The way I have been sculpting goes pretty much like this:
-
Start sculpting on a simple shape. Sometimes I start with Voxel and just use Elastic Deform to get a simple base shape for the head. That bird model I made weeks ago I used that method for the beak, but after that I went straight to Dyntopo.
-
Once I begin using Dyntopo, I use Constant Detail. I usually set it to a medium amount of detail and just keep building the rest of the body with the Inflate brush and Accumulate on. I don’t want to go overboard with the details too much or it will become harder to control the shapes. I only add more details to areas that need smaller details like the face, hands, and feet.
-
Once I am happy with the overall shapes, I begin adding more detail to the body with Dyntopo so it is a bit more consistent across the board. That way I can add some extra free details before activating Voxel and avoid getting too jagged shapes that will be retained using a high density Voxel remesh.
-
When I activate Voxel, I don’t use Adaptivity, but just remesh the sculpt at a high enough resolution that my shapes don’t get lost. Usually this will make your sculpt have a couple of million verts, but it could be higher depending on the complexity of your model.
-
I start refining the details and shapes by checking all the anatomy to find any major faults. Because of the mesh density I can sculpt without needing to remesh over and over again. However, I will do one final remesh once I am happy enough with this stage of the sculpt so I can start adding textures.
-
I increase the Voxel Remesher density some more. This will add some additional million vertices to the sculpt, but this will allow you to finally use texture brushes. This way you can get a lot of free texturing and avoid issues like the extremely slow Smooth brush with MultiRes down the line.
-
After I am happy with the anatomy and texture work, I at last start doing manual retopology, or simply use QuadRemesher/QuadriFlow depending on the need. I then add MultiRes to the new mesh and Shrinkwrap on top of that after I subdivided to the point where the mesh has reached a similar amount of verts like the high resolution Voxel sculpt had. I change Shrinkwrap settings to Project with Positive and Negative checked on to get the best results.
-
Finally, I just keep working with MultiRes to add textures and doing a few fine adjustments. Nothing too crazy, since you know how unstable it is with Apply Base and lower subdivs.
Here’s an example of a sculpt I made using these steps. The images in the OP are primarily the Dyntopo stage, then you get to see the Voxel stage. I’m currently in the MultiRes stage with ideas of redoing the base mesh so I can get better UVs for texture painting later.