Shape keys and applying subdivision surface modifier

Question

I have been creating a dragon model lately (it might come familiar to someone) and I've done a lot of work so far, I've also spent alot of time with creating shape keys for my eye. However, I see now that I have forgot to apply the subdivision surface modifier for my eye and when I remove the modifier it looks totally crappy.

I'm very-very confused now as I've looked help in Youtube and I've seen that almost everyone create their shape keys with subdivison surface modifier on (without having applied it earlier). I'm clueless how can this be, do they later on somehow do some kind of workaround to apply the subdivision modifier??

Am I really supposed to create shape keys for my eyelid or eyepupil when I have some crazy polycount (it's very hard to form the shapes I had previously).

I'm a bit in panic, so please pardon me. Maybe you could offer some ways to get through this :)

Answer 1

This answer comes from a previous one here Cannot Apply Array Modifier. I simply add some more details in the explanation and a way to do it for all shape keys at once.

Here is a simple shape, which has a subsurface modifier and shape keys :

Make one copy of your initial object for each shape key and set for each the shape key to "1" :

Select this copies and Alt + C and "mesh from curve/..." :

The modifiers are applied and the shape keys too (these objects now have no shape keys or subsurf modifiers).

Make another copy of your original object, remove the shape keys from it, and apply the subsurf modifier :

Now select the "shape keyed" objects, and the last copy we just made (in this order), and in the shape key panel use "join as shapes" :

We now have the final object, with the subsurface modifier applied and the shape keys available :

Answer 2

I scripted @Lemon 's answer!

the code is kinda long, but I think it is (I hope it to be) self-explanatory. Just copy the code into a new file on blender's text editor, call the text file apply_with_shape_keys.py, and then, select the object with the modifiers and shape keys, (make a backup, just in case). Now from the console execute it like this:

>>> import apply_with_shape_keys
>>> apply_with_shape_keys.super_apply_modifiers()

the script will apply all the modifiers conserving the shape keys! and will create a new object called "backup" without the modifiers applied (just in case you wanna add new shape keys in the future).

If the object has an armature, the script will explicitly not apply that modifier. Also you have to make sure that the skeleton is in the original pose, otherwise strange things will happen.

here is the code!

import bpy

def reset_shape_keys ():    
    for name, shape_key in get_active_block().items():
        shape_key.value = 0

def get_active_block ():
    block_id = bpy.context.object.active_shape_key.id_data.name
    return bpy.data.shape_keys[block_id].key_blocks

def select (selection):
    bpy.ops.object.select_all(action='DESELECT')
    selection.select = True
    bpy.context.scene.objects.active = selection

def select_last_shape_key ():
    shape_key_count = len(get_active_block().items())
    bpy.context.object.active_shape_key_index = shape_key_count - 1

def remove_shape_keys (object):
    selection = bpy.context.object
    select(object)
    shape_key_count = len(get_active_block().items())
    select_last_shape_key()
    for i in range(0, shape_key_count):
        bpy.ops.object.shape_key_remove(all=False)
    select(selection)

def apply_modifiers (object):
    selection = bpy.context.object
    select(object)

    for key, modifier in object.modifiers.items():
        if key != 'Armature':
            bpy.ops.object.modifier_apply(apply_as='DATA', modifier=key)

    select(selection)        

def super_apply_modifiers ():
    original = bpy.context.object
    bpy.ops.object.duplicate_move(OBJECT_OT_duplicate={"linked":False, "mode":'TRANSLATION'}, TRANSFORM_OT_translate={"value":(0, 0, 0), "constraint_axis":(False, False, False), "constraint_orientation":'GLOBAL', "mirror":False, "proportional":'DISABLED', "proportional_edit_falloff":'SMOOTH', "proportional_size":1, "snap":False, "snap_target":'CLOSEST', "snap_point":(0, 0, 0), "snap_align":False, "snap_normal":(0, 0, 0), "gpencil_strokes":False, "texture_space":False, "remove_on_cancel":False, "release_confirm":False, "use_accurate":False})
    backup = bpy.context.object
    backup.name = 'backup'
    remove_shape_keys(original)
    apply_modifiers(original)

    for key, shape_key in get_active_block().items():
        select(backup)
        bpy.ops.object.duplicate_move(OBJECT_OT_duplicate={"linked":False, "mode":'TRANSLATION'}, TRANSFORM_OT_translate={"value":(0, 0, 0), "constraint_axis":(False, False, False), "constraint_orientation":'GLOBAL', "mirror":False, "proportional":'DISABLED', "proportional_edit_falloff":'SMOOTH', "proportional_size":1, "snap":False, "snap_target":'CLOSEST', "snap_point":(0, 0, 0), "snap_align":False, "snap_normal":(0, 0, 0), "gpencil_strokes":False, "texture_space":False, "remove_on_cancel":False, "release_confirm":False, "use_accurate":False})

        meshed_shape_key = bpy.context.object
        select(meshed_shape_key)
        reset_shape_keys()
        get_active_block()[key].value = 1
        bpy.ops.object.convert(target='MESH')

        select(original)
        meshed_shape_key.select = True
        bpy.ops.object.join_shapes()
        select_last_shape_key()
        bpy.context.object.active_shape_key.name = key

        select(meshed_shape_key)
        bpy.ops.object.delete(use_global=False)