Organization of Asseto's storage and ref. codes in gamedev?

Question

git

How to properly organize the storage of project files, say for Unity3d for joint development. My vision of the situation is as follows:
1. Git repository with git Ignore written .
It turns out that there will be a develop branch on git and master in the last branch will include all approved versions of the source code and models. (in fbx format) so as not to take up much space (Is it right to store fbx,png,jpg on git?) as well as textures for them .
2. Whatever Google Drive, with PSD sources, and 3dsmax files for models. For each model, images are a separate folder with files named by date. And a readme file describing them. And in git there is already a link (web shortcut) to a file in the cloud.
Can you suggest a better option for PSD,3dsMax?
Did I understand Sergei's idea correctly?Sergei in this topic?

and in general, it is easier to store huge files somewhere far away, and versions in the form of links.
The bottom line is that all version control systems are designed for text - so that you can do a diff, but it makes no sense to do a diff for media files, so it's easier to just drop them under some number (date), and store the history of changes in a text file / config .

There are also all sorts of perforce which, as I understood at first glance, allow you to store both source codes and media files in one repository. and make a checkout for the latter. But the prices for indie developers are cosmic, namely:

20/20 Program: Free for Up to 20 Users
Perforce Helix SCM and content collaboration solutions are free for up to 20 users or 20 workspaces.
And then comes
Users 1-50 $320 per user

I'm guessing stuffing everything into git is stupid. And git will do comit for a very long time, although I didn’t check it, and I heard about git lfs , but as I understand it, this github extension cannot be raised on my machine.
How would it be more correct, or how would you organize joint development in this situation, let's say, having ~ 30-35 employees in the state.

Answer 1

Keeping binary data in a repository is not a very good idea. (voluminous)
Everything that has ever got into a commit will forever remain in the repository, even if you physically delete the file.
I don’t know what to do, here everyone is looking for their golden mean. :)

Answer 2

I settled on Git + git lfs. raising everything to vps. At first I looked in the direction of git-annex, I really liked that it saves symlinks. But unfortunately, it does not have support for windows (more precisely, it does, but in direct mode, which is extremely inconvenient). Since the git lfs server is open, and github is developing, it has stopped on it. In addition, it is very convenient that you do not have to use third-party commands to work with large files. You just need to set the filter.