top of page
Writer's pictureATAO

Solution thinking AI mindset

- midjourney, how do you create such beautiful pictures?

- I steal ideas from people's endless photo libraries, collage it and recreate like mine


But the problem is that midjourney does not always follow the task from the client, because he is engaged in creativity, and collages from what he has. This is similar to the work of studios with a focus on art, such as mir.no or kvant1.no . As for the rest 99% of studios that has a focus on a commercial product, service, the organization of the process there, looks like this:



1. Coordination of camera angles and the building model/master plan itself on gray material

2. First round of drafts without retouching

3. The second round of drafts, with retouching (replacing greenery, sky and adding people, smth one or all three)

4. The final render in 4k or additional rounds of edits


In more creative studios, such as mir.no or kvant1.no, the workflow is different:



1. Communication with the client and architects about the concept idea, sending a brief to a client

2. Matching camera angles by drawings and sketches inside the studio

3. Completion of the selected sketches for an interim presentation to the client and another round of discussion of the client's/architect's ideas

4. The final work of the artist, 3d, collage, drawing up a task for the photographer, inserting 3d into the photo…

5. Presentation of the final image to the client. After which the client accepts the job and pays or does not accept and the working relationship is over.



The archvis market is not oversaturated with work, such as the IT sphere, so the correct algorithm for working on a picture is still in the stage of improvement and also differs not only from the specifics of the studio, which is emphasized on art or commercial pictures, but also there is a difference in the composition of the team that works on the project. In some studios, modeling, creating the image itself in 3d, and retouching are three different processes performed by 3 different people, while in others it's all the task of one artist.

But there is an answer to what the right workflow should be. IT which is oversaturated with work and money tells us about it. The essence of working on a project in IT is to minimize all the risks by reducing work on a project to a series of short and fast cycles, short-distance races, where you do not do a whole stage in 1 day, but do a pre-stage in 0.5 hours and, after agreeing with the client, you calmly finalize it, without the risk of doing wrong.


In other words, the stage of matching camera angles on the grey material is something that cannot be in the correct process of working on the picture. Since it is impossible to adjust the light and angle without an approximate understanding of the color. If you have a red bike path, and you need to show a green entrance group, it is obvious that the camera should not capture the bike path, or if you have a black façade or opposite, the chrome metal panels facade, you will not be able to set up a beautiful light for it on a gray material. Therefore, it is correct to coordinate camera angles according to a series of fast drafts. Fast draft is a draft that lacks all the final content, people, cars, and so on... The task of the fast drafts is to choose the camera angle from a picture with more or less customized building materials and lighting.

It's quite another matter if there is a desire to put pressure on the client with the professionalism of the studio and there is no desire to show bad-looking fast drafts for the choice of angle. But in this case, we harm ourselves by violating the principle of short fast cycles and increasing to ourselves the risks of agreeing an unsuccessful angle on the grey material. If we compare our grey material angles and light on it with our final renders we will se, that they a different, and to make this little difference artists spend a lot of time, because the camera angle is agreed without materials and color, and all the problems appear, when this color was added… So, it's not easy to correct something, when you already agreed this with client and art-director. So, the right workflow for a studio that provides not art, but service, commercial pictures, looks like this:


ATAO workflow algorithm

1. Communication with the client and architects about the concept idea, choosing the references for each picture.

2. Coordination of camera angles and the building model/master plan itself on fast drafts with basic textures, and final light settings but without final filling of the scene

3. First round of drafts with retouching

4. Second round of drafts with retouching

5. The final render in 4k or extra rounds of edits


Do not neglect the choosing of references, because the creators of the algorithm for midjourney bet on it and did not lose. The correct reference is the key to the success of a beautiful picture.

An important point is that if in the process the client agreed on a reference where the morning with long shadows is shown, but then asked just to raise the sun to make the shadows shorter, you need to go back to the first stage and agreed the reference again, let it be completely different, but in the end, the raising of the sun on your picture will also make it completely different.

bottom of page