Summing up: how our tools compare for cost, speed, ease of use and quality

This page has 0 threads | Add post

Blogs

Finding our blogs useful? Please consider helping us to share more!

Comparing the 5 leading AI tools for image generaton from text Part seven of an eight-part series of blogs
Having recently compared the 4 main AI tools for text prompts, I thought I'd do the same for image generation tools. In this blog series we compare Dall-E (via ChatGPT and separately via Copilot), Firefly, Midjourney and Stable Diffusion for 3 pre-defined tests to see which ones score most highly for cost, ease-of-use, speed, editing ability and above all for quality of image. A deep-dive review of text-to-image AI tools The 5 AI image generation tools that we will compare The 3 pre-defined tests that we will run for each tool Test 1 - photo-realism (a picture of mice being judged) Test 2 - following instructions (a treasure island map outline) Test 3 - being creative (a toy sheep as superhero) Summing up: how our tools compare for cost, speed, ease of use and quality (this blog) Conclusions and recommendations

Comparing the 5 leading AI tools for image generaton from text

Part seven of an eight-part series of blogs

Having recently compared the 4 main AI tools for text prompts, I thought I'd do the same for image generation tools. In this blog series we compare Dall-E (via ChatGPT and separately via Copilot), Firefly, Midjourney and Stable Diffusion for 3 pre-defined tests to see which ones score most highly for cost, ease-of-use, speed, editing ability and above all for quality of image.

A deep-dive review of text-to-image AI tools
The 5 AI image generation tools that we will compare
The 3 pre-defined tests that we will run for each tool
Test 1 - photo-realism (a picture of mice being judged)
Test 2 - following instructions (a treasure island map outline)
Test 3 - being creative (a toy sheep as superhero)
Summing up: how our tools compare for cost, speed, ease of use and quality (this blog)
Conclusions and recommendations

Posted by Andy Brown on 22 July 2024

In this blog

Speed
Ease of use
Cost
Editing ability
Quality of results

Having done the 3 tests I'd prescribed, let's compare the 5 AI tools being reviewed for speed, ease of use, editing ability, cost and quality of results.

I'll just repeat at this point: I'm looking for a tool which can generate images for blogs and newsletters by following my instructions, rather than magnificent artwork with a zillion-pixel resolution to enter into competitions.

Speed

For me speed isn't a big issue - I can just do something else while I'm waiting - but for the record here is how long each test took to complete in seconds:

AI tool	Test 1	Test 2	Test 3
ChatGPT	19	27	19
Copilot	24	22	27
Firefly	17	14	N/A
Midjourney	40	58	41
Stable Diffusion	21	19	15

So there's no clear winner here, but there is a clear loser: you'll typically wait twice as long for Midjourney to produce any images as for any of the other tools.

Ease of use

Once I'd got them set up OK, 3 of these tools were straightforward to use: you just pasted in the prompt and said go. ChatGPT and Copilot were probably the easiest tools to use as they made uploading images so straightforward.

However there was one very clear loser! Step forward ... Midjourney:

See my full list of whinges about Midjourney below!

I gather there is a web-based version of Midjourney in the wings, but a) I can only judge what's before me and b) I don't know if this will solve all of the problems I've listed below.

Here are some of the things which make Midjourney harder to use than the other tools:

You have to sign up to Discord, a discussion app, then join the Midjourney group, then find a newbies channel to join.
You have to type /imagine before every prompt. This is annoyingly unnecessary, especially as I keep forgetting to do this and wondering why no image is appearing.
You have no privacy. While you're waiting for your image to appear, you can also see the exact text of what other people are typing in, and the images they are creating (and by the same token, they can see yours).
Adding images (as we did for the third test) is a pain: you have to upload the image, then copy the link to this and paste this into your prompt.

If you already use Discord for other work, maybe you'll find all this intuitive (I do, but still don't).

Cost

You'll need to pay for ChatGPT (probably) and Midjourney (definitely) before you start using them, but you'll probably end up having to spend money whichever tool you go for. Here are the details:

AI tool	Pricing	Link
ChatGPT	You may be able to get away with free use of ChatGPT 4 for a bit, but this depends on the whims of OpenAI - it seems safest to assume you'll have to pay $20 a month from the outset.	See pricing plans
Copilot	Copilot offers a one-month free trial, but thereafter you'll have to pay £19 per month.	See pricing plans
Firefly	You can get 25 monthly generative credits for free; thereafter you'll have to pay £4.49 per month. Each time you generate an image from text you use up one generative credit.	See pricing plans
Midjourney	$12 a month (it's a bit cheaper if you pay annually)	See pricing plans
Stable Diffusion	$27 a month, but you can get a 3-day free trial.	See pricing plans

Editing ability

I only really compared this for the first test, and I don't think any of the tools shone! Then again, perhaps the fault lay with the human in this case, so I think my conclusions should be taken with a pinch of salt.

Quality of results

Just to reiterate: what I mean by this is how closely each AI tool was able to produce what I asked for.

Copilot won every test, with ChatGPT coming a fairly close second in each case. They were the only tools which seemed able to do what was asked of them.
Stable Diffusion was shockingly bad; the images it generated seemed to bear little relation to the prompts I gave.

However, what was really bad about Stable Diffusion was its 500-character prompt limit (and seeming intolerance of some characters like apostrophes):

This got wearisome!

Please do get in touch if you work for Stable Diffusion and feel your employer has been unjustly treated. I've tried throughout this (long) blog to be as fair as I can be, and would hate some prejudice or ignorance on my part to be distorting the results.

So given all of the above, let's finally look at my concluson and recommendations!

Parts of this blog
A deep-dive review of text-to-image AI tools The 5 AI image generation tools that we will compare The 3 pre-defined tests that we will run for each tool Test 1 - photo-realism (a picture of mice being judged) Test 2 - following instructions (a treasure island map outline) Test 3 - being creative (a toy sheep as superhero) Summing up: how our tools compare for cost, speed, ease of use and quality (this blog) Conclusions and recommendations

Parts of this blog

A deep-dive review of text-to-image AI tools
The 5 AI image generation tools that we will compare
The 3 pre-defined tests that we will run for each tool
Test 1 - photo-realism (a picture of mice being judged)
Test 2 - following instructions (a treasure island map outline)
Test 3 - being creative (a toy sheep as superhero)
Summing up: how our tools compare for cost, speed, ease of use and quality (this blog)
Conclusions and recommendations

Some other pages relevant to the above blogs include:

Courses in Artificial Intelligence

This blog has 0 threads

Add a new post

BLOGS BY TOPIC

BLOGS BY AUTHOR

Speed

Ease of use

Cost

Editing ability

Quality of results

Head office

London

Manchester