just want to bring my lil grain of sand to this topic.
I very recently switched to 2.9 for a big project ( see my portfolio for more info ) where blender usability stability and workflow is a critical path to me.
I use blender for modeling sculpting and texturing and check-rendering but also also some few times for animated toons creation.
I previously used 2.79.6 wich was perfect except for cycles ( wich has too early replaced BI IMHO ).
BI was not very easy to setup but it had all features i need and it worked properly for all its features.
Cycles is noisy and buggy.
Why i switched to 2.8+ ? because cycles renders are far less noisy than on 2.79 and direct & indirect clamp work properly.
I needed textures bakes less noisy that the one you get in 2.79 cycles. And 2.9 was the answer.
Now what about all the changes made in UI, shortcuts, 3Dview look, etc… ?
My 1st solution was customization. I erased all useless things i could on the 3Dview.
My 2nd solution was to dive in shortcut system to retrieve my Z Key toggle between wireframe and solid.
My main pain was to find a way to have textured meshes without render. It’s done in solid options.
I simply added a PAGE UP shortcut for toggling from solid to textured when in solid view mode.
I simply never use pie menus ( that IMO are boring and useless when you use a mouse+keyboard )
Aside this i really appreciate the multi-objects edit mode. It’s particulary convenient when editing and unwrapping multiple objects on the same UV. Now this exists it lacks of multiple objects simultaneous handling but this is another story.
Finally i slowly get used with 2.9 hoping the dev team won’t decide to chang the future UI again in future versions…
A big deception though is the non retrocompatibility of files from 2.9 to 2.79. I can easily understand the 2.9 much bigger .blend files handle more things but an ‘to 2.7x’ exporter in file menu would have been the least thing to dev. Maybe as an addon ?
Now i confess i like to use 2.9 even if i use very few of the later features.
Thx for the effort, I know the X-Ray thing. And I know I can press Z to switch the shading. Wait: to show a pie menu, where I can then select the shading.
So,
before: wireframe display -> Alt + Z -> textured display.
now: wireframe display -> Z -> Pie Menu -> mouse and select -> Materials preview.
Why should I switch? To spend double the time it took before?
This, obvious enough, makes me even more stick with older versions.
Not my cup of tea really.
I will most likely stay for a very long time with latest blender 2.79 , which is super powerful and mature for my needs. Because i use blender mainly for my 2D cutout animation kit:
Up to now all blender versions from 2.80 and up have brought me no improvements, but only unsolvable difficulties. I have already invested too much energy and time in a possible change, but only with unsatisfying results.
So far i don’t like the new blender UI neither the increased resource hunger. What I find worst is the omission of previously useful features like the good old screens (now workspaces) dropdown menu - the new horizontal one is a cumbersome joke.
At least I can understand all users advocating a higher, further, faster Blender development. Especially Eevee & Grease Pencil are promising for my purposes. But to have fun with those new features I would need the latest CPUs & GPUs. Actually I can’t afford new hardware for it every few months though.
I’m sorry to hear the newest versions of Blender does not run on whatever it is you have available to you. With that said to imply Blender requires top of the line hardware to run is absolutely ludicrous.
I use blender 2.79 . My pc is not compatible with new versions.When I use modifiers my blender 2.79 crashes and nothing renders. Please tell me how can I solve this problem.
You probably should create a thread in modeling section of support forum.
Modifiers may be demanding. If your pc is not compatible with 2.8,2.9 ; it is probably not very powerful.
If you pushed too far a setting of a modifier, that may be the cause of the crash.
As I said , a pie menu is a quick way to switch to another mode and use same shortcut for several things.
That is not slower than a shortcut made of a combination of 3 keys.
If you make mouse move before of release of Z key, switch may be quick.
You are not forced to make a switch.
I understand that you are not making a switch for pleasure of customization.
But probably, several 2.8/2.9 features could speed up your workflow (EEVEE’s rendertime, Cycles improvements, Sculpt mode refactor, Library Overrides, Multi-Object Editing, Grease Pencil Object, VR support…).
There may be valid reasons to switch. And often, simply, a little effort of customization is sufficient to feel comfortable and enjoy what is really important.
After that effort done, you may experiment a gain of productivity, although some stuff could stay frustrating.
Sure, “Every few months” maybe a bit of an exaggeration, but it describes the core issue.
As long-time blender user (since version 2.37 - can you keep up with that?) and community member, I have already noticed how fast blender has developed since 2.80 in terms of hardware requirements. Before that, I could easily use blender on the same hardware over many version jumps.
I was just sharing my experience and opinion on the thread topic.
I wish you a merry Christmas, we don’t need to start a blender fanboy debate here - not on Christmas.
A pie menu is something you have to reach with your mouse - both if you click on it or not.
Sorry if I insist, but - for me - clicking Alt + Z is WAY faster than having to press Z and, after that, select a voice in the pie menu.
And clicking Z to cycle between shading modes (in a direct way like it was before) is certainly way faster than having to use the mouse and move it on the pie menu choice.
Should I seldomly need a different shading, it’s THEN that I want to have options. Not every single time that I need something much simpler, like switching between 2 shading modes.
Imagine doing this hundreds of times in a normal day of work. Why should I?
And, just to clear this: I would like very much to switch to the new blender, but the workflow is way worser and slower than in previous versions.
The new blender can be better for newcomers with a new pc. If they start from scratch they will never know how fast the old workflow was. And bells and whistles will always be there to make them happy.
But a tool is something different from a storefront or a book cover. Ton made it very good in the past.
Was it ugly? Yes, in a way. Was it badly organized? Yes and no. Was it fast? Yes, it was blazing fast.
still using 2.79 for modeling and everything other than rendering.
because it’s just slow. feels like blender found a new friend and ditched all the old users that has been using blender for many years. (came on when cycles came online)
i have 3ds in my workplace, and i still use blender because it was fast, everything you had to interface was in the keyboard, i barely have to read anything when using 2.79, when you remember something, it’s just there, don’t need to fish my mouse through the the drop down menus icons, or click into a submenu of a submenu of a submenu. everything was just lighting fast, even faster than in 3DS.
now blender is becoming 3ds max, still remember when i had to use the pie menu, and within that there’s another pie menu, and then a drop down menu untill you reach there.
one thing might be that i just need to learn the new locations, but then stop changing it!! another thing is that there’s just too many drop down menus, and too many extra buttons that i need to click to do the same thing.
it’s good to add periphery functionalities, but it’s like having a child dropping everything to pick up a road side candy. how’s anyone ever gonna work with that?
my suggestions: have a 5 month study, make a plugin that tracks all the functions ever called in a session of blender, and upload them, i’d give up my privacy for blender, i know they won’t sell it. then make the top 85% of them a one click function and never touch them.
another gripe is the grouping systems and the layer function, yeah it’s easier to manage alot, but not at cost of me trying to fish out something and wasting even more time.
so, to sum it up: blender used to be a muscle memory software, you don’t even have to turn on your brain to model, that is how smooth 2.79 is.
now it’s a chore, and if i wanted these extra functions, i’d go 3DS and port it back to 2.79
I think thats going too far. Maybe a better alternative would be a community driven, “back to 2.79 future keymap” project, where we make a GitHub repo with the Key map file and people can make pull requests for adding the shortcuts that are most likely to be used, along updating a cheat-sheet for reference, with the goal of having the most 2.79-ish keymap possible maybe with a few extras or tweaks and an add-on if necessary.
Brecht tried to modernize Blender Internal with the render25 branch, it worked well enough for Sintel, but it still had its usual issues of unexpected bugs along with features that just couldn’t play with each other (ie. anything after the first bounce having a yellow tint).
Now you might try to cite UPBGE as an example of something old, crusty, and messy being revived, but that team even now has had to deal with the old code mess rearing its ugly head (like some features getting major new bugs almost at random). If you don’t believe me, you can take a look at UPBGE’s issue tracker on github. Sometimes, the idea of keeping old project files compatible forever is a lot more trouble than it is worth.
but that’s not the problem, i do have the keymaps remapped and sorted in my 2.91 version (which i only use for rendering)
the problem is that because there’s alot of re-sorting functionalities, so that hierachies makes sense now for new user, everytime i call on a function there’s just more clicks, and more importantly, they’re mouse clicks.
when you use your mouse, you need to dart you eyes back and forth to find the thing you want, that totally breaks the workflow, and that makes it tiring… when i was trying the new 2.81 version, i found myself just using the search function all the time, that’s how bad it’s become.
to understand, this is what a “responsive application” should work like:
Your eyes are always looking at the thing you’re modelling, not once do you have to remove your attention from the subject, so your attention is undivided on the subject untill you’re done.
if that’s still not clear, imagine using a word processor, but everytime you have to a capital letter, you go to special characters, then capital letters, then scroll through A to Z untill you find it. you need a comma?
swim through a drop down menu, make things bold? drop down menu/hidden menu, underline something? drop down menu.
now can you type quickly like that? no, and that’s what it feels like using post-2.80
it’s not keymapping that’s the problem. there a limit to sensible keymappings, but often used things should not be in hidden menus or submenus, they almost need to be as accessible as quick settings.
That’s why we need a more complete keymap, a single keystroke will always be faster than anything. And as long as most commonly used operators have a shortcut I’m fine with darting my eyes around for that rare occasional operator I almost never use.
But they removed my shortcuts, and the keymap editor is a pain to work with
It seems I’ve opened up a can of worms that I both like an am afraid of. I don’t want Blender split into two different versions, but according to much of the discussion in this thread that’s exactly what’s happened to some level. A version that follows the philosophy of awkward but simple (2.5+), and a version that follows the philosophy of industry standard but laborious (2.8+). Often times I find things that I like be it hardware, software, or services, and I specifically like them because they are not following everyone else’s standards and are thus easier to learn, flexible, or fit my needs without extraneous features. That’s what Blender has been for a long time to me, but that’s something 2.8 started moving away from in the UI and workflow.
I just hope we can find a middle ground (if I’m being honest, trending more towards 2.5/2.7) and that some of this gets back to the developers.
I very much highly doubt Blender is going to split. It’s got more momentum support and buzz behind it than ever before and a lot of that is because of the direction it’s been on since 2.8.
I thought the old way of using Blender was great once got used to and very fast and efficient and I was a big defender of it myself as well. But Blender will increasingly follow more common industry standards now. Part of that is because studio pipeline integration is now a top priority.
There will always be very genuinely heartfelt debates in the forums as there was about cutting the BGE and the open movies vs game development. We all understandably feel deeply about the important tools we use for our work. And Blenders community based support makes these debates even more passionate. But I think the general agreed direction of travel is pretty clear now.
When 2.8 came out I quickly adapted myself to the new workflow as that was obviously going to be the way in the future. As I also quickly adapted to the 2.5 workflow when I first learned Blender.
Talking about older versions though. I’m still using 2.79b a lot and love it. Mainly for the in house studio work I have been doing with an artist sculptor and printmaker based in London where we are designing large scale gallery installations and creating short films and animations.
The two Mac’s we have at the studio are quite old and won’t run 2.8. But 2.79b works great and is solid as a rock. I also have such a nice Blender Render pipeline going with it now for our specific work.
Everything is mostly composited in 3D with render layers as we don’t have the time or the hardware for rendering multiple passes each time. But I have been amazed at the complexity of the composites I can achieve simply compositing in 3D with multiple render layers. And at the quality I can get into the BR renders this way as well. It’s been these recent projects that have helped me really get to grips with the Blender compositor and it’s potential.
I have also switched up to 2.8 to 9 in my own work. But I have still not yet achieved this level of flexibility in Eevee. Although I am sure it will mature. I love the new motion blur that has suddenly appeared.
But 2.79 still works great. It’s plenty good enough for what we are doing right now. And that’s one of the beauty’s of Blender. It’s always going to be there. Always be around to download. And no tangled licensing issues and torturous hoops to jump through to legally instal and run an older version and pipeline. So right now we have a 2.79 pipeline that is working really well for us. We don’t feel under any pressure to change it up for change’s sake just yet.h
Yes. But movement you need to do to make a selection is really tiny. 12 pixels by default.
If you get used to placement of item of pie-menu, it can become a reflex to move mouse to left to be in solid view in just the time to press and release Z key.
That can be faster than trying to reach Alt key and combine it with Z.
That is what I observed.
It probably would be a less perturbing choice to set a pie-menu to switch from one Select mode to another instead of setting 1,2,3 keys like in other software.
There literally were months of discussions about new keymap.
You can’t satisfy everybody on every point. At least, G,R,S were not remapped.
Remapping could have been a lot more massive.
But nobody is forcing you to use defaults, if you don’t like pie-menus.
You can disable pie-menu and change shortcut to have it exactly like in 2.7x.
Default Keymap had to change to let people discover all the ways to customize Blender.
The 3 and 4 key combinations were awful gymnastics for people frequently using those shortcuts.
Users probably changed those shortcuts by themselves while using 2.7x.
In 2.8, default keymap lets F-keys from F5 to F10, free for customization.
2.7 was showing what it is to have all keys of keyboard used.
But it was one way to do things. It was not the most efficient one, for everybody.
Such thing does not exist.
So, 2.8 lets everybody fills some blanks.
But there is no reason to consider that same ability of customization + more to be slower, at the end.
The access to properties and buttons under subpanels & popovers or things stuck at a certain placed that can’t be moved elsewhere are valid reasons to complain.
But the default keymap is not really one, if you have ability to easily create your own.
Ant that is the case.
That is a lot more painful to rewrite the whole UI to hide unused panels and buttons.
If they follow WIP videos from pros, they will.
That is just as simple as saying to them : if they want to be fast ; they have to create a keymap using keyboard at its full potential.
Keymap configurations are files that can be shared.
Yes. 2.8/2.9 supports a more recent version of OpenSubdiv library, with more options for crease and UVs. But currently, that is just supported by CPU.
GPU acceleration is still a todo.
Work for Vulkan support has started. It should continue during the whole year 2021.
There is an official 2.7x keymap bundled with 2.9 for this goal. If there are problem with it, we should fix it, first.
That why I liked 2.79’s 2 column layout. Why I found Grid Layout unsatisfying.
Why I zoom out in my 2D view and use Page Up/Page Down shortcuts instead of scrollbar. Why I wish I could snap Properties editor at a top/mid/bottom level instead of using those shortcuts.
Why when I start a task, I try to close irrelevant subpanels and open relevant ones, once for my whole session.
But I have to say that if we have more clicks to do : that is not just to keep things clear for new comers.
That is also because we have a ton of new tools, new settings and new options to deal with.
That is a way to manage expansion of Blender.
If 2.79 UI had to deal with same amount of settings, you would have to scroll more, too.
Render tab had to be split in half because EEVEE is handling at scene level more things that Blender Internal. Some that were World properties (AO,IL) but are, now, considered Render stuff. Some that BI was unable to handle (Volumetrics, SSS, SSR, Bloom).
We can discuss pertinence of subpanels versus more panels, pertinence to put a frequently used setting into a subpanel. But we can not neglect the fact that refusing more panels has as consequence longer ones taking all vertical space.
That is why I am annoyed by Adjust last Operation panel position.
That is where things are getting tricky with the inflation of Blender community.
People are no more agreeing on standards, on what is easy to learn or flexible or what is needed and what is an extraneous feature.
If you want to be satisfied, you have to take part of discussion.
But respecting expectations of others means that you have to consent to presence of things ; that you don’t use, but may ignore after some customization.
That would be a little bit naive to think that developers can make a satisfying result just by collecting feedback from divergent voices.
Users have to discuss between themselves to come to an agreement that can be submitted to and understood by developers.
Dissatisfaction can happen at any step.
Users are not able to agree. Or they agree but agreement is not translated into a proposal. Or developers simply don’t understand what the proposal is about or why it is a priority. Or they do but they don’t have manpower or time to make it real, now.
To me, produced design tasks are already a middle ground, but not enough of them have become real, yet. And I think that is often a misjudgement to blame what has been done for what is missing.
Because what is missing is often just missing. And what has been done is not necessarily the reason of the lack.
Of course, when there is a replacement in order to experiment a new way, that is legitimate to blame that replacement for a lost of a loved old way.
But that does not mean there is no teaching to retain from the experiment to do better than the old way.
I think that developers have always seen 2.8 as an experiment before 3.0.
I think they have underestimated the workload, related to all those changes and their ability to reverse what was not working well, in time.
