Simplification and Rendering Organic Objects
The most obvious thing about manga is that it's basically simplified realism. Sure we play with proportions for style's sake, but the end result still reflects a good amount of realistic influence. But what people seem to have trouble doing is translating their knowledge of realism into their manga drawings. I see a lot of artists who do realistic anatomy studies, but the studies don't really show in their work because they have trouble integrating the know-how.
That's why I want to demonstrate simplification of form.
This is a whole spectrum of simplification up to your typical manga level of abstraction. We start with anatomical realism and start grouping together the forms until we narrow it down to an abstract of the original. The best way to group forms together is by spatial priority and function. By spatial priority, I mean how large and prominent the forms are. The purple drawing shows most of the forms grouped by spatial priority. The deltoids are the largest muscle group in the arm in terms of sheer volume. A man with large arms usually has massive deltoids, especially in anime. That's why the deltoids get their own group. The pecs and traps also get their own groups, as well as the various extensors in the forearm. A good way to determine what muscles get high spatial priority is just to look at your own arm. What forms stick out the most?
These muscles also are grouped by function as they all essentially work in tandem within their group. Therefore, if that area of the body moves in a certain way, that particular group of muscles will change shape. What simplification does overall is that it allows you to memorize shapes instead of individual muscles. By mentally grouping the muscles as you draw, you can portray exactly how each area of the body will look in a specific pose from a specific angle and just combine the pieces to create full body poses.
The next subject is rendering organic objects. People have told me they have trouble portraying objects with complex surfaces and contours. It's hard to imagine how lighting interacts with something that's not a basic shape, considering most man-made things are composed of basic shapes. That's why simplification also plays a role here.
In industrial design terms, this is a "draw through" of this random organic object. Though it may be difficult to interpret the lighting on the object itself, we can make it easier by turning that object into more familiar ones: cubes, cones, cylinders and spheres. We're all rather familiar with how those interact with light, so it allows us to get a rough idea of the surfaces of the object, which are hard to read in the original sketch. Once you discover these surfaces, you can start filling in the areas in between to create the whole form. Since you now have an idea of how the shading is structured, you can just bleed it into the missing space. For example, the conical and cylindrical areas have a core of dark down the middle, highlight on one side and refracted light on the opposite side. Using that knowledge, you can blend these light and dark areas into the rest. I finished it off by adding "seams" of light and dark where the forms collide with each other.
This technique can be carried over to human anatomy, which can be difficult to shade due to its complexity.
Another technique you can use is by reading the organic object as a rigid object with hundreds of faces.
Here we see a human head on the left, which I have reinterpreted as a block with many sides. Though the complexity of the head may be overwhelming at first, if we reduce it to only a dozen faces or so, we can get a vague impression of the light interaction without having to worry about smoothness and details. This is a vital skill when working with silhouettes or speed paintings as you need to be able to indicate objects with lighting very quickly. Establishing good values while remaining generally vague is a key skill.
If you were to remove the lines in the second drawing, you would have a good basis for a value sketch as the general shape of the head is already blocked out and rendered. But, to achieve the level of smoothness that a real human head has, you have to continue adding faces upon faces until the block shading essentially becomes gradients and the faces become unrecognizable. Though this is good practice, you don't literally have to draw in every one of said faces. Eventually, you'll be able to do this mentally.
Soon enough, you can take shortcuts and start with soft rendering for big forms, only applying block rendering where lights and shadows are harshest. But you have to build intuition to be able to do this.
This is a quick soft render of a female torso and, on the right, the surfaces I noted as I rendered it. By noting these surfaces, I could tell where light would be hitting directly or be entirely absent. But, because this is based on intuition, the surfaces were integrated into the overall shading.
Keep these things in mind and rendering things will become much easier for you. Good luck.