This NEW Leonardo AI Update is a Print-On-Demand Gold Mine (Transparency)
Вставка
- Опубліковано 20 бер 2024
- #leonardoai #leonardoaitutorial #leonardoaiupdate
This new Leonardo AI update is a feature that will be incredibly helpful, especially for creating transparent images.
Try Leonardo AI: app.leonardo.ai/?via=eric-pal...
Now, this isn't the same as the background remover tool in Leonardo AI. This feature's specific purpose is transparency, which will be a goldmine-type feature for print-on-demand creators.
🥇PROMPTS
🎯 Ultimate Leonardo AI Prompts Database (NEW Features): metricsmule.com/leonardo-ai-p...
🤖 Mega Prompts Database: metricsmule.com/buy-chatgpt-p...
🔥 Hundreds of Prompts Here: metricsmule.com/blog/
When you use this new feature, the cool thing is that now all of your images will automatically be downloaded as a PNG file (not JPEG). That way you're endured to receive a generated image with no background (transparency).
This new feature will be perfect for print on demand products like t-shirts or for creating logos.
And one new, unique suggestion Leonardo AI states this is useful for is using this for generating ready-made 2D video game assets.
---------------------------------
Let's Collaborate:
📧 info@metricsmule.com
Schedule a 1-hour Consultation
📆 metricsmule.com/consulting/
Follow me on Twitter/𝕏
𝕏 / metricsmule
🍿 WATCH THESE NEXT
• Why You're Prompting W...
• 12 New Leonardo AI Ele...
• Mind-Blowing Leonardo ...
• This is the Most IMPRE...
🙏 SHOW YOUR SUPPORT
If you find my videos helpful, please consider hitting the THANKS button in any video or buy me a cup of ☕️ with a small tip: metricsmule.ck.page/products/...
What have you been using AI the most for? Let me know and...
🙂Don't for get to subscribe
Disclaimer: Any affiliate programs I participate in are only for ones I highly recommend, use myself, and strongly believe can help others on their AI or digital marketing journeys.
*Some product links are affiliate links which means if you sign-up, I may receive a small commission to buy a 🍪 cookie.
Your all round production keeps improving, I love your delivery, clean, crisp, adult, and to the point, no sarcasm, no ego, just the facts. Well done to you and your team and keep them coming. Much love from Louis in Cape Town SA.
Guy is a total tier legend, it's the little things he says that make huge differences, I remember scratching my head on all the models, he said Albedo, Dynamic pretty much by default, saved me so much wasted time and flip flopping. It's mainly all I use outside of Photoreal, 3D etc. Top guy, top chanel
Thanks so much!
I appreciate that very much, thank you!
This is huge. I've had to use go arounds via photoshop or other websites to create transparent backgrounds after generating images, but now this makes a lot of work more efficient.
Agreed! Honestly other than this new update, the best background remover tool IMO is Canva by far. I always get far better results removing backgrounds in Canva then with Photoshop.
@@metricsmulealso Clipdrop and Starryai have great background removers
This is a great development. I've been using Leonardo AI for just a few weeks and came across the transparency option just yesterday. After creating an image with "minimal background", I went to the menu at the bottom of the generated image (initially to download the image) and noticed I could remove the background. So I tried it, and it worked great. The saving in time editing in Photoshop is definitely worth the value of the extra tokens to do it here.
Your video earned my subscription... and the attention to user experience will soon lead me to trying out the paid plan on Leonardo AI.
Thank you!
I am relatively new to the world of AI tools but I think it is a plus for our creations, thank you very much for your videos, always a pleasure to watch and learn
You're very welcome!
Great vid, thanks. Interesting results by remixing older images and turning transparency on compared to Canvas which was a bit aggressive removing BG. Looks good and at the moment, anything that will reduce the growing pile of apps can't be a bad thing. thanks again
Glad it was helpful!
This update is AWESOME.. I just saw it yesterday when I logged in.
Thank you. I've been getting great results when I use the words: vector, logo, octane render, 3D render, or sticker. It does very well
Thanks info
I like leaonardo upscaling:)
Me too.
Leonardo continues to lead the way in AI generation.
Agreed!
Cool transparency but it's still small right? Does the upscaler work with transparency?
Good question. Yes, it does. What I've been doing is first creating the image and then going directly into universal upscaler and choosing that same image from my recent generations. It works great.
Wow! My jaws just dropped by watching your video! what you just demonstrated is simply amazing! 😍👍 Thank you for sharing this. I didn't know that Leonardo was that good. I usually use Midjourney. But Maybe I will reconsider my mind. 😊 I am glad I have subscribed to your channel.
Thank you! I love Midjourney too, but if I had to pick one, it would be Leonardo AI for sure. The constant updates and improvements did it for me, especially one of their newer updates, Motion.
04:10 am and I ponder what will LeonardoAi give me in exchange for 'Peppa Pig inside of a meat grinder, cartoon' ... What did LeonardoAi give me you ask? ... GOLD
😂 Should I try it?
@@metricsmule don’t lower yourself to my levels man 😂 we need you to be the pro you are to show us all the ways we can create, learn and monise this tech, not mutaliate children’s cartoons like me 😂❤️
Hey there, thank you for the video! I've watched a few of your videos and I'm really liking your content. I subscribed to Leonardo AI and your videos will really come in handy.
I'm loving Leonardo AI, although I still have some issues when it comes to fixing deformed hands. Do you know what would be the best approach for this task or do you have any videos about this topic? Sometimes when it's just a finger that is deformed or so, it's not so hard to fix it with Canvas, but there are many times where images are just close to perfect but hands are a bit more heavily deformed, like for example, there are many fingers severely deformed, or some extra fingers or missing fingers, fingers placed in unrealistic positions when holding certain objects, and so on.
I'm having a hard time finding good info about how to approach these scenarios, and while I certainly made huge improvements since the day I tried Leonardo's free plan for the first time, it's still hard and I usually end up spending a lot of credits. I usually create several images with the same prompt until I get a couple in which hands are ok, but I think this approach is a perfect way to destroy my credits, not very efficient. What I found so far is that there is no tool that consistently performs well in this task, and that which tool will shield the best results will vary all the time depending on the particular image / deformity / problem. A missing finger in one image might be easily brought back with the Inpaint tool, while a missing finger in other image (or another missing finger in the same image) might require the use of the Sketch2Img mode. So far these are me findings.
May you please give me some insights on this matter? What would be the best tool / canvas mode to fix hands in tough scenarios? Or is there no specific way and it depends very much on the situation, just like I found? When a hand is severely deformed, maybe would it be better to use the erase feature and just generate a new hand instead of trying to fix it finger by finger or by small portions?
Once again, thank you for the videos. You have a new subscriber. Have a great day!
Yeah, I will too every now and then but it has gotten much better. But yes, I do have a video to help with this. I created a negative prompt generator you can use that will provide you with negative keywords specific to the type of design you are creating. Once I get the negative keywords, I'll then plug them into the negative keyword section in Leonardo AI and it usually solves the problem. Here's the video: ua-cam.com/video/7DwS63S-Qxk/v-deo.htmlsi=KuQlgZlKzUoamapG
@@metricsmule Thank you for your answer. I just watched the video. My only concern is having to disable PhotoReal, but considering what you say, is it then better to disable it in order to have the ability to add a negative prompt and this would yield better results? I'm certainly going to try this out. Thank you very much.
For sure. I disable PhotoReal all the time, especially when using AlbedoBase XL and Leonardo Kino XL. Try those two models with PhotoReal off, then turn on the negative prompt box and add your keywords. This method works pretty well. Let me know if it worked for you.
@@metricsmule hey there. I've been trying this method, still a lot of issues with hands, it might have helped a bit but I don't think it made a big difference, at least in the type of images that I'm doing. Maybe it works better for images in which hands are not so important I think. However, it was useful to find out, thanks to this method of yours, that PR is not really necessary to achieve amazing results, I'm saving some credits plus I can use a negative prompt. Thank you!
In case you're interested to know and try it, I might have came to a potentially good method to fix hands more easily. Now what I'm doing is the following: When I generate the image I want and it comes out with bad hands, first, if hands are really, really messy, I take the image into image to image and proceed to generate similar images, I generally leave the strength slider at 0.3 for this purpose, or maybe raise it just a little bit. Sometimes I do it one time with just one model, sometimes I do it several times using different models. It doesn't take long until I get a very similar image with much better hands. However, I think this step can be skipped, the important one is the following.
After that, I take the image I like the most and take it into the Universal Upscaler. I proceed to upscale it leaving the creativity slider on 5 or 4 (I don't think this is relevant for the purpose of fixing hands though). Once I do this, I take the upscaled image into the Canvas Editor and fix the hands. The thing is: The Canvas Editor seems to be way more accurate with images that have been upscaled through the Universal Upscaler. I guess it might be that they are interpreted by the Canvas Editor in a much more accurate way due to having way more detail, idk. But doing this, the Inpaint / Outpaint feature seems to make waaaay less mistakes, and I usually end up with a good result in one or two tries for each section of the image I want to fix, while it usually took me several tries before. And I'm using this method to fix images of groups of people working in the office, so... dealing with lots of hands atm.
Keep in mind that this is something that I'm still testing so I might be wrong, but so far, it seems to be working really well for my images.
Hi Mate. Why the last prompts you put in the database don't have fine-tuned model. Is an error from the database? Thanks
I'm in the process of doing a big update today and adding a lot more prompts so there may be a few that temporarily don't have the models or presets assigned yet. They'll be in there today though.
@@metricsmuleThanks mate. No worries. Still doing the great job you do
Your content is always OFF THE CHAIN!1🙏🧞♀️🙏
Thanks so much
Do we need to upscale these to use in POD.. and if so does it maintain the Transparency to export as a PNG.. I haven't tested that yet.. but I am sure others have that question
Good question. You can actually upscale the upscaled version which produces amazing POD results.
The file format to png is because jpeg doesn't support transparency. I always use png anyway so it kills 2 birds with 1 stone for me.
Yes, good point. The only time I really recommend jpeg is when you're uploading an image on a website and then compressing them.
Ah how great it would have been if only Leonardo was making good images. I just hope that now Midjourney does the same.
PS I have soo many credits for Leonardo which I don't use that I decided to give it a try, as usual to have a decent result, I had to give an image reference and .... the image was better, but the transparent background does NOT work anymore, it came on white background.... how frustrating is that?!
If that ever happens to me, then I'll also just use their background remover tool as a compliment and it usually does a great job. What kind of images are you not getting the results you're looking for? Maybe there are some unique keywords I can provide you with that may help.
@@metricsmuleThank you for the advice, I was trying with illustrations clipart type images, watercolor style, air-brush style. :)
Can you check your mails?
Will do.
not consistent maybe 4-5 years later.