I think ..without adjustments and settings ...MJ is a force... but SDXL has tons of fine grained controls. Thanks for sharing your viewpoint. I highly appreciate it.
I agreed with their ranks on everything except maybe the fashion prompt as I slightly leant towards the SDXL result. Which result didn’t you agree with?
@@lamsmiley1944 To be onest, MidJourney due to it's training data set is the winner for me also. But the scoring pictures the other far worse and I don't agree with that.
@@TrimTask I immediately recognized that ai voice when I watched your video. Recently, I stopped using Elevenlabs and switched to imyfone Voxbox. I get great voices without the generation errors of Elevenlabs. While Voxbox has a lifetime deal, I am not thrilled with their understanding of lifetime. Lifetime to them meant 2,000,000 characters and voice cloning seems to be extra; I am still waiting for an answer from them. Regardless, I think for the quality and cost, Voxbox blows the pants off of Elevenlabs.
SD is definitely getting closer to MJ lately but I overall agree with the conclusion here. I even liked MJ's portrait better than the SDXL one to be really honest. The speed at which MJ is improving is also incredibly impressive. They sure seem to be hard to beat. Not unsubscribing any time soon!
The comparison is absurd. You generate an image of 1024 pixels with a model trained on 512 pixels in Leonardo. Of course, you have your subject duplicated... Furthermore, you tested only one model, and not the best one on Leonardo. As for the prompts, the syntax is different and you use the MJ syntax on all prompts. In short, a comparison that makes no sense. You may know MJ, but not SD, that's obvious. Inpainting in MJ ? No ? Upscaling by X4 or X8 with more details ? No. Controlling the composition (controlnet) ? No. Using LORa ? No. With SD, and lot of work, you could achieve amazing results. The images produced by MJ are mostly aesthetic, but there are many subjects and compositions that are impossible to achieve, except by hoping that by continually generating images in Discord, we will eventually get what we want. You compare apples with pears in five minutes. I'm using both (MJ and SD with WebUI), for different purposes : you really need to learn how SD works. Install Automatic1111, do your homework, and come back in a few months.
If you carefully watched, there was clarification given that we are avoiding any settings.. additionally, whatever you mentioned requires expertise and gpu based systems My target audience is not expert users like you... but simple casual users who use image generation for regular basic usage.. ... prompts were not specific to anything. I am no expert of MJ, Leonardo or SD... but a simple casual comparison keeping simplicity of usage in mind. Surely there is no comparison ... but I did it anyway because that's how a common ..casual user looks at it... anyways thanks for watching and feedback... I use SD ..Leonardo regularly for different purposes too.
![30 ชั่วโมงคนเดียว บนรถไฟชั้น 3 โซเวียต [Around The World Ep. 2]](http://i.ytimg.com/vi/MlE9boAVqDA/mqdefault.jpg)
Woah, I'd say it's totally biased. But. The comparison and the making is marvellous.
I think ..without adjustments and settings ...MJ is a force... but SDXL has tons of fine grained controls. Thanks for sharing your viewpoint. I highly appreciate it.
I agreed with their ranks on everything except maybe the fashion prompt as I slightly leant towards the SDXL result. Which result didn’t you agree with?
@@lamsmiley1944 Yeah the more your looked at it the more you liked it right?
@@lamsmiley1944 To be onest, MidJourney due to it's training data set is the winner for me also. But the scoring pictures the other far worse and I don't agree with that.
I'd be interested in a video about how you do your videos. The AI narration was super impressive!
Thanks! That you liked the narration! I used following method : th-cam.com/video/mBtS2Rmhr6k/w-d-xo.html
@@TrimTask I immediately recognized that ai voice when I watched your video. Recently, I stopped using Elevenlabs and switched to imyfone Voxbox. I get great voices without the generation errors of Elevenlabs. While Voxbox has a lifetime deal, I am not thrilled with their understanding of lifetime. Lifetime to them meant 2,000,000 characters and voice cloning seems to be extra; I am still waiting for an answer from them. Regardless, I think for the quality and cost, Voxbox blows the pants off of Elevenlabs.
SD is definitely getting closer to MJ lately but I overall agree with the conclusion here. I even liked MJ's portrait better than the SDXL one to be really honest. The speed at which MJ is improving is also incredibly impressive. They sure seem to be hard to beat. Not unsubscribing any time soon!
Totally with you here .
I think for the ramen one, SDXL is a clear winner, the MJ one looks like a desert somehow.
Thanks for sharing your observation and opinion!
The comparison is absurd. You generate an image of 1024 pixels with a model trained on 512 pixels in Leonardo. Of course, you have your subject duplicated... Furthermore, you tested only one model, and not the best one on Leonardo. As for the prompts, the syntax is different and you use the MJ syntax on all prompts. In short, a comparison that makes no sense. You may know MJ, but not SD, that's obvious.
Inpainting in MJ ? No ? Upscaling by X4 or X8 with more details ? No. Controlling the composition (controlnet) ? No. Using LORa ? No.
With SD, and lot of work, you could achieve amazing results. The images produced by MJ are mostly aesthetic, but there are many subjects and compositions that are impossible to achieve, except by hoping that by continually generating images in Discord, we will eventually get what we want.
You compare apples with pears in five minutes.
I'm using both (MJ and SD with WebUI), for different purposes : you really need to learn how SD works.
Install Automatic1111, do your homework, and come back in a few months.
If you carefully watched, there was clarification given that we are avoiding any settings.. additionally, whatever you mentioned requires expertise and gpu based systems
My target audience is not expert users like you... but simple casual users who use image generation for regular basic usage.. ... prompts were not specific to anything. I am no expert of MJ, Leonardo or SD... but a simple casual comparison keeping simplicity of usage in mind. Surely there is no comparison ... but I did it anyway because that's how a common ..casual user looks at it... anyways thanks for watching and feedback... I use SD ..Leonardo regularly for different purposes too.