with a shit ton of trash ‘illustrations’ that now I have to sift trough 10 times harder than I did before AI
Can PS do a decent job converting "photo to comic book style art?" That is, with thin black outlines, better if the lines are of variable thicknesses as if hand-drawn wth a pressure sensitive tablet. I get rather spotty results on Leonardo AI, once in a while surprisingly brilliant ones.
yeah because chatgpt is a language model, there are tool that is specifically designed to edit photos and videos and chatgpt is used for text based academic help
Hello Unmesh, you saved my day. Please more of these videos. AI must be unmasked. Lots of people are blinded by it. But AI is and will remain stupid. The term artificial stupidity would be much more accurate. I hope the hype dies down soon and we return to giving the spotlight to creative people. But honestly also includes the fact that Adobe is just as bad in the field of AI. 90% of the results are completely inadequate for me. I don't know how dependent you are on Adobe. But it would be nice to reveal the inadequacies of AI in Photoshop. This company has also allowed itself to be clouded by this hype. It has gone so far that the CEO of Adobe said in an interview that we will save ourselves the photo shoot in the future - make our AI. 😂 I could show dozens of examples of how bad Adobe's AI model is. People, be aware that AI is only a help in small areas. It does not free you to continue working on your skills. AI never beats human creator, not now and in the future, who knows.
What if....we start feeding ChatGPT of actual videos of master Photoshop users doing the editing?! Hmm....it can't do it yet because of lack of relevant data... Resistance is futile...
this is kind of misguiding chatgpt is not meant for all this. comparing with photoshop and saying you are the control is actually understating its potential. compare with comfyui then especially controlnets using flux models.
This video only shows "Chatgpt" is not capable of editing a photo "yet". And it is not its priority. There are many trained or training AI's out there and they can easily understand what red blemishes are.
No. I think ChatGPT just sends the photo off to DALL-E to do the actual editing. This is how many specialized ChatGPT add-ons work. They're just front-ends to other AI programs. ChatGPT doesn't look at the finished product once DALL-E has completed it.
AI generation is cool but you have no control over it. Photoshop and other tools let you create something with absolute full control. Like an artist. That is something that can never be replaced. It’s really a question about weither you want to be the artist or if you want to let AI be the artist. If you want to be the artist and to actually create something, then AI can never give you that. AI is is just an automatic tool that serves you something. You tell it something and it gives you something back. That is not creative or creation. Photoshop on the other hand is a tool to use to be able to realise your artistic ambitions. Like a brush is to the painter.
I have an image that was initially generated by Midjourney and then upscaled using Topaz Gigapixel. Here are the details of the workflow and the issue I'm encountering: Original Image Details: Dimensions: 10752 px x 7168 px Size: 441.0 MB Resolution: 300 pixels/inch Print Size: 91.03 cm x 60.69 cm I resized the image to 50 cm x 40 cm. The new pixel dimensions are 5906 px x 4724 px. The resolution remained at 300 pixels/inch. The file size changed to 159.6 MB. Export Settings: I exported the image as a JPG with the following settings: Format: JPG Quality: High Image Size: 5906 px x 4724 px Scale: 100% Resample: Bicubic Automatic Color Space: Convert to sRGB Issue After Exporting: Upon reopening the exported JPG file in Photoshop, the image size changed to: Width: 208.35 cm Height: 166.65 cm Resolution: 72 pixels/inch Dimensions: 5906 px x 4724 px File Size: 79.8 MB My question is : Why does the print size of the image change to 208.35 cm x 166.65 cm with a resolution of 72 pixels/inch when reopening the file, even though it was set to 50 cm x 40 cm at 300 pixels/inch before export? Could the upscaling process in Topaz Gigapixel affect the resolution and dimensions in a way that leads to this discrepancy? How can I ensure that the image maintains the correct print size of 50 cm x 40 cm with a resolution of 300 pixels/inch throughout the entire workflow, including after exporting and reopening the file?
![I Replaced ALL my ADOBE APPS with these [free or cheaper] Alternatives!](/img/n.gif)
"You will be the Master". Couldn't agree more. This talk is more proof that AI won't ever get as good as a skilled artist.
„Tofuburger-Menu“ hit me hard. 😂
I have learned more about Photoshop from your videos than anywhere else in the last 20 years. Thank you so much! I love your videos.
colorizethis AI fixes this (AI image colorization). ChatGPT controls Photoshop edits!
I have an image that was initially generated by Midjourney and then upscaled using Topaz Gigapixel. Here are the details of the workflow and the issue I'm encountering:
Original Image Details:
Dimensions: 10752 px x 7168 px
Size: 441.0 MB
Resolution: 300 pixels/inch
Print Size: 91.03 cm x 60.69 cm
I resized the image to 50 cm x 40 cm.
The new pixel dimensions are 5906 px x 4724 px.
The resolution remained at 300 pixels/inch.
The file size changed to 159.6 MB.
Export Settings:
I exported the image as a JPG with the following settings:
Format: JPG
Quality: High
Image Size: 5906 px x 4724 px
Scale: 100%
Resample: Bicubic Automatic
Color Space: Convert to sRGB
Issue After Exporting:
Upon reopening the exported JPG file in Photoshop, the image size changed to:
Width: 208.35 cm
Height: 166.65 cm
Resolution: 72 pixels/inch
Dimensions: 5906 px x 4724 px
File Size: 79.8 MB
My question is :
Why does the print size of the image change to 208.35 cm x 166.65 cm with a resolution of 72 pixels/inch when reopening the file, even though it was set to 50 cm x 40 cm at 300 pixels/inch before export?
Could the upscaling process in Topaz Gigapixel affect the resolution and dimensions in a way that leads to this discrepancy?
How can I ensure that the image maintains the correct print size of 50 cm x 40 cm with a resolution of 300 pixels/inch throughout the entire workflow, including after exporting and reopening the file?
This is a massive white pill for digital artists. AI will never be able to truly replicate the human eye for detail. It may surpass us in many ways, but it will always fall flat creatively. Thanks Unmesh. Great video, as always!
I'm 36 years old now and throughout my life I've heard again and again "technology will never do this or that" and then it did, and it did, again and again
This video only shows "Chatgpt" is not capable of editing a photo "yet". And it is not its priority. There are many trained or training AI's out there and they can easily understand what red blemishes are.
@@phozelno they dont, ai image generation engines are extremely dumb and inefficient when it comes to understanding the physical world because they are not physics engines, they have no clue about anything that is physical. All that ai is capable of doing is remember what it was told about a specific number of pixels in a specific situation. Maybe in the future when they will teach AI about physics, things will change but it doesnt seem like we are going in that direction, all they money are currently dumped in the diffusion technology
Lol sure naive one
I wouldn’t bet on that
G.O.A.T of Photoshop
Thanks, man. I will sub. You are my photoshop mentor.
Hey ... Please create a tutorial on how to alter hands or alter hand gestures
Thank you for continuing to advocate for the importance of learning Photoshop. Too many people have grown complacent with relying solely on AI.
Wow, blown away by how much better your edits are than the ChatGPT edits. Haven't used Chat GPT, and after watching your video, I'm not sure I will be trying it any time soon.
Absolute fantastic - you made my day with "Photoshop is simple if you just understand the concept"... I'm not sure if I understand the concept, but I follow all your videos, hoping that one day I'll understand the concept! Thanks a lot for your work and videos
1:06 thank you Umesh for considering us vegetarians too
Everything is piece of cake for unumesh
Awesome dear
God bless you
I haven't laughed so much in a long time. You are divine.
🤣😂🤣😂🤣😂🤣😂🤣😂🤣😂
you should try some image generator.. like midjourney, Leonardo, Comfyui..
You are best because of you, I have learned lots of things in Photoshop 😃
WOW! What a comparison. Who would use anything other than PS. Thank you for sharing.
Soon 1000 vids on the YT-channel Congratulations in advance! Been loving your tutorials for YEARS (what is it, 7? 8 now?)!!
Can we create plugins via chatgpt like retouch for me
In my experience, ChatGPT calls on DALL-E (or maybe some other AI art program) to make the requested changes to the photo. But ChatGPT never looks at the resulting edited photo to see if it's what you asked for. And DALL-E prompt engineering is an art all its own, which ChatGPT isn't very good at.
For example, I give ChatGPT a photo and ask it "Add a second car in the background." Even though DALL-E failed to do this satisfactorily, ChatGPT unknowingly replies "I have added a second car to the background."
chatgpt is a large language model which is just one type of ai and isnt designed for photoediting, there are other specialized tools for that
i liked your tutorials = best over all i saw at you tube. Great work
Love your tutorials. Thanks to you, Ai will be bettered (if that is a word) by tomorrow'ers (if that is a concept) algorithm. In all seriousness, your work is superb!
And there you go Gentleman and Ladies... Unmesh the Photoshop Magician
07:27. OMG! That output looks so much like my work. I must sue open AI!
I asked AI to delete a tree stump that was next to a moose. It did delete the stump, but it put a African wild dog in its place. So now I have a picture of a moose licking an African wild dog.😆
loved the thumbnail, the okay 😒was hilarious
Fantastic, but if you want a real challange try it against someone using a local ai, like stable diffusion with all its built in plug ins. :) (It will be a lot more even, though ps will still win some of the challenges or will be needed for slight improvements)
Hey Umnesh, I think your prompt looked a bit too simple compared to what I usually see. To get better results you gotta give it a scenario, the task, and the goal clearly. I think that would yield a better result and help decide whether we'll all be jobless or not. 😅😅
Please make a thumbnail tutorial also😊
Unmesh : I Let ChatGPT Control My Photoshop Edits!
Chatgpt : Takes notes
All that AI did for me as a graphic designer was flood the vector websites with a shit ton of trash ‘illustrations’ that now I have to sift trough 10 times harder than I did before AI
Sir you're great
Thanks for everything 💞
Love you
Wow, thank you.
I thought AI would eventually takeover.
Thanks for the proof.
large language model means it can generate text, he is misleading and is biased
Can PS do a decent job converting "photo to comic book style art?" That is, with thin black outlines, better if the lines are of variable thicknesses as if hand-drawn wth a pressure sensitive tablet. I get rather spotty results on Leonardo AI, once in a while surprisingly brilliant ones.
Use clip studio paint
@@rohitraoyaduvanshi7330 Thank you. Will try it out there.
Love the thumbnail 🤣 ChatGPT is upset cause you shouted at him 🤣🤣
SIR GIVE THIS ONE PIC IN DESCRIPTION
Wow! That ChatGPT edit is hideous! 😮
yeah because chatgpt is a language model, there are tool that is specifically designed to edit photos and videos and chatgpt is used for text based academic help
Hello Unmesh, you saved my day. Please more of these videos. AI must be unmasked. Lots of people are blinded by it. But AI is and will remain stupid. The term artificial stupidity would be much more accurate. I hope the hype dies down soon and we return to giving the spotlight to creative people. But honestly also includes the fact that Adobe is just as bad in the field of AI. 90% of the results are completely inadequate for me. I don't know how dependent you are on Adobe. But it would be nice to reveal the inadequacies of AI in Photoshop. This company has also allowed itself to be clouded by this hype. It has gone so far that the CEO of Adobe said in an interview that we will save ourselves the photo shoot in the future - make our AI. 😂 I could show dozens of examples of how bad Adobe's AI model is. People, be aware that AI is only a help in small areas. It does not free you to continue working on your skills. AI never beats human creator, not now and in the future, who knows.
Thanks of lot🥰🥰
The one and only Photoshop Gigachad
Why can you edit the original picture in chatgpt? For me it just gives a completely different picture and I'm on the paid version.
And still People say 'AI will take over the world' 😂😂😂😂
photokshop is like a language the more u use the more you know
Again, one of your vids makes me wish that YT had a double-thumbs-up option.
Thanks for the reassurance. AI ain't got nothin' on humans!
so, that's how AI gonna took over huh!
ChatGPT needs to learn Photoshop from you 😂😊
Watching the chatGPT version of that blemishes photo, I was like "Hheheheheheheh..."
Whhooaaaaahhhhh Thank you sensei
2:51
The best part is Chat GPT is fucked by Pixim silently. 🤣🤣🤣🤣🤣🤣
YOU are the master not AI!
So sending your image to India for retouching is still better than ChatGPT.
What if....we start feeding ChatGPT of actual videos of master Photoshop users doing the editing?! Hmm....it can't do it yet because of lack of relevant data... Resistance is futile...
👍
I hate AI out of photoshop in my case the AI gave me another picture 🤦♂️
5 more vids to 1k
Here's a challenge - How to fix deformed Ai hands and fingers in Photoshop.
this is kind of misguiding
chatgpt is not meant for all this. comparing with photoshop and saying you are the control is actually understating its potential. compare with comfyui then especially controlnets using flux models.
This video only shows "Chatgpt" is not capable of editing a photo "yet". And it is not its priority. There are many trained or training AI's out there and they can easily understand what red blemishes are.
Did OpenAI people teach ChatGPT to learn photo editing using Photoshop CS6?
Even if they did, this is still inexcusably bad, lol.
No. I think ChatGPT just sends the photo off to DALL-E to do the actual editing. This is how many specialized ChatGPT add-ons work. They're just front-ends to other AI programs. ChatGPT doesn't look at the finished product once DALL-E has completed it.
Sir hindi me videos banao plzz🙏
First comment brother, please reply 🌺😌
First
AI generation is cool but you have no control over it. Photoshop and other tools let you create something with absolute full control. Like an artist. That is something that can never be replaced.
It’s really a question about weither you want to be the artist or if you want to let AI be the artist. If you want to be the artist and to actually create something, then AI can never give you that. AI is is just an automatic tool that serves you something. You tell it something and it gives you something back. That is not creative or creation. Photoshop on the other hand is a tool to use to be able to realise your artistic ambitions. Like a brush is to the painter.
Chat GPT is just a too much overhyped and overrated tool😂😂
I am unsubscribing your channel, because chat gpt and flux ai will do everything for me....good bye to creators like you.
Chatgpt😂😂😂
❤❤❤🇧🇩🇧🇩
Here comes the goat with another goated video 🐐
Bro just roasted Chatgpt in everyway possible 🥲
I have an image that was initially generated by Midjourney and then upscaled using Topaz Gigapixel. Here are the details of the workflow and the issue I'm encountering:
Original Image Details:
Dimensions: 10752 px x 7168 px
Size: 441.0 MB
Resolution: 300 pixels/inch
Print Size: 91.03 cm x 60.69 cm
I resized the image to 50 cm x 40 cm.
The new pixel dimensions are 5906 px x 4724 px.
The resolution remained at 300 pixels/inch.
The file size changed to 159.6 MB.
Export Settings:
I exported the image as a JPG with the following settings:
Format: JPG
Quality: High
Image Size: 5906 px x 4724 px
Scale: 100%
Resample: Bicubic Automatic
Color Space: Convert to sRGB
Issue After Exporting:
Upon reopening the exported JPG file in Photoshop, the image size changed to:
Width: 208.35 cm
Height: 166.65 cm
Resolution: 72 pixels/inch
Dimensions: 5906 px x 4724 px
File Size: 79.8 MB
My question is :
Why does the print size of the image change to 208.35 cm x 166.65 cm with a resolution of 72 pixels/inch when reopening the file, even though it was set to 50 cm x 40 cm at 300 pixels/inch before export?
Could the upscaling process in Topaz Gigapixel affect the resolution and dimensions in a way that leads to this discrepancy?
How can I ensure that the image maintains the correct print size of 50 cm x 40 cm with a resolution of 300 pixels/inch throughout the entire workflow, including after exporting and reopening the file?