I was sarcastically trying to point out that we blender heads have a tendency to be dismissive of negative criticism when a new design task comes up to the point that we end up committing the very faults the Cassandra’s were trying to warn us off.
Not sure what you mean by “least”. The topic is honestly more split in the middle than a clear majority favouring one over the other.
I’ve tried to make my case as to why Edit Mode being the dedicated workspace for retopology is a mistake. The people in favour of Edit Mode have so far primarily talked about tools, tools, and more tools when the biggest concern should all be about workflow. A well designed workflow is not the primary focus here and it is that lack of focus that makes me think that people don’t know what they really are missing out on in other software with focused workflows.
All this talk about tools and people seem to completely forget that retopology tools do not have to all be exclusive to a supposed Retopology Mode. You can still have a dedicated mode to work with and still allow for some tools to carry over elsewhere where it is appropriate. It is not an either or scenario we’re talking about here.
3 Likes
Depends on which side of the argument is the Cassandra here. The way I see it the current course is practically the same course that’s been taken for retopology for years by sticking to Edit Mode. The course was trying to change since the way things were had its fair share of problems, but some people just don’t want to listen to the prophecy. 
I am interested.
But I have to ask. Is this because it is the tool you have used the most? Or if you were to compare it side by side to Maya, what would be the breakdown plus and minus?
I do not mean this as an invitation to argue by the way.
Issue with me for 3D Coat was it did not look to do it better than Maya. So I could not see buying it only for retopo. Not that it does not have a zillion other goodies. But there is only so much time in the day…lol
Can you give a rundown on what you have found compared to Maya?
1 Like
I’m sorry, I should have written “3D-Coat has the best manual retopology tools I know”, because I’m not familiar with Maya’s retopo tools. I just know that 3D-Coat was recommended by the majority of retopologizing modelers in several forum discussions when I bought it a few years ago. I’ve worked with it and it’s very powerful and inituitive. 3D-Coat also has an auto-retopologizer that comes close to ZRemesher, and great PBR material creation tools. I’d use it more if I wouldn’t be trying to condense all my activities to Blender only.
If you want to know more, check out the dedicated YouTube playlist.
got it.
What happens to me is I get 10 mins into a video like this:
And already I see some glaring things that are not as good - in my opinion - as Maya.
To the point of discussion.
In Maya you get into a zone and you have all kinds of tools that you would also use in Maya, but in the retopo toolset they are assigned to different and much more focused condensed set of simple key stroke and mouse options.
Without ever reaching for a menu you can draw quads, add edge loops, smooth, select and delete polygons and edges. Probably some things I am forgetting.
So the thing about a mode is that this process is a discipline. Just like painting and sculpting.
An artist can be busy a week or more on one character. Just doing nothing but retopo.
And this is a huge part of the pipeline when you have sculpting. I think people complaining about a mode for this don’t grasp this.
So let me say it again. How far will I go for the best retopo tools - away from other modeling tools?
A completely different app is how far.
So a mode in Blender?.. pftt…
As mentioned above. It is about workflow.
3 Likes
I make an add-on called PolyQuilt.
This is an add-on for low-poly modeling, but can also be applied to retopology.
Currently, when you switch tools, kd_tree, bvh_tree, and viewport transform caches are built for surface snapping and highlight.This will be an unbearable pain for you when dealing with high resolution meshes.
Therefore, it is necessary to have a mechanism to share these between tools, but it is not suitable in edit mode. A new retopology mode is required to handle these.
7 Likes
Nobody said that current situation in edit mode is satisfying.
Proposal to solve that was to make a new mode.
I saw things differently and asked for a new, better handling of snapping and highlighting by edit mode.
At a moment, a tentative to fix performances under Edit mode should happen. Not just with retopology as a target but for regular poly modeling, too.
If it is possible to do it, now, what would be the interest of another mode for few tools that are by essence also interesting for modeling ?
Nobody said the opposite.
But a workflow can be handled by a workspace.
The mode is not crucial.
The whole point of 2.8 design with mode switch and addons per workspace, high customization of viewport, is to let user forget about the modes, being less annoyed by them.
I see a new mode as a new barrier, a forced switch between complementary tools, a new module that would require a module owner (could possibly lost him and become outdated).
Blender does not have same manpower and same resources than Maya or a dedicated app.
When Pablo will pass his time on retopo tools, he will not pass it on sculpting tools.
If those tools are part of a mode that other developers are familiar with ; they could be maintained by more people.
An edit mode for low poly modeling and another one for high poly modeling that sounds really bad to the ear.
If you want to work an mesh of several millions of polys, abandon all hope to use dozen of Edit Mode tools and go into Retopology Mode, to use only one tenth of mesh tools.
That is what I want to avoid.
I am not a developer. I don’t know. Maybe, we are condemned to have bad performances with high poly meshes into Edit Mode.
But if there is one possibility to make it as fluent and more resilient without a mode, I think that’s best choice for Blender.
In other words, I think that copying Industry’s default workflow for Retopology could disperse Blender forces and keep things unsatisfying during a longer period.
Anyways, if what I am imagining is not possible : they will not to do it.
But if they are agreeing that could have benefits for users : they should try.
2 Likes
I’ve only been using Blender a few months, for learning and fun, with Max being my DCC tool. This is an area that I have great interest in as manual retopology is - bizarrely enough - one of the only areas in the pipeline that hasn’t really evolved over the last 10 years. I haven’t actually done any manual retopo work in Blender(watched a lot of videos), but have production and personal experience retopologising in Max graphite tools and Topologik plugin, Maya quad draw, Modo, Topogun, and Zbrush.
Each of these solutions have pros and cons, with each having or missing a tool from one of the others. Ideally, I would like to see a mix of Maya quad draw and Max Topologik plugin. Quad Draw base tool and the all important workflow is excellent, but it lacks the advanced features of Topologik, while Topologik suffers from a clunky programmer designed workflow. Quad Draw also suffers from lagging on more complex meshes. Personally, I’m really looking forward to the Topogun 3 beta and it will be interesting to see what is in store after 5 years of development.
Workflow has been mentioned above and this really is the key:
ease of setup(Maya live surface is a good example) live conform
ability to use standard edit mode tools in conjunction with retopo tools
visually pleasing(Blender’s in Front, Maya QD, and Topogun overlay display)
a core single tool like Maya QD with modifier hotkeys
procedural and nondestructive tools like Topologik
fast and responsive viewport feedback on multi-million meshes
One of my favourite things about Blender’s unique modeling workflow is the ‘live’ concept of interacting with an active tool dynamically and calling different operations through hotkeys. It is so powerful and a perfect environment to create an all-round manual retopology toolset/workflow.
Here are some of the advanced features of Topologik:
7 Likes
Nice to see you here ! Ah, well - if it’s technically impossible, then I guess we’ll have a mode. Which is ultimately fine, although I have the same concerns as Zeauro.
Speaking of (manual) retopology, Topogun will also get a new life soon:
… But lazy as I am, I’m eagerly awaiting… Here we go again… Drumroll… QuadRemesher! 
reading what they said.
TopoGun had to be rewritten from scratch, to make room for the new features.
that must be something big maybe one magic button or fewer clicks, otherwise it won’t be worth it, too many competiton already.
1 Like
Surface snapping in Retopology mode is not simple.
Depending on the operation, the function may need to find a point on the nearest face, a ray in the normal direction, a ray from the viewport, and so on.
After all, most retopology tools require a bvh_tree, but if there is a way to share it in edit mode then there should be no technical problem.
In addition, the current edit mode lacks many functions to create a retopology add-on. I expect them to be added for the new retopology mode.
2 Likes
second pass of improvements…going maya’s quad draw style.
https://twitter.com/pablodp606/status/1164953947628539906
11 Likes
Looks promising. I guess we are at the zspheres retopology level now. 
Edge extrusion next plzzzzz.
If an mode switch is necessary because of different handling of mesh data…can’t the mode switch be hidden, under the hood so to speak, while maintaining the illusion of keeping it in the edit mode context?
I don’t care what happens under the hood, i only care about the result and from an UI perspective i think another mode is bollocks.
Lets take the middle way, best of both worlds.
I don’t have any resistance to having universal tools. Or integrated workflows. I have been a huge opponent to walling things off. Very outspoken about this as Blender 2.8 was ramping up.
But I am going to just speak from a user of retopo.
I am not exaggerating when I say one of my artists spends several weeks on retopo alone. In fact I am being conservative. I have some characters so complex it has taken over a month.
What I don’t think most people understand is that retopo is just as much of a discipline all by itself. In exactly the same way that sculpting and painting are.
Really. It is. Don’t be fooled by the fact it is considered modeling.
Now. Also I find my self directing my artists to Blender for other aspects of the retopo process that can be done with modifiers. Could do this in Maya but the Blender workflow is nicer for this kind of thing.
So I think that another aspect of this is to appreciate that many artists especially professionals and studios use the best tools for the job. When possible.
So for instance in my studio we have Zbrush Substance Painter, Maya and Blender.
This is not ideal from a pipeline perspective. But it is ideal for the artist doing each task.
It is not like you are sculpting for a few hours and then move to retopo and then to UV and the next day you are painting.
So the idea of a fear that things can become segregated is unfounded as they already are in terms of time.
It is not realistic to consider that having things separated will cause confusion or waste time.
The workflows are already set up like this anyway.
All this said of course if things can be integrated wonderful. But it should not be integrated for the sake of integration. This is a wrong target.
That target needs to be the best retopo workflow first.
Inside a blender file.
But the Elephant in the room is viewport performance.
Need to address that if you are bringing in sculpts to retopo. Even decimated ones.
Blender handles large meshes quite poorly.
4 Likes
I have to disagree with the notion that another mode is “bollocks”. Essentially what is happening with the “modes” is that you are designating a task. This can never be a bad thing. Switching tools or switching modes, its practically the same concept. The benefit of the mode is that you can actually have the UI match the workflow and offer a set of tools rather than just one tool.
One of Blender’s biggest problems in the past was that it started squeezing in too much when there was very little UI real-estate to hold it all. In a way we have to get away from the old blender mindset of having everything thrown at the user in a messy format, and start clearly streamlining and defining task specific workflows.
Imagine when switching to a retopology mode, the background mesh goes semi-transparent automatically while allowing task specific tools to quickly retopologize over that mesh. Maybe a grease pencil tool specifically for automatically drawing along the surface of the high poly mesh, automatically creating topology between lines. Without the task specific mode the user might require a handful of steps to set up that ideal environment for retopology, including snapping, transparent mesh, pen behavior…ect
What could also happen is simply adding a topology “pen” or brush to the edit mode or sculpt mode so that you can still “retopo” in those modes but on a very basic level. This is essentially what Modo does.
Maya via NEX tools has also been brought up a few times and I think they take the same kind of approach Modo does, with a set of tools that have behavior geared towards retopology. I think both the mode and a few basic tools outside of the mode can exist. For the mode itself, tying it to an actual workspace tab is the most efficient way to go about the task of retopology.
Finally with a separate mode you can actually specify viewport behavior that can potentially handle much higher polygonal meshes to speed up performance. Modo did this with “reference meshes”, which allows extremely high poly meshes to exist with very little viewport slowdown, specifically with the intent of retopologizing on top of them with a new mesh layer.
As I also mentioned earlier, if we feel there are too many modes already… 3-4 of them can be merged. Meaning, the 3 paint modes do not need be separated, as their workflow is generally the same. Better to actually have select layers you can paint on in that mode, such as vertex paint layer, texture paint layer, or weight paint layer…ect
5 Likes
Nah, It’ll be fine without all that jazz.
You have brought not one single reason why a separated mode should be preferable.
You said so yourself:
Why use a way which needs more effort? One click more is one too many.
The change of viewport could be instantly initialed whenever you activate the retopo tool, snapping can be automatically activated whenever you pick the object to snap to. Just cut down all the unnecessary steps and automate them.
You pick the tool, pick the HighPoly mesh and start with the retopo. There is no need for another mode, not even another workspace.
But you are right with one thing, the paint modes should be merged.