Variations with noise injection KSampler (in pills) 

Okay this an experiment I'm doing with a new kind of video format. I try to pack my videos with a lot of information and sometimes it might feel overwhelming. I was thinking maybe you'd also appreciate shorter videos dedicated to just one concept or even just one node. Please let me know what you think and if you'd like more of this kind of videos. Ciao!

probably one of the best guys doing tutorials about comfyUI

Long or short, I’ll still watch your videos start to finish at least twice (usually more). Thanks for making such high quality guides. You’re the best out there.

this is cool, but the more in depth are cool too, please don't stop them....

The advantage of these is not having to scroll and remember where was that small thing I faintly remember from what video, thank you!

Just be flexible with your video lengths. If you think a subject needs more explanation than take your time because it's greatly appreciated. I find your videos to be priceless and easily digestible :)

Even your short videos are extremely impactful! 🥰 Another tutorialist did a video where he just let us watch his process in understanding the lora stacker node when it first came out and there was no information on it. It was absolutely wonderful to sit on the shoulder of someone with more knowledge and watch how they tackle the unknown. It helped me a lot in understanding the underling structures, just like all of your videos do! Thank you to everyone spreading the Comfy love!

short or long videos from you always bring value to the community, looking forward to next update. thanks for sharing your insight.

I like it a lot, I am also guilty of not always completing a video before I start trying out what you are teaching.

Please keep explaining how it works under the hood! It is what makes you special!

Hi Matteo..thankful for your presence.We need your knowledge and ideas,every single piece of experience.Waiting for more

Honestly if it's valuable it should last for hours for the deep insight. I don't mind shorter pills as well! Good job!

I am looking forward to the next pill, and all other episodes. Great work.....

personally enjoy all the formats of your videos. A series of focus on each node and how it fits in with others is greatly welcomed! Especially enjoy how you explain the settings and what they do and impact things further down the noodle. Thank you for your time creating these useful extensions and explaining them!

This is great! Very Midjourney like where you can select an image you like and then iterate on it.

I like both the short and long form

Wow I learn a new thing with every video I watch. Thank you for ComfiUI.

I love following along with the longer videos! They teach so much

When you show.... You have my full attention, Maestro Latente, no matter the length!!

this the kind of content we want .... keep it going.

Yet another great tutorial, very nice explanation of the process. Thank you very much!

I really enjoy your more in depth videos but the shorter format, limited topic video also works well - I think a mixture would work well, depending on the subject matter and how much time you have available!

This small pill format is a fabulous way to create small learning chunks similar to advance chess games. Being able to learn the basics while learning small advance nodes it is fast way to learn a subject. Thank you so much again for this video.

This is honestly a great concept. Somewhere between like a short and a 20 minute video on a comfy ui concept. I could have used some 3-5 minute video on any of the following and it would have made my life so much easier from the start: Templates, clicking github links on the custom node manager to learn directly about the nodes, changing base directories, how to custom save files using the "Save image extended" node and the y/m/d folder format, how "uploading" input images actually works and how to not constantly upload the same image multiple times, how guide size functions in detailer nodes, lite normal and plus ipadapters functionality, interpolation to cut down on animatediff/video frame processing for an even smoother look, etc, etc.

Thanks a lot for decode the process, Nice video

personally don't care about the length as long as it is useful, and yours is, and you have a cute accent, don't worry about it

Awesome and also many thanks for the mini tutorial!

Great short video! Thank you for your job!

Definitely not leaving because of your voice ! lol. The videos are dense and every sentence contains important information. So they need to be re-watched, slowly and thoroughly. Thank you for doing these. They are unmatched. I hope you do a series (or even a paid course ) on training a Dreambooth or LORA. Cheers, b

Fantastic guide, as always. Long or short, I'll watch your videos in full. Often more than once. But pills are great.

Wow sooooo simple, easy to follow in setting up one node concept workflows. Often I find myself going back to basic and efficient workflows, thanks for the work you share with community Mateo! Building grass roots knowledge!!!

I watched to the end on all your videos 😊

both, both are absolute bangers filled with knowledge

I for one like your long-form videos

yes, focusing on just one concept is great. thank you

Fantastic video! The advanced sampler was something I have seen plenty of times but never used before. I appreciate the short exposition of how to use it. Long or short, your videos are explanatory powerhouses. >:D

You do the best videos, so much knowledge, and your voice is great. Keep it up!

These technical, under the hood approach really beautiful. Many thanks 🙏🏻

Thank you for always great lectures!😍😍😍😍😍😍😍😍😍😍

to be honest you did great tutorials, don't change anything. just keep it up 15-25 minutes is perfect. ive seen all your videos from start to end, top content

No matter your video, I always end up watching twice! Once to get the idea and once again to follow along. Long form or short form, I iwill always apprecatie your insights! You ae a treause!

Amazing node, tutorial and format! thank you Matheo!

amazing! love the format!

Super dope node, Matteo! Love the new format too, although I love all of you videos)

Great format of videos!

Thank you for these, all lengths are great. I personally really enjoy both and just pick the ones that make sense for the time I have

yes, it's great. thank you

Thanks a lot. Thanks to you, I migrated most of my A1111 workflows. Keep them coming.

Always great videos. I like the new format

Very cool format, looking forward to more latent pills 💊

If I had to choose between long and short format, I would choose long, but this "pill" format is a nice addition for specific tasks or nodes. Anyway, your content is still the best!

The lenght doesn’t matter, I’m always looking forward for a new tutorial from you! This short one is nice for small stuff like “hat have you tried this funny feature?” But the long ones stuffed with all bunch of info helped me the most at my beginnings with understanding Comfy in general ;)

Long or short they are both very insightful, id take either of them. Just do whats easier for you as we know making these videos are not easy and very time consuming. Thanks!

for me it's okay long or short videos, if I have to watch a 15-minute video, I find the time to do it, but it's obvious that 5 minutes can always be found, and that's the pro of short videos. But I also like the technicalities, understanding how things work, but I understand that in a 5-minute video you can't have a full explanation, so that's the cons. However, it is always a pleasure to see your videos.

While you could have just explained your node, I really appreciate the depth you went into in explaining the inner workings of how and why it works. Very smart to take a queue from your analytics. While I personally would watch you do this for hours in one video, I have no issue watching shorter videos that focus on one area at a time too ❤

This is a great node, it eases the process for image-latency to image, thank you a lot for assembling all this 😀

Long format or short format, these are the best comfy videos. Maybe longer when doing a project, like the card creation, and shorter for singular node features like this?

What a wonderful node and tutorial, thanks!

I love your long videos matteo. Im just eagerly waiting for you to release a video and I want it to be a long one if possible

I really love this new format. I also really like that you are covering basic and advanced topics. Many other channels just talk about a topic, but don't actually explain how it works and (more importantly) how to use it.

Love youre vids Matteo....both formats are perfect 👌🏻 👌🏻

Do it the way that is most convenient for you. Your knowledge needs to be seen many times This is because viewers watch slowly, multiple times if necessary.

Short and sweet, thanks for this!

I learn so much from you. Thank you mate 🙏

you are amazing!...some people don't understand how precious is the things you are doing for us, the community...Thank you so much!

outstanding as always!

Great info as usual keep up the hard love your expertise and willingness to share!

Suffering from character consistency, this is a great pill

Great job 👏🏾

Great video, thank you!

Super interesting 🙏🏽

Those "pills" are a good concept, but I also really like the longer in-depth tutorials, becaus e they help me to better understand the underlying priciples... but maybe it's just me...

The best tutorial ever for me❤❤❤

Thank you, sir.

Great stuff! Thanks you and keep up the good work!

Love it, thx

thank you for your wonderful work ! one thing that i miss the most with comfyUI (and i don't know if it's available out there) is good shortcuts. I come from blender that i use with the node wrangler addon, and i've gotten so used to the useful shortcuts it can sometimes be frustrating to use comfyUI

Your analytics might say that but you have an audience of people like me who will watch the entirety of your videos no matter how long ! We might not be the majority but we will always be here because you make the most useful content and therefore the greatest content. If you believe it's your voice the problem, i'm pretty sure you know how simple it is to train an AI voice for live modification ;) keep up the good work we believe in you !!!!!

Outstanding!

amazing 🤩

Shorter is good, but there should be more shorter videos! Love your channel

Really good thanks!! More Zero To Hero please!

I love it again thanks!

MATEO YOURE THE BEST

I love the video, but I also could watch an hour long video if you released it! ❤

You are LEADING the pack - Job well done - Now the Joke part - I only watched the first third of the video, the noise injection steps seemed obvious (lol - just kidding!)

Very good thank you

Please longer videos with more explanation

The ones that can't listen for longer than 5 mins can't take the heat! Bring on the 60 min in-depth tutorial! 🤣😆

Nice video, quick and easy

Hey Matteo, I appreciate your videos, they're the only stable diffusion related things I watch. I think shorter videos is probably a good plan, even if you have more to cover than you can manage in one short video, it might make sense to break things up. I know I can get distracted, sometimes I have to pause, etc. Might help the metrics to do it that way. Also, regarding your essentials node pack... As it grows, it might be a good idea to break that up as well. I think a lot of node pack authors are deep into working with ComfyUI, but as a more humble user it can be overwhelming to install a pack and end up with a jumble of nodes you have no idea about, especially if you end up installing node multiple node packs. It makes a mess real quick. I don't know if that's something that applies to your collection, but I think it's worth considering.

I'd to see more work on producing body consistency

I'd rather have a video from you than not, so I'm 50/50 on the new format. I'm prone to rewatch a video, incase I missed something the first time, or if I can refine a workflow for a discreet task in Comfy.

all your format are belong to us

Your GOAT

Nice! Is it possible to mask where the noise is applied? I played with the Blenderneko noise injection nodes to try and get different noise for upscaling the face vs the background, so I could control how much it changed while keeping the face similarity, but as you noted, extremely complex to setup and follow.

Ciao Matteo e grazie per il tuo lavoro è straordinario… ad ogni modo vorrei porti una domanda alla luce dei nuovi aggiornamenti, in sostanza attraverso touchdesigner vorrei fare un noise audio reactive e darlo frame per frame in pasto a comfyui, secondo quale protrerebbe essere il settaggio delle parti al fine di avere ovviamente una continuità nella consistenza dei risultati ? Grazie in anticipo

I'm trying to get your IP Adapters to create a three unique face composition but not having success. I tried to use the New Attention Mask video and workflow but it doesn't preserve the face very well. I suppose FaceID could be used to influence the image. Can you create a tutorial for showing how this could be achieved to have 2 or more different faces in the same generated image. I was wanting to use this kind of workflow for Mother's day, maybe other people benefit from it too.

Very useful, I have been using a similar method in which I add noise to the out put image directly by blending the output image with a noise image or generated noise.. I assume you could use a mask to only generate variations in the figure or background?

👌👌👌