F16: Let's improve the user interface ;-)

I'm thinking that the main problem with the fusion user interface and even with the Davinci user interface is that Blackmagic tries to maintain a clean user interface design "millennial style with large buttons (a la fischer price toys)", but in my case ( and I speak only for myself) I want as few clicks as possible in my work ...

just take a look at the new redesigned Blender 2.8 user interface ... oh man, that was wonderful ...

It's time for BMD to think about tabs, less waste space, fewer pages, text and / or icons (good icons) user interface, and take advantage of every space you can to implement fewer steps in the workflow ...

make the use of the pages something for very particular and special operations, not a forced step for the simplest task that could be done in a menu or button redundant but very useful in the same page in which we currently work.

But what forces the toolbars to be in the same order?

Not sure what you mean here by 'same order'?

And it still doesn't address the issue of the icons being unreadable in use in the flow. There's plenty of room for improvement on the icons in Fusion 9, but at least you can read them when the scale is >= 50%.

I think everyone is agreed on this and I am sure this will change.

But what forces the toolbars to be in the same order?

Not sure what you mean here by 'same order'?

We agree that icons can work great when combined with a spatial mnemonic. It's not just an orange cog, it's also the orange blob second slot from the right, and you don't really even have to know it's a cog out of the corner of your eye because you know based on it's position what it is.

That's also why a "copy" icon doesn't have to be perfect if it's next to a "cut" icon that looks like a pair of scissors. We get a hint of what the "copy" icon is for because it's next to the scissors.

What does a square mean as an icon? Nothing. But when it's between some isosceles triangles, we know it means "stop".

But as soon as you remove the spatial context, the icons have to become a LOT stronger. If you wanted to make an icon for "stop" that could work without any spatial context, you might make a red octagon, for instance.

But what if you have no spatial context and you don't have a standard or meme like an international convention on stop signs? What would a universal icon for "pMerge" be? Probably not this:

pMerge.PNG (716 Bytes) Viewed 76839 times

So what's the fallback when icons fail? Almost universally, it's text. We have a library of symbols for road signs like stop and yield and pedestrian crossing, but the overwhelming majority of street signs use the name of the street written out in the local language. It's just not practical to expect drivers/users to learn a enormous catalog of icons when text (itself a collection of glyphs that the users must learn) suffices. Stop signs are still red symbols, and that's fine because we both need them to be instantly recognizable even when we aren't focused on them, and it's a common enough sign to be included in the small catalog we expect users/drivers to learn.

And it still doesn't address the issue of the icons being unreadable in use in the flow. There's plenty of room for improvement on the icons in Fusion 9, but at least you can read them when the scale is >= 50%.

I think everyone is agreed on this and I am sure this will change.

I'm not so sure. These issues were raised pretty clearly in the Resolve 15 beta and not only were they ignored for that beta, they were never addressed in any point releases and here we are with Resolve/Fusion 16 and all they've made it worse, not better. At the very least they doubled down instead of addressing the issues.

Greetings, Everybody.

I just registered to this forum, because I was really baffled when facing the new Fusion UI and I am very concerned that it could stay this way.

My first steps in (Digital) Fusion date back to the late 90s on version 2.52, if I recall correctly.

Back then, I was using various other tools and platforms like Discreet Logic's Flint/Flame/Inferno, Quantel's Henry and Edit Box, as well as learning some basics in Softimage3D that eventually became XSI.

Experiencing the freedom, flexibility and precision that Digital Fusion offered in those times already, resulted in a fundemental change of my approaches and workflows to any given task.

Ever since version 3.x came out and got rid of the admittedly funny, but rather toyish "tubes on a checker board" design, the usability and efficiency improved greatly and even the somewhat fundamental UI change from 4.x to 5.x was pretty easy to get used to.

...

However, after launching Fusion 16 for the first time, I was facing this new Resolve style interface.

Knowing Fusion for that long, it was a given to me, that I could easily adjust it back to where it came from, in order to suit my needs and my decades long grown workflow habits.

As this whole thread points out, it is a very diversified topic and my hopes were too high.

I can agree to most of the positive and negative postings, but Chad pretty much hit the nail right on its head.

I also fully understand BMD's approach of trying to unify the UI, especially since Fusion has become an integral part of Resolve's package.

...

As it is right now, I cannot get any work done in time when using Fusion 16.

Fortunately it peacefully co-exists with Fusion 9, so that I can keep comping away and meet my deadlines.

In an ideal world, I'd really like to see the already mentioned option to switch between the old and the new interface, including all possible customizations that your long term userbase loves and has gotten used to over all those years.

Obviously, I have no idea how many resources it would take up, but maybe it could be an idea or a reasonable compromise to keep the stand alone version's UI as it was before and use the new interface only within the resolve package, until it's finished and polished ?

Hoping the best for my favorite compositing package, I simply don't have the time to work my head around this new UI and thus, I'll stick with Fusion 9 until this issue gets resolved (pun intended) in any future version of Fusion.

Best regards,
Maggi

Something to consider; Resolve now supports half a dozen languages and in many locations text just won’t fit where icons will. To grow Fusions user base we need to appeal to the many non English speakers and good icons do that. We clearly have a number that need improvement and all the repeats need their own but a text only I/f isn’t easier translatable given the sizing considerations on screen.

On the surface this seems pretty reasonable, but... I think it's justifying a poor solution to a problem that doesn't even need to exist. Many other DCC apps handle this problem quite elegantly without any kind of graphic design wizardry or show-stopping localization problems.

Let's take a look at 2 other node graphs:

Nuke:

(also: better underlays & stickies... that's another post though)

Houdini:



Plenty of other examples you could use, but these are my core competencies so it's easiest for me to talk about them.

You can use icons or not, but if your primary means of identifying a node is textual then there is nothing to "design" around when you localize. Menus, tooltips, manuals, etc already need to be localized. If you have to translate those node names into some kind of text anyway... why not use that same, system-rendered text in the interface as well? A text field can be longer, shorter, whatever the language requires -- but they are the closest things to a universal interface object that we have, so if you want a functional interface that's easy to localize... why not use text?

But you don't have to stop with text, nor should you. Modern Houdini & Nuke use multiple levels of design to convey meaning simultaneously:

+ Text identifies the specific node operator and/or label (NOT region-specific acronyms, but actual words)
+ Shape and/or color tells you the general logical grouping of the node for zoomed out recognition of purpose
+ Icons are a bonus to give some common nodes an easy indicator (nothing requires an icon though, so you don't have to spend design resources on every new tool)

This gives you as many as 5 ways to identify a node simultaneously, all of which can respond dynamically to zoom or user customization in different ways.

To compare, I'd say classic Fusion falls somewhere in the middle for functionality:




+ You've got easy to read text labels but those acronyms used to identify node types only work intuitively in English, making localization harder. And because they're in-line with the label they're often hidden by the small space they are constrained to.
+ Icons do a pretty good job describing what each node does, but only when the tile picture is visible and not being used as a thumbnail; so it's mostly a feature of the Bin, not the Flow.
+ Very little shape differentiation for logical types, which limits quick readability to a single factor: color, which overlaps with user customization.

To back up a step from this granularity about icons, the UX changes in Fu16... just don't succeed on more than a handful of minor points. I don't mind the little panel toggle buttons from Resolve and the new UI looks subjectively more... modern? But what does that get you if the functionality is a regression?

These are clearly solvable problems and the lowest hanging fruits ask for little creativity on the part of the developer: Nuke is not pretty but it does the job of letting you know what's going on and putting the buttons + information you need right in front of your face.

At NAB 2019 Grant Petty made a big thing about reducing the number of clicks and the amount of time editors spent performing common tasks. BMD set out to try a whole new editing interface and even a new piece of purpose-built hardware to solve the problem. I don't know yet if the Cut Page + Editor Keyboard will catch on or not, but it's great to try things and most importantly, these new ideas don't break the existing tools. The don't fall into the Apple trap.

The Fu16 interface is adding clicks & making information harder to find. Unifying the UI with Resolve is fine in principle, but not if it ruins the Fusion UX. Hobbling UX to for the sake of branding feels like a development goal that comes from marketing, not from users.

I'm trying not to be overly critical of a still functional tool, but the original (loose) timeline for full Fusion integration was ~18 months, and that goal was stated around NAB 2018. We're getting to the end of that timeline and I'm worried this is what the goal looks like.

You've got easy to read text labels but those acronyms used to identify node types only work intuitively in English, making localization harder.

Why would it be harder to localize? You wouldn't translate the OpIconString directly, you'd make a new one based on the translation of the OpDescription.

You've got easy to read text labels but those acronyms used to identify node types only work intuitively in English, making localization harder.

Why would it be harder to localize? You wouldn't translate the OpIconString directly, you'd make a new one based on the translation of the OpDescription.

I'm only saying relatively -- it's just as easy to do as translating the words, but... one more step? This is just responding to Peter's justification concerning the work of localization being greater with text than with icons, it feels pretty arbitrary to me whether you'd have to translate a word vs. word + shortened word but I'm not doing it, so...

You've got easy to read text labels but those acronyms used to identify node types only work intuitively in English, making localization harder.

Why would it be harder to localize? You wouldn't translate the OpIconString directly, you'd make a new one based on the translation of the OpDescription.

I'm only saying relatively -- it's just as easy to do as translating the words, but... one more step? This is just responding to Peter's justification concerning the work of localization being greater with text than with icons, it feels pretty arbitrary to me whether you'd have to translate a word vs. word + shortened word but I'm not doing it, so...

Localization is a long way off for Fusion, I have no idea where it is for Resolve, I can't seem to find localized versions to see how it's implemented there.

There's no localization support in Fusion, that's for sure. And it would be a huge pain for anyone using a localized version to do any scripting. Heck, even things like SimpleExpressions or CustomTool would not work localized.

I guess I'm just thinking that adding 0.1% more things to translate wouldn't be that much to ask, especially when you're expecting localized versions to have parity with the English one, and losing a feature like OpIconString would be noticed.

Unifying the UI with Resolve is fine in principle, but not if it ruins the Fusion UX. Hobbling UX to for the sake of branding feels like a development goal that comes from marketing, not from users.

Especially because Resolve isn't just different than Fusion in key UI areas, it's objectively worse. Take keyframing, for instance. If you were looking at Resolve 14 and Fusion 8 and saying "we're going to make a unified interface, what should we go with?", the choice to use Resolve's keyframing UI is a poor choice. The GUI in Resolve isn't bad in how it handles keyframing, it's simply working within the limits of the underlying animation features. You can't make the keyframing work better with a better GUI, but you can make the GUI suitable for the limitations that exist. Resolve did/does that. But you shouldn't carry those GUI conventions over to software that doesn't have the same underlying limitations. You should either break the convention or remove the limitation in Resolve and carry the changes in the other direction.

But you shouldn't carry those GUI conventions over to software that doesn't have the same underlying limitations.

Well put.

Jed,

You put up a screenshot of Houdini, but that's an interesting one for the localization issue, right? Task OPeraters is English, the abbreviation is TOPs, and the icon is a realistic top hat. It's all great for English users, but Task Operators in any other language isn't going to abbreviate to a word that also means "top hat".

I don't know any non-native English speaking Houdini users, but I wonder if they say "sops" and "dops" just like native English speakers do or do they have different ones that would be unrecognized by the non-localized Houdini users.

I don't know any non-native English speaking Houdini users, but I wonder if they say "sops" and "dops" just like native English speakers do or do they have different ones that would be unrecognized by the non-localized Houdini users.

I feel like there's a point at which the root language is kind of abstracted away into proper nouns -- like saying "Nike" while speaking in Chinese vs. English. It's not a Chinese word but it's also not English. Technically it's ancient Greek, but really it's just a placeholder for something both languages understand equally well.

I think the silly Houdini operator acronyms are like that too -- TOPs DOPs, POPs, CHOPs, SOPs, ROPs... when you look at the modern Houdini UI they don't even use most of the OP names to describe contexts anymore. SOPs now exist under /obj, which can be properly localized.

Fully understanding the English behind the cute acronyms doesn't really matter because the meaning is conveyed elsewhere. The real function of the CHOPs context or the 3Xf node in Fusion or whatever is better conveyed by the localized language of the thing itself, so why not just lean into it? Houdini never displays a node without the full name of it's function somewhere in the title, and if you change the name to be something else it pops up a little secondary title in the local language to make sure you will always know what it is when you look at it.

I'm a native English speaker as well so this is something I shouldn't really be spending so much time talking about without knowing the pain of failing at it -- my only reason for bringing this up was because it was the one thing I've seen posted by BMD staff, anywhere, to give context for any of the design decisions in Fusion 16 and even that small detail doesn't feel like it makes sense for the changes that are being made.

So, I'm just clutching at straws here...

Yeah, I feel like better context sensitive help and tooltips would go a long way to helping all users with any localization and English proficiency do better.

The stuff that is persistent in the UI is important, but if a tool name or parameter name doesn't make sense to a user for any reason, they should be able to get documentation for that quickly, and if it's in their native language, even better.

Icons are worthless. I prefer all text, even the short three
Enter text buttons are better than icons. SoftImage and Lightwave had it right when the decided to use text for everything. Icons for for your phone, not pro software.


That said, if you are going to use icons at lease make them distinct. Put bits of color into the icons. This all grey garbage has gone too far, and has no impact on color vision at the scale present in icons.

The nodes also need much more distinctive colors. The colors should ideally match any colors used in the icons (but don’t use icons, use text).

The text on the nodes should be bigger, and the viewer dots and inputs as well. Fu9 was pretty perfectly sized here, 16 too small.

Give us an option to disable the drag on number fields. This is so counter to every other program that it just gets in the way more than it helps.

The timeline should go back to the bottom of the entire screen.

Full file render indicators should actually be full tile render indicators! The current indicators are way to small, mostly useless.

Pleased to see progress. Miss the good old days of emailing directly with the devs. Relationships are important.

Get the Re:Viison FX guys the SDK so they can update their plugins.

Thanks.

The text on the nodes should be bigger, and the viewer dots and inputs as well. Fu9 was pretty perfectly sized here, 16 too small.

Give us an option to disable the drag on number fields. This is so counter to every other program that it just gets in the way more than it helps.

The timeline should go back to the bottom of the entire screen.

Full file render indicators should actually be full tile render indicators! The current indicators are way to small, mostly useless.

For the love of all that is good in the world, please, YES. These slider text areas are driving me insane, the sliding should be an option to enable, not disable. It breaks the most basic of UX principles and expectations. And the lack of actual full tile indicators is also putting me off.

Cheers,
Mike

The text on the nodes should be bigger, and the viewer dots and inputs as well. Fu9 was pretty perfectly sized here, 16 too small.

Give us an option to disable the drag on number fields. This is so counter to every other program that it just gets in the way more than it helps.

The timeline should go back to the bottom of the entire screen.

Full file render indicators should actually be full tile render indicators! The current indicators are way to small, mostly useless.

For the love of all that is good in the world, please, YES. These slider text areas are driving me insane, the sliding should be an option to enable, not disable. It breaks the most basic of UX principles and expectations. And the lack of actual full tile indicators is also putting me off.

Cheers,
Mike

No doubt, but people have pointed this out since the beta, and it seems to fall on deaf ears.

I hate to say it, but it is the users who must be listened to with software. Just because "Resolve does it like that," doesn't mean it's good. In fact it is silly, and since the users all object why not give them the option to turn it off? I appreciate that it was well intentioned, and some people may even like it, but the fact remains that it is driving most users nuts.

The speed improvements are great. The better use of the GPU is great. Mimicking Resolve's UI, which few people care for... not so much.

I know adding a pref to disable that feature is not difficult programatically. Hell, if Stuart did it it would take him about 10 mins, with 9 of those being updating the docs and dealing with the RCS.

I for one would LOVE to pay an upgrade fee for the next new version to have a fixed up UI (un-Resolve it) and more features. I love Fusion, and am willing to support its future. But it's gotta be for some real progress.

I know the Fusion team is capable of truly great advances if BMD will let them at it. The Planar Tracker is pretty darned good, and the 3d Tracker is improving quite a bit. I just used it in production to do an object track and was surprised how well it worked. Personally, though, I think 3d tracking generally belongs in its own app, and would rather see time spent on greater speed in compositing, deep compositing, better flare tools (think at least Knoll lens flares or better for Fusion), a saturation control for the Soft Glow node so we don't have to do it manually (or how about just a Merge Over checkbox like Highlight), better noise removal, more updates to the 3d system (meshing of point clouds, GI, perhaps Vray support, etc.).

Long live Fusion!

I just stumbled over BMDs own Fusion tutorial page.

Besides the fact, the page is labelled Fusion 16 and all contents is based on Fusion 9, I was wondering, how any potential new customer could get the slightest grip, when launching Fu16 for the first time and is faced with those videos.

Does BMD intend to redo all those tutorials ?
How many resources would be required and how long would it take to do so ?

Just a side node.
Maggi

PS: I'm still not allowed to post links, but I guess you all know, which website I'm refering to.

Just because "Resolve does it like that," doesn't mean it's good.

What's bonkers is Resolve doesn't do it like that.

Couple things with Resolve to keep in mind when it comes to the controls there...

1) Resolve doesn't like precise values in it's inputs. Routinely rounds to integers or one decimal point. This doesn't matter a lot since Resolve isn't used for technical work, nor is it used for animation where the rounding steps would show. So in Resolve, you're never worried about typing (or pasting) in a value to 5 or 6 decimal places.

2) Sliders in Resolve don't have incremental nudging. If you click on a slider, the knob goes to that exact position on the slider. So no click to add a small value, click, click, click. Likewise, there's no modifier to change the scale of changes.

3) Most controls have no sliders. More like virtual spinners. Without buttons for the spinners, Resolve uses mouse drag to set the value interactively.

4) Some controls, like Color Wheels and RGB Mixer, have no numerical input. It displays the numbers, but they're read-only.

5) You can't animate.

So what we have is something completely new.

I'd love to see Fusion get something like Resolve's spinners. Just a slider without the slider. Then the current behaviour would make a ton of sense.

5) You can't animate.

Actually, the vast majority of options in Resolve's Color page can be "animated" with keyframes. There is a Keyframe panel where you can set and manipulate keyframes for node parameters, and a toggle to have keyframes added automatically as changes are made.

5) You can't animate.

Actually, the vast majority of options in Resolve's Color page can be "animated" with keyframes. There is a Keyframe panel where you can set and manipulate keyframes for node parameters, and a toggle to have keyframes added automatically as changes are made.

Animated in quotes, indeed. The controls themselves aren't animated, the entire set of them is. You only can set keyframes on "Color Corrector", for example, en masse. And you only get preset interpolations between keyframes. And worst of all, you can't see the interpolation on the inputs.

Give us an option to disable the drag on number fields. This is so counter to every other program that it just gets in the way more than it helps.

The timeline should go back to the bottom of the entire screen.

I like virtual sliders but, like almost in any tool that employs them well, they should not be on the input field but on a label. Or if they're on the input field, they should require a modifier to activate: sometimes you want a virtual slider and sometimes you just want to click an input field. I'd rather not have to enable/disable it at a system level.

I sound like a broken record, but I really enjoy the way Houdini handles this with the step-ladder to scale your input:


As for the timeline, I actually prefer it in the middle for some workflows, but this gets to the whole point of the Fusion UI needing flexibility: this "single page for all users" idea misses the fact that a "Fusion artist" could be any of a dozen professions, each of which require a different layout of tools to work efficiently (before you even account for user & monitor preferences).

This idea isn't alien to BMD either: it's the main UX principle in Resolve. Re-configuring the UI to match a given task, page-by-page. Fusion requires even more diverse setups than a sound mixer vs. editor, so if we can't have full customization we at least need some kind of modular pane system that allows swapping & duplicating panes.

Meanwhile, yeah:

Animated in quotes, indeed.

That is a rough way to animate anything.

Virtual sliders on the label would be good, or with a modifier as you say. However, the default should be off without modifiers in text boxes. The Houdini method is nice, and would come in handy in Fusion, as sometimes I need big steps and sometimes (often) VERY small.

Impossible to read all comments but Chad and Bryan hit it on the spot. I am using F16 daily to see how much is “getting used to” . I assure you the icons are ridiculous and nobody will ever get used to them. Ask any UI designer and they will all tell you how bad the current UI partially is. The real estate is something you can get used to. Isn’t NUKE also kinda similar in that regard? Someone said simple click on the number’s field instead of dragging. Well as Bryan said, try that with a Wacom. In fact I got used to this madness and now when I use Fusion9 I have this worry that I accidentally drag values. Very very bad design and as Chad states, we enter values or expression. Otherwise we use CTRL drag for fine increments. For me another big issue is that I am left-handed using a Wacom monitor. I drag right on the monitor and currently I always have to cross the screen with my whole arm. Horrible. With that said, I still see a bright future for Fusion.


Sent from my iPhone using Tapatalk

Look under 'support' on the BMD home page. It should be there under the Fusion download section. Just scroll down (a lot).

Look under 'support' on the BMD home page. It should be there under the Fusion download section. Just scroll down (a lot).

THANK YOUUUUUU.

just take a look at the new redesigned Blender 2.8 user interface ... oh man, that was wonderful ...

It's time for BMD to think about tabs, less waste space, fewer pages, text and / or icons (good icons) user interface, and take advantage of every space you can to implement fewer steps in the workflow ...

Agree about Blender. I was pleasantly surprised when I saw their revamp.

Also agree. Too many pages. Prefer if we just had more UI configurability and some default workspace configurations shipped out of the box.

I honestly prefer separate apps to one mega app that hamstrings you when you move off the page designed for its historical primary function.

I find resolve to be a complete organizational mess for editing, because things you expect on the edit page were on pages you'd never expect them to be (like Stabilization on the Color Page - somewhat rectified in 16).

Fusion 9's configurable UI (and better dual monitor use) is the main reason I tended to use it in isolation, with Premiere Pro, instead of going all in with Resolve, which makes my screen feel 30% smaller/less dense.

Probably just an effect of the design differences, but that's how it feels to me.

Have no problems with icon. They are easier (for me) to learn than dozens of text abbreviations, IMO. One can easily use text and icons, though. But I guess that's where the localization issues comes in...

To me (a Fusion novice, but use it for simple stuff regularly), this is the most annoying thing. A row of identical icons for tabs that have completely different meaning. It used to be text, clean and simple. Now it's clean and impossible.

Screenshot 2019-04-09 at 21.48.53.png

Yaarr, happy I am not alone on this. Mentioned this in my first video about Fusion in Resolve.
I still cant get used to this Icon disaster. It gives me scurvy. Don't know what the big deal is. Please add text to the icons, very simple! Other soft like Houdini does it and guess what, it works!
Fusion is about speed and efficiency...it is just a pity if one is working so fast, then bottle necked by some bloody icons.

Vito

Not sure where it comes from either. BMD's other products all use text instead of icons. Check out the menus on their cameras and monitors. Look at the buttons on their panels and ATEMs.



This entire panel has 5 icons on it, a forward arrow, a back arrow, play, play reverse, and stop.





This has a few more icons, like volume up, volume down, command, etc., but most of those have text too, even if it's just F12 or CMD.




Two icons, up and down arrows. That's it. Oh, the back of the device has two icons as well, both are molded into the fuse covers on the power connectors. The rest of the device, front and back, is text.




Look at the screen. No icons at all. Look at the camera body. Four icons, forward, reverse, play, and power. That's it. Rest is text.

Interesting observation Chad. That really makes one wonder where those Icons come from. One thing about hardware though is that icons could probably be harder to print and more expensive?

A big step if they want to keep the icons (other than the correct answer, which is all icons suck, use text) would be to color code them, like the node in Fu9 are colors code, light blue for 3d, green for loaders, pink for particles, etc.

But make these colors fairly prominent, not 4 pixels at the tip of a magic wand, but rather 25% of the icon should this color, ideally the background or outline or something so the color is in a consistent location.

I assure you tiny bits of color of icons are not affecting our color perception because we are only focused on them for an instant.

Usability is far more important than trendy monochromatic interfaces, and this is coming from a guy with neutral grey walls, daylight balanced lighting (if it’s even ever on), etc.

One thing about hardware though is that icons could probably be harder to print and more expensive?

Actually less, if it means you don't have to print localized versions. It's also cheaper to have one SKU for everything instead of localized versions.

So there must be some other reason why the vast majority of BMD products use text almost exclusively. I suspect that it is because even if you can't read English, it's easier to understand/remember a short sequence of latin characters than it is to understand/remember a set of new icons that you've never encountered before.

If you don't know a function, you can easily just transfer the sequence of latin characters from the device to a latin keyboard and search the localized PDF docs to get an explanation of what the characters mean.

I don't speak Dutch, but I can "read" the word fluyt, and even if I don't know what it means, I can look it up quickly and learn, whereas an icon of a fluyt would be completely indecipherable to me. I might recognize it as a sailing vessel, but if an icon of a schooner was next to it, I wouldn't be able to tell the difference between them, but more importantly, I would have no way to figure out the difference from the documentation.

I can't add much to the conversation besides that Fu16 is downright unusable.
The whole UI is 1 step forward and 20 steps back.
The number inputs are not usable with a tablet, at all, and every Fusion user I know uses a tablet.
The icons are useless, for any user, regardless of language barrier.

We use Houdini as non native english speakers and every user I know knows what 3/4 letter word operator does what, without fully grasping what is's an abbreviation for, just the context.

What also is missing since 8/9 is the ability to customize the skin, we are not always doing color critical work (creating textures, icons, etc for use in realtime applications) and work in a lighter environment than a film compositor would, so switching to a lighter / higher contrast skin helps a lot. With Fu16, if you are working in anything but a basement, the UI is downright illegible.

The whole UX and UI design seems likely based on a heavy dose of cognitive bias and not knowing your actual user base.

The whole localization argument I can't argue much on besides 'what localization?'

The fact they keep removing feature after feature while stacking style over substance features is really really grinding my gears.

This whole Fu16 update feels like a massive case of malicious compliance. Maybe they make it so bad that in a year they can discontinue Fu standalone and when we complain they can point a finger at their analytics and say 'see, we gave you a new version and no one is using it, ergo, Fu for Resolve is the way to go.'

I do not like the new interface too.
I can not imagine working with that.

You all have positive points in the discussion, but do you think Blackmagic is listening?
No sarcasm here, just to be confident about the future.

I believe they are listening.

A few simple things like going back to text buttons or at least making very contrasts, distinctive icons and increasing legibility of things would be easy to implement and be a big step in the right direction.

Hell, just increase the contrast of all the icons would be a big step (photoshop script, anyone?)

They’ve done a lot under the hood, and the new updates are much nicer and faster. It’s just the UI that needs some cleanup.

Or just go back to the Fusion 9 interface and move forward from there.

Or just go back to the Fusion 9 interface and move forward from there.

From us users' perspective this would be THE solution. Highly doubt that will happen, but just for fun, what would people want to improve in the F9 interface? That list is inevitably a lot shorter than the one for f16...

Or just go back to the Fusion 9 interface and move forward from there.

• From us users' perspective this would be THE solution.
• Highly doubt that will happen,

What's wrong with this picture?

what would people want to improve in the F9 interface?

I like that notion.

Nuke-like up- and down-arrow manipulation of the numeric entry box: Put the cursor behind a number, and you can change just that number.

Customizable node shapes, allowing greater semantic meaning to tiles in the Flow.

Indicators, visible in the Flow, of animated properties (present in v16 and in Resolve; thanks!), expressions, and concatenation.

Syntax highlighting (there in 16's CustomTool, but not quite nailed down as well as could be desired) for expressions, Custom Tools, and the Console.

Drop-and-replace nodes on the flow.

Publish control to Flow View to permit editing values right on the node. Suppose you have a CC (dozens of controls), but you're only manipulating the Gain, or maybe Red, Green and Blue Gains. You could right-click the control, choose "Show in Flow" and get a slightly larger node with numeric entry fields on the tile. Maybe that would be a good place for scrubable numeric fields (totally senseless to have them if there's a slider right next to the field—it's great for point controls, but that's the only place that actually needs them). Node could have a minimize button to show just the name to prevent accidentally changing things when you're just moving it around in the graph.

Modifiers available as nodes in the Flow. Maybe with darkened wires leading to nodes connected to them.

The ability to tear off/pop out comp tabs so two comps can be more easily compared side-by-side.

Improved custom menus (this is a very niche case). I have a script that I like to run directly from the menu bar. I hacked it onto there, but for some reason the text is tiny:

Untitled.png (3 KiB) Viewed 76152 times

It really shouldn't work at all, but it does after a fashion, so I'm keeping it.

The Edit Hotkeys dialog could be improved. It's a little bewildering. I have to relearn it every time I want to use it, and most of the time I just make a .fu file by hand instead.

An Edit Macro interface, where running it on an existing macro pre-loads all of the controls currently on that tool so that they could be tweaked a little instead of having to completely start from scratch. Again, I almost never use the Create Macro dialog because it's far easier and faster to just write it in Notepad++. It could be much more accessible if it had options for rearranging and renaming controls right in the editor.

Customizable rulers and guides in the Viewer, similar to how Adobe products work. Also, snap to guides, with a toggle button in the Viewer window.

More options for UI Scaling. 100 or 200% is not enough choices.

An option to keep Bins on top of the main window (UI Manager scripts do this; why don't the Bins?), and restore drag-and-drop into the Bins.

The Fusion 9 interface definitely has plenty of room for improvement, but we'd really like to see it move forward, not backward and a little to one side. I don't care about whether or not nodes have borders or rounded corners. I care a lot about being able to read and manipulate them!

Of course, many of these wishlist items could apply also to Fusion 16's current UI.

I agree with all of Bryan's suggestions. And as an aside -

"The ability to tear off/pop out comp tabs so two comps can be more easily compared side-by-side."

- was, I believe, present until Fu8, when it was downgraded to the present system. Little did we know then that this was the thin end of the wedge . . .

I agree with all of Bryan's suggestions. And as an aside -

"The ability to tear off/pop out comp tabs so two comps can be more easily compared side-by-side."

- was, I believe, present until Fu8, when it was downgraded to the present system. Little did we know then that this was the thin end of the wedge . . .

I believe that was a consequence of shifting the interface to Qt. That is, it was an engineering-driven change, not a UX-driven one. And since it came along with UI Manager, which is so much better than IUP, I'm fully willing to have given up side-by-side comps and that script debugging window (I'd also like that back, but I'm not sure if it's strictly a UI request?)

That said, I don't think it's impossible to do in Qt, but it might be more complex than it appears on the surface.

Totally agree with what's being said here.
Lot's of improvement in v16, feels faster, but few UI issues disappoint.
1. Icons instead of names of tab in inspector that all look the same - disaster
2. Too simplified icon set of the viewers, b/w icon for color channel.
3. I really miss "fit" button on the panel. Yes, with more clicks I can do it, but it was so superb to fit image with one click.

Another thing not related to UI is that I dream about porting ResolveFX to Fusion standalone. It's so nice to have them in Fusion of Resolve, but for single shots it's much more fluid to work in standalone. Loader/Saver limitations in Resolve suck.

But overall it feels like a great upgrade,
thanks Blackmagic

All you got to do is this to see what's wrong and what should be done

I really wish I could spend more time in the Beta but I have some tough deadlines and I am finding it all to slow with the current UI to get my work done. Which is a shame as the actual processing under the hood feels a lot faster (and probably more stable) but at the end of the day the UI issues outweigh this.

Hi,
installed Fusion Studio 16b and am not ok at all with this UI!
How can I get the normal skin as in Fusion 9?

Hi,
installed Fusion Studio 16b and am not ok at all with this UI!
How can I get the normal skin as in Fusion 9?

Sadly, I think the only answer right now is to use Fusion 9.

Hi Frank,
thank you. But this UI is way too inconvenient for work!
We need normal Fusion 9 skin.

As it is only beta of F16, its only test UI, or they think to leave it in official version?

Unless enough people provide enough concrete examples of changes to improve the new UI it is likely destined to end up as the final.

Something tells me they are not going back to the old one at this point. A lot of work went into the new UI. However, there wasn’t much communication with users during the design phase. I think it can be salvageable with some of things mentioned in this thread. But these things have to be done before this gets out of beta. I worry there will be many, many defectors if users try to use this new interface in production.

Another thing that need fixing: right click -> Edit on an animated parameter needs auto open and display the Spline Editor... and this should be at the same size and position it was last used, not the default.

The more I try to use Fu16 in production the more I concur with others (esp Chad) who want to get rid of all the silly pop up menus and get the multi-buttons back. Those pop ups are slow, and don’t present the options to the user all at once for quick access and quick identification.

Here is another change in Fu16 that destroys usability... easy to fix too.

In Fu9 when a spline is animated you get a blue or green marker next to the Right click here for shape animation line. Now you get nothing. So there is no way to tell if a spline is animated.

This may have been mentioned before, but it’s a big one and an easy one to rectify. Thanks.

U less enough people provide enough concrete examples of changes to improve the new UI it is likely destined to end up as the final.

We have. Lots of examples, including text and images.

Something tells me they are not going back to the old one at this point.

We're not expecting them to, we're just pointing out that in nearly every actual use case, the new UI is regressive relative to the existing Qt-based Fusion 9.

If you change 100 things between Fusion 9 and 16 but 95 of them are negative, the choice is whether you should fix 95 things or 5. The labor isn't just in the core UI design, either, it's in plugins, fuses, macros, documentation, tutorials, etc.. And there's still a ton of labor to get the Fusion 16 UI functioning properly even if we thought the design was good. It's buggy in ways that Fusion 9 isn't.

A lot of work went into the new UI.

A lot of work went into the Fusion 8/9 UI, too.

A lot more work remains to get the new UI functional. They're a year and a half into sunk costs, sure, but there's still a ton of work remaining.

However, there wasn’t much communication with users during the design phase.

But there has been a lot of feedback from users on the new UI before the Fusion 16 beta. A lot of the issues raised here were also raised during the Resolve 15 beta, we were just told then that it was transitional. I would feel a lot better if between the Resolve 15 beta and Resolve 16 beta there were more improvements to the UI, but there haven't been, nor has there been much during the Resolve 16 beta. It's like it's been nearly locked for a year and a half. The missing comment icon? The non-functional menu buttons? Been there since Resolve 15. The color picker changed, yes, but it's still not as good as the Fusion 9 one.

I with you, Chad. Trying to encourage MORE users to add MORE ideas and thoughts on the new UI so that BM realizes the importance.

If there is some way for them to easily keep the under the hood changes, but revert the UI to Fu9 trust me, I am all in favor of this approach. Fu9 was not perfect, but it was useable. However, I find when software is acquired the new owners typically want to put their mark on it, like a dog with a fire hydrant.

This change for change sake is rarely productive and typically destructive. We see it in advertising all the time. New VP in company decides he wants to mix things up to justify his existence and hires a new agency that ruins a great campaign.

I also find most managers are too proud to admit their mistakes and go back to the old version/method.

I think the longer and more detail ridden this thread is the more likely BM is to listen.

I don’t mess with Resolve. So I was not involved in that side of things. Sad to hear the UI has been this messed up for so long.

The longer this thread stays at the top of the forum the better. I hate to see all the hard work of the eyeon team dragged down by the new UI.

I don’t mess with Resolve. So I was not involved in that side of things. Sad to hear the UI has been this messed up for so long.

That's the catch. The UI in Resolve was wholly unsuited for production compositing work, so most of us didn't use it and stayed with Fusion 9. We were told it was a WIP and just let it be. There were issues with Resolve that made it unsuitable aside the UI (metadata, format support, plugins, project files, etc), so we just waited for the standalone to get updated with the new GPU enhancements. We didn't at the time realize that the UI would also come over. Especially since it was still a WIP on the Resolve side. I assumed it would be siloed. I don't understand why they would simultaneously change the underlying pixel processing code, which means there's a HUGE amount of testing to do, including breaking all the plugins and OCL fuses, along with the UI, which means it's incredibly difficult to test the changes to the pixel processing code.

Heck, we're so busy muddling through the new UI, none of us are commenting about the new tools.

And I'm not saying the new UI is all bad, either. I think the new hidden inputs is nice, and other than it not being per-input (and thus not applicable to user controls), I'm happy with that new addition. I also like the idea of moving the passthrough toggle to a more prominent space on the tool controls. I don't like that the pin got moved to a harder-to-select location, but I don't have a suggestion about how to keep them both prominent. I like that the loop/ping-pong button toggles on and off instead of trimode cycling through off/loop/ping-pong. That's a personal preference thing, though, and I can see how other users might not like having to right click to change the mode.