The easiest way to begin a commissioned artwork is to ask what style the client wants.
I think that is usually too early.
For Twain Editions, the more useful first question is:
What is the story?
A style can make an image attractive.
A story gives the image something to hold.
That distinction became the foundation of Portrait Dossier.
A commission is not a filter
Twain Editions creates premium commissioned digital narrative art.
The word “narrative” is important.
The goal is not to take a photograph and apply a recognizable treatment to it. The goal is to build a composition around a person, relationship, memory, place or period in a way that feels specific to that story.
That means references are not just likeness references.
A wedding date can matter.
A building can matter.
The geography of a place can matter.
A small object can matter because somebody remembers it.
The order of those details matters too.
The artwork has to decide what is central, what is supporting evidence and what should remain absent.
That is composition before rendering.
Start light
The early stage of a piece should be cheap to change.
That sounds like a production principle, but it is also an artistic one.
If the first concept already looks finished, there is a strong temptation to protect it.
So I prefer beginning with extremely light, non-photo blue-pencil studies.
At that stage the drawing is allowed to be incomplete.
The purpose is to resolve structure:
Where is the emotional centre?
Which figure carries the story?
What belongs in the foreground?
What should sit in memory?
How does the eye move?
Where should the composition breathe?
That stage is closer to storyboarding than illustration.
And that is deliberate.
Resolve one problem at a time
Generative tools make it easy to request a complete image too early.
The result can look impressive while still being wrong in several important ways.
A stronger process isolates decisions.
First the composition.
Then the relationship between major forms.
Then likeness.
Then architectural or documentary references.
Then tonal hierarchy.
Then rendering.
The exact sequence can change from piece to piece, but the principle stays the same: avoid asking one image to solve every problem at once.
This is slower than generating endless finished alternatives.
It is also faster than repeatedly repairing finished images whose foundations were never resolved.
Likeness is not the whole portrait
When a piece includes real people, likeness matters.
But likeness alone does not create emotional accuracy.
A face can be recognizable while the image still feels wrong.
The posture may be wrong.
The age relationship may be wrong.
The distance between two people may be wrong.
The environment may be carrying too much weight.
The composition may be describing the event without describing the memory.
That is why I treat likeness as one layer inside the narrative rather than the narrative itself.
The artwork has to feel like the person belongs inside the story.
Style comes downstream
Once the narrative architecture works, style becomes useful.
Graphite can create intimacy.
Ink can create urgency.
A restrained limited palette can separate memory from present time.
A polished digital finish can be useful for one subject and completely wrong for another.
The style should serve what the image is trying to say.
Not the other way around.
That is also why Twain Editions should not become visually uniform for the sake of brand consistency.
The brand can have standards without forcing every client story through the same aesthetic template.
The work should have family resemblance.
It should not have a house filter.
The product is the process as much as the output
Portrait Dossier is sold as a finished digital artwork.
But the quality of that artwork depends on the process around it.
Collecting the right references.
Understanding what is emotionally important.
Separating fact from embellishment.
Building the composition slowly.
Knowing when a piece is finished.
And, just as importantly, knowing when to stop “improving” something that already works.
That last point is difficult with generative tools because another version is always available.
The discipline is recognizing when iteration has stopped solving problems and started manufacturing new ones.
Story before style
The principle sounds simple because it is.
Do not begin with the effect.
Begin with the reason the image should exist.
Understand the people.
Understand the memory.
Understand what the client is actually trying to preserve.
Then compose.
Then draw.
Then render.
Style is powerful.
But it should arrive after meaning.
