Fixing Photoshop
Photoshop, my love
I’ve been a Photoshop enthusiast for nearly a decade now. When I made myself familiar with Adobe Photoshop CS2, its features were so overwhelming that I knew there was no going back to my then-treasured Macromedia Fireworks. Photoshop was simply so powerful, robust and stable. Since then, I’ve seen its steady evolution, for better or worse; during my school years I saw the advents of CS3, 4, and 5, and when I stepped into the world of graphic design as a profession, my trusty steeds of choice were CS6 and eventually CC.
Promises were big nearing each new release: New tools, content-aware image manipulation, video editing, 3D extrusion, 3D brushes, 3D printing even. Though delivery was made, every iteration was practically, in its core, the same. It seemed as if Adobe’s primary goal was to put the new, fancy features on a pedestal to lure customers into upgrading to a marginally better program. Yet here we are, some 25 years later, fooling around with our puppet warps while locked to a fixed keyboard increment. Spicing up our 3D models while unable to distribute space properly.
The point I’m trying to illustrate, devoid of all malice, is that fresh Photoshop releases let new features overshadow the old, foundational, yet broken ones.
Someone smart once said: ”It’s felt more and more like paying for bug fixes and unnecessary features.”
Since I began writing this article, there have been two additional releases, CC 2014 and CC 2015, neither offering anything game-changing.
I’ll try my best not to only list every little thing that has irritated me through the years, but also to propose a solution. I don’t know all the technical details and I don’t care. Buckle up, there’s a lot to cover.
Problem #1
Alignment woes.
For some reason, Photoshop has only six object align functions and six object distribution functions to offer. This is a problem, as seen in this example scenario:
If I need to have the same spacing between different-sized objects, I’m in trouble. Both Illustrator and InDesign have two additional object distribution functions for Vertical and Horizontal Distribute Space, respectively, all of which can be combined with the ‘align to…’-variable.
Why Photoshop lacks these features is beyond my understanding.
Illustrator and InDesign are programs that operate in vector space, but that probably isn’t the issue here. I have no choice but to use Trevor Morris’ scripts, that do work well, but are still a bad excuse not to have them in Photoshop’s core.
Solution #1
It’s literally right here. Swiped from Illustrator.
I’m aware that Photoshop has no concept of “Artboard” (up until CC 2015) or “Key Object”, so to fit the paradigm it would be logical to change those to “Canvas” and “Top layer.”
Problem #2
Increment inconsistency.
If you change a value with the arrow keys, the amount is increased tenfold if you hold Shift. This applies to color value fields, font sizes, adjustment layer values… to my knowledge, every user-editable numerical field. It might seem obvious that moving a layer with an arrow key raises its position by one pixel, ten if shift is held. This, however, is an illusion. The amount of pixels a layer moves to a given direction upon pressing an arrow key depends on your zoom level, unless you’re holding shift, in which case the layer always moves 10 pixels.
This is problematic because it gives a false sense of control. Just about every program in the Creative Suite follows a set of rules on how the modifier keys function when used with tools. Why the sudden exception?
This can be averted with scripts as well, but having to start an action or a script every time I want to move something just doesn’t seem right. After Effects has this all figured out: the amount that an object moves while Shift is held is always ten times the amount it normally moves, depending on your zoom level.
Solution #2
Add this into the options panels of the Move tool and Path selection.
Or just put it on the general preferences window. Like in Illustrator. Additionally, make holding Shift correspond to this.
Problem #3
Displacement map vagueness.
Displacement maps can prove to be extremely useful if you’re manipulating photos, making cool effects or even visualising data, yet actually using the filter can feel very tedious. There are three main reasons why:
- Official Photoshop Online help has only a very short article describing displacement maps. It is literally the same explanation that can be found in Adobe Photoshop CS2's manual.
- The dialog is counter-intuitive. Almost every other tool uses a tabbed panel layout with an immediate preview, but the Displace window has two radio buttons and two input fields with no explanation. There aren’t even tooltips.
- Lack of preview. Using a displacement map requires you to go on for several steps, blind, before you see the changes you made. Even if you only want to make minor edits to the displacement, you must go through the same frustrating steps all over again.
This triad of troubles is likely to drive users (especially new ones) away from using the tool despite its versatility.
Solution #3
Let’s review the old way:
Very tedious! But since displacement maps are grayscale in the first place and have only a few properties, why couldn’t they be treated as adjustment layers? Here’s a mockup of a new method.
The fact that the changes are instant would surely encourage users to play around more and learn more. I could imagine someone turning down this idea for being too processor-intensive, but, you should remember that Photoshop CS6 gave us a GPU-accelerated Liquify, which increased its performance a great deal. If the idea is not viable, the feature could at least be disabled for low-end machines.
The old Filter > Distort > Displace method should still be available for Smart Filter compliance, but having the chance of using it like an adjustment layer too would open a whole lot of new doors.
Similar problems persist with a couple of other Filters too, but more on that in problem #8.
Problem #4
Perspective Warp dumb-down.
I’m very aware that the perspective warp is a new tool and possibly subject to change, but let this serve as an example of a great new tool being introduced while still lacking the basic behaviour that every Photoshop tool should have.
When you begin setting the perspective warp mesh up, modifier keys stop functioning completely with the exception of Shift when dragging an edge of a net. Again, this, like problem #2 is a case of giving a false sense of control. You’re using a tool that looks like something familiar but acts like something totally different.
Other peculiar design choices are the spheres for corners and three-pixel wide edges. The spheres are so big and the lines are so thick that they sometimes obstruct the underlying image. Case in point, why the sudden exception? Hitting caps lock won’t do anything either, even though it gives you “precise cursors” with some tools.
I had to manually reverse-engineer the PSIconsLowRes.dat binary file inside the Application package, locate and overwrite the image data for the corner spheres just so I could change them to something nicer. Apparently, the icons stored in the binary are 24×24 pixels, but are scaled to 22×22 pixels by the program, so making them as sharp and simple as every other cursor is impossible.
Solution #4
Change the behaviour of the tool to work as shown on the left. The design should also be something between the basic Transform Controls and the classic Warp tool, not only for consistency but also that you could actually see what you’re doing.
It wouldn’t hurt to borrow some features from Affinity Photo’s Perspective Warp tool either, namely the choice between using the perspective net as your source or destination. The tool has so much potential but it’s now gimped down to just a nifty bonus. Also, the final result’s interpolation is always Bicubic, even if you’ve set it to Nearest-neighbour in the preferences — more on this in Problem #6.
Problem #5
Radial Blur is outdated.
Here’s the radial blur dialog in Photoshop.
I really shouldn’t have to tell you that this is completely unacceptable, given that it is practically impossible to align the reference point accurately. The so-called preview is a 128×128 monochrome image, and there’s no preview checkbox, so you really have no idea what you’re doing.
This can be averted by using Smart Objects, but even then the alignment is a chore. Not to mention that the quality options are as vague as possible. The amount field can only go up to 100 units (pixels? percent? nobody knows..) and the Spin-option is ultimately useless because Blur Gallery is a thing now.
Solution #5
Simply add the Zoom option to the Blur Gallery. This should be a no-brainer. There is no excuse to leave it to this decades-old dialog. In fact, remove the Radial Blur option entirely. It really serves zero purpose.
Technically the characteristics of a radial Zoom blur could be achieved by adding a few additional options to the Iris Blur, but there’s really no excuse here.
Problem #6
Interpolation lies.
If your Image interpolation has been set to Bicubic Automatic in the general Preferences, you have every reason to believe that that is your Image Interpolation algorithm, unless otherwise stated. However, straightening a photo gives you this, even though you just set it to Bicubic Automatic.
Where do the rough edges come from? As it turns out, there are two separate Image interpolation drop-down menus in Photoshop. One in the General Preferences and one in the upmost Options-panel whenever you’re transforming something. They have the same options but they are not linked. This can be very misleading and cause a lot of unnecessary trouble, especially if you’re working with big images, zoomed out, unaware of the rough edges.
So to avert this you need to first begin transforming something, then change your Image interpolation from the Options-panel, apply the transformation (even if you didn’t do anything) and finally you’re good to go.
It is worth noting that this drop-down menu is only 64 pixels wide, so the only words that fit inside it without being truncated are “Bicubic” and “Bilinear.”
Solution #6
This is a tough one. What if you really do want to have separate Interpolation methods for different tasks? I know I do, but sometimes it can get frustrating and other times I don’t have a choice. And what about other Interpolation methods? Sometimes I want to use hqx, but adding it to the list isn’t a universal solution because it’s only a scaling algorithm and works only with integer factors.
I propose a few overhauls. Not only should the Preferences panel get a new sub-category, but an “Interpolation method” -option should also be present in every dialog and in every options panel where actual interpolation is about to happen.
Also, especially in the case of Smart Objects and Distort Filters, being able to actually change the interpolation method afterwards would help a lot.
Problem #7
“Smart” Objects.
Smart Objects are great. “Things that you can edit later and won’t lose any quality even if you twist and turn them” is something people usually think about when the subject is Smart Objects. While true, there are actually a few other “Smart Object-like” objects and they each have subtle differences. A reform is needed.
There are three different kinds of “Smart Objects.”
Traditional Smart Objects
- Can be fully distorted.
- Bounding box orientation is preserved after the transformation is applied.
- Predefined and custom mesh warps work.
- Filters can be applied and are editable as Smart Filters.
Vector Smart Objects.
- Can only be scaled, rotated and skewed.
- Bounding box orientation is preserved after the transformation is applied.
- Mesh warps do not work at all.
- Filters can be applied and are editable as Smart Filters.
Text Objects.
- Can only be scaled, rotated and skewed.
- Bounding box orientation is not preserved after the transformation is applied.
- All numerical values (rotation, skew, scale) are reset after the transformation is applied.
- Predefined mesh warps work, custom mesh warps don’t.
- Filters can be applied but they automatically rasterise the text.
- The interpolation method drop-down menu does absolutely nothing.
- While not really Smart Objects per se, they should still be taken into consideration because of the ease of re-editing them after applying typical transformations.
Because the Traditional Smart Object is the only full-feature one, you’re forced to do convoluted Text-to-Smart Object conversions even for simple perspective tricks.
You lose the ability to directly edit the text completely and you must open the Smart Object in a new window to fix your text there, and even then you’re stuck with the original dimensions your text was in. You can resize the canvas of the Smart Object, but you’re doing it blindly in a separate document.
All this trouble for such a simple task.
And even then, you’re out of luck. Once a given corner goes too far, something weird happens.
And what about after you realise you’ve made a mistake? You can’t reset any Smart Objects, and you’re especially screwed with text objects…
…Because resetting a transformation of a Smart Object is impossible.
If you’ve got a Smart Object you’ve distorted, scaled, skewed and rotated, the only numerical fields you can access are the width and height percentages, rotation and warp type and you can manually set those back to 100%, 0° and “none”, but that’s it. With text objects you can’t do anything except change the warp type.
Solution #7
First off, unify the Smart Objects. Give all the currently Smart Object-like objects the same properties as true Smart Objects.
Finally, show all the fields you’re manipulating as numerical, editable text boxes, add a big “reset transformation” button and an “Interpolation method of this particular layer” dropdown menu into the Free Transform’s options panel.
Also — since version CS6, videos have also been treated as standard layers. Although transforming a video requires the additional step of converting them into a smart object, the video’s timeline and audio track still remain intact. I can warp, distort and preserve my bounding box even with large videos, but I have to do the arbitrary task of hitting the “Convert” button in the “Transforming a video layer requires converting it to a smart object layer” dialog. To my knowledge, this doesn’t lose any of the video’s original properties, so unless there’s a good reason not to treat Video layers as Smart Objects by default in the first place, why would converting be necessary?
Problem #8
“Smart” Filters.
We’ve had Smart Filters since Photoshop CS3, but they still have a few flaws. The most annoying of them is that the transformations are applied first, then the filters. You can avert this by going back and forth inside a Smart Object inside another Smart Object and back again.
I’ve made a 512 × 512 Smart Object. Can you guess what happens when I apply an Offset Filter with +128/+128 values?
This is not the case only with the Offset filter, but with every other filter as well.
The transformation/warp is applied first, then the filters. So, there isn’t really a way to apply a Mosaic Filter (for example) with a block orientation along the distorted perspective.
And even so, what if I wanted a row of Fibers that go along the distorted axis but a noise filter that doesn’t?
And what about masks? You can only apply one mask to a set of Smart Filters. You can also set individual opacity percentages or blending modes for Smart Filters but you have to do it one by one.
Solution #8
Instead of a total overhaul, just improving upon the current one would probably be the best way to approach the issue. First, perhaps a slider to give quick access to individual Smart Filters’ applying method.
Great, so what about individual mask controls? Well, having the chance to actually select (multiple) Smart Filters would help. In the mock-up below I’ve selected the Radial Blur filter and given it a bitmap mask from the Add Layer Mask-button in the bottom.
Finally, simply by giving the user the chance of Smart Filter multi-selection, the Blending Mode and Transparency problem is solved without any extra effort; simply use the already existing Blending Mode drop-down menu and Opacity Slider to set the values. With a set-up like this, the original Filter Blending Options button is no longer needed.
Adding a “Copy/Paste/Clear Smart Filter(s)” to the right-click menu would be a logical choice too, as the Layer Style menu already has them.
Problem #9
Frequency Separation.
Blur, Shake Reduction and Masking are essential, very common tasks in Photo Retouching and it is logical for Adobe to include powerful tools built from the ground up for these tasks. CS6 gave us Blur Gallery, CC brought in Shake Reduction and CC 2014 introduced the Focus Mask.
Sadly, one common yet extremely powerful task, Frequency Separation is still missing from Photoshop as a dedicated tool. Affinity Photo has a great action for this, but even that has some issues.
The gist of it is this; you separate your image into two frequencies, the high frequency (the sharper details like hair, skin follicles, etc.) and the low frequency (a blurred mash of skin tones and other colors) that together make the original image. The advantage here is that you can easily fix the underlying colors but still keep the fine details intact, and vice-versa.
This is great, and can give fantastic results quickly. But no matter which frequency you’re editing, at the end of the day, you’re still destroying the image. Both of the Frequencies can be achieved by using Smart Filters but after that you’re stuck. Using the Clone Stamp or the Healing Brush will rasterise the image and there’s no going back.
Solution #9
Like the Blur Gallery and Shake Reduction before, a new, well thought-out feature should be considered. What makes Liquify so special is the fact that despite being a Smart Filter, it can alter the underlying Image in practically infinite amount of ways, as opposed to the traditional Smart Filters like Gaussian Blur or Unsharp Mask, where you’re limited by the sliders/given values. Liquify, however, gives you a GPU-accelerated playground where you can smudge and smash your image as much as you want, losslessly.
The same approach should be taken with Frequency separation, and I’ve made a quick mock-up below.
Something along these lines — build a separate workspace to concentrate into, have the typical tools you’d use in Frequency Separation, and output everything without losing quality.
The true power of Liquify lies in the fact that you won’t lose any data when you use it as a Smart Filter. Frequency Separation as a dedicated tool has some pretty deep implications and could very well be a game-changer.
Problem #10
Render filters.
Adobe Photoshop CC 2014 brought three new render filters to the table. Flame, Picture Frame and Tree. Each one of them is unresponsive, can not be treated as Smart Objects, their UI is a mess and your color sampling options are limited.
Let’s start with the Tree filter.
Tree
After two seconds of waiting, I was able to select a custom color for the leaves through the System color picker. You have no option to change this. There is no live preview, so if you think that a particular color is fine, you have to close the color picker, wait for your preview to change and then see what it looks like. Three clicks and three seconds. To see a color get changed.
There’s no undo either, so when you choose a color you might like, better write it down.
The settings in the dialog are confusing and limited. You have 34 Base Tree types and 16 Leaves types. You can’t create or edit the base trees nor leaves yourself by traditional means.
The controls give no context, and you’re limited to arbitrary ranges. Light direction ranges from 0 to 180, Leaves amount is 0–100, Leaves Size is 0–200, Branches Height is 70–300, Branch thickness is 0–200 and Arrangement is 0–100. They are all sliders and lack a unit. None of the textbox values can be set with the arrow keys (optionally combined with the shift key) either, but the sliders do a jerky motion if you select one (you get no indicator anyway.)
Finally, your preview is a small window. So, as with problem #5, you have no idea what you’re doing. Because these filters can’t be used as Smart Filters, you’re ultimately stuck with what you have.
Flame
The same problems are present with the Flame filter as with the Tree Filter. The values are all sliders with arbitrary ranges, everything is slow and your color picker is always the system picker. Your preview window is a tiny box with a black background so if you want to have black flames, you will have to do it blind.
The flame filter requires you to make a Work Path, so one could think that once you’ve hit “OK”, you would still be able to edit the work path and the flame would move with it. This isn’t the case. As with the Tree Filter, you’re stuck with what you accepted.
Picture Frame
Most of the same problems are present with the Picture Frame Filter, too. The user interface is a mess, the units are given without context, your preview is a small window, everything is slow and once changes are committed you’re stuck with what you have. The margin and the scale sliders can only be set so high, and the final output doesn’t always match with what you see in the preview, not to mention the “Arrangement” slider (1–200) is a set of 200 different types of borders, so if you’re looking for a specific one, you must go through every single value. You’d think you can draw a curve like in the Bevel setting in the Blending Options menu, but no.
Additionally, the preview window in all of these filters is always an RGB-parsed image. If you’re in any other color mode, the preview is misleading.
Solution #10
Before I propose a solution, I must emphasise that these features are subpar at best. The program has shipped with these three filters, made by the people at Adobe, marketed as official new features, but upon closer inspection they don’t seem to be more than an afterthought. The reason behind the strange UI, unresponsiveness and raster-only output is that these aren’t true features, coded, compiled and incorporated to the Photoshop’s core, but in fact JSX scripts in the Photoshop > Presets > Deco folder.
I’m speechless. But as I was saying, the solution…
First off, re-program the filters from scratch.
From what I gather, Photoshop is mostly done with C++. Having some things open is great and the scripts that have shipped with the program since 2003 are nice, little things. People have made wonderful stuff with scripts, and some are well worth the money. These three filters aren’t.
These are official features made by the same team that has access to Photoshop’s source code. Either do it properly or don’t do it at all. Why not put the best programmers on the job to guarantee the features to be as robust as possible and truly integrated within the program? The usual stuff: being able to edit the Objects afterwards, being able to sample colors properly, having a “preview” button… I don’t think the same level of versatility and speed can be achieved with mere scripts.
Second, re-design the User Interface.
This should be obvious. I’m sure this mock-up I made is far more intuitive than the current Tree filter.
Have all the same settings in there, add some context to the sliders, have a consistent UI, output as a Tree Smart Object that upon double-clicking opens this same dialog. The same overhaul for the Flame and Picture Frame filters is clearly needed, too. Obviously, add a “custom” option to every drop-down menu as well. Now I’m just stuck with the leaves and branches that Photoshop provides.
Lastly, don’t do this again.
I don’t want to sound harsh, but these filters are against everything that Photoshop used to stand for.
Problem #11
Blending modes.
All Blending Modes aren’t always necessarily available.
There are 29 Blending Modes in Photoshop. After Effects has 38.
- Depending on your Document Color Mode, what you’re doing (painting, handling layers, etc.), your Bits per Channel and a handful of other factors — you’re only allowed to use some of them. See the chart below.
Similar problems persist with Adjustment Layers, but more on that in Fundamental Problems. - You can only use the Blending Modes that ship with Photoshop — if you want custom Blending Modes you’ll have to resort to Plug-ins.
- Previewing Blending Modes always requires you to commit the action, whether it’s releasing your mouse button on the Layer Blending Mode drop-down menu or starting to paint with the Brush tool. You can cycle through blending modes by holding Shift and pressing + or -, but a quick preview would be a real timesaver.
Here’s what’s available right now:
Solution #11
First, unify the Blending Modes wherever possible. Obviously Color won’t make much sense in Grayscale Mode, but there’s no reason not to have Difference in Lab Mode. Adding at least Clear to the Layers Blending Modes would be beneficial, too.
Fixing the 29-Blending Mode limit could be averted by allowing user-defined Blending Modes. After all, they’re just color math, and even if the average user couldn’t code one’s own, experts could distribute their custom Blending Modes and not break compatibility. The current situation allows only those with the appropriate Plug-ins to use additional Blending Modes.
On the subject of previewing the Blending Modes, having an (optional) on-hover sample when using the Brush tool with Blending Modes, not unlike the preview you see when using the Clone Stamp tool (and recently, the Healing Brush tool) would save a lot of headache. The way that Affinity Photo handles Layer Blending Modes is clever, and Photoshop could learn a thing or two from it.
Fundamental problems
Flaws that reach deep.
These are the things that nobody really complains about, but I think these points should be addressed not only because ignoring them will give the impression of incompetency, but fixing the issues might quite possibly lead to better code quality and to a much more versatile product overall.
1. Most things have a limit
But why? Is there a reason besides “it’s always been like that” for these seemingly random limits? In some cases like Fill and Opacity it’s obvious that 0–100% is the only option, but the more you think about it, the more you start questioning the limits. For example, the Gaussian Blur is 0,1–1000 pixels. Lens Blur is 0–100 something. Polar Coordinates is always 100%. Add Noise is 0,10–400%. Vibrance is from -100 to +100 somethings.
It’s a mixed bag and there really aren’t any guidelines or clear explanations on what the filters truly do. I know Lens Blur does a pretty nice Lens Blur effect if the Amount slider is at 25 and a stronger one if it’s at 50, but for what reason? What is the 25 and why is it like that? Why is 100 the maximum, and why does Median max out at 500 pixels? Are the sliders linear or logarithmic?
There probably isn’t a single real solution to this, but at least After Effects lets most values go way overboard and MAXON’s Cinema 4D has a unique take on sliders. Maybe the solution lies somewhere in between.
The perfect scenario would be total control over everything. Everything in Photoshop has either a finite slider limit or no choice at all; you can only apply Diffuse, you can only apply Despeckle, you can only set Polar Coordinates. You have no control over how much.
I was delighted to hear how Photoshop CC 2015 lets me apply (among others) multiple Drop Shadows. But then again, “multiple” in this context means “up to ten.” Such a shame.
Okay, maybe the limits are there for a reason, and setting a Zoom Blur with a value of 90 000 might be a bad idea. But the least the Photoshop developers can do is go through each user-editable value and ask “Is this limit enough in this day and age? Does this slider and its implied units give enough context to the user? Are there tooltips and a preview checkbox?”
Funnily enough, even the Large Document Format (.PSB) format’s maximum bounds, 300 000 × 300 000 pixels are based on an arbitrary limit. According to the spec, the width and height are four-byte values, so technically you should be able to have 4 294 967 295² pixel image. But that’s a minor issue, I just wanted to point that out.
2. Most things are inconsistent.
I foreshadowed this issue in Problem #11. Blending Modes aren’t the only thing that were designed with only 8-bit RGB in mind. Here’s another chart regarding Adjustments (not necessarily Adjustment Layers, because you can’t have Layers in certain modes).
Again, the Black & White Adjustment Layer doesn’t make much sense if you’re already in Grayscale mode, but being able to Vibrance my Lab document would make sense.
The odd one out here is Shadows/Highlights — it can be used as a Smart Filter, despite being in the Adjustments menu. Wouldn’t it make more sense to make it behave as an Adjustment Layer? (To be fair, the same confusion is present with Puppet Warp and Perspective Warp; they’re also technically Smart Filters despite being in the Edit menu)
Your set of Filters get more and more limited the further you go in Bits/Channel or Color Modes, too.
The last five adjustments, HDR Toning, Desaturate, Match Color, Replace Color and Equalize are one-shot only. You commit the changes or don’t — you can’t combine these with Smart Objects at all.
That alone is one of the most irritating flaws Photoshop has, and this is especially evident to anyone who has used After Effects. Photoshop has these destructive functions here and there, and actively encourages you to break the image data. Smart Objects/Filters are usually the secondary method. Even the Eraser tool still exists. And in the end, your undo stack is singular, linear and limited.
You can set the Opacity, Fill, Layer Styles and Blending Modes of multiple layers at once. But you can’t apply most of the destructive things. If I want to desaturate twenty layers, I will have to do it one by one.
You can set the feather amount in the Lasso selection tool but not in the Patch tool.
Like in the limit problem, there probably isn’t a single all-around solution for this, but again, here are a few questions the developers should ask themselves:
- Is it justified not to include this feature in this color mode/depth, even if the average user doesn’t benefit from it?
- Could we turn this feature into a “Smart” one, giving the user the chance to edit its values afterwards, even if the feature was previously just a single command (i.e. Desaturate)?
- Can we add add sliders/checkboxes/fields to Filters/Adjustments that are missing them (i.e. “Repeat Edge Pixels” to Gaussian Blur or “Feather” to Patch tool)?
- Does the dialog have its own Undo stack? Why not?
- Can this action be applied to multiple layers at once? Why not?
Forget about big-ticket features. What Photoshop needs is some finely tuned, pixel accurate love. –Bjango
The Filter/Adjustment inconsistency isn’t the only one, though. The User Interface is a mixed bag, as brought forth in Problem #5.
Excluding all the one-shot Filters (Blur, Blur More, Solarize…) and the Filters that build separate workspaces (Lens Blur, Lighting Effects…) I’ve gone through all Filter dialogs. None of them are particularly overwhelming, but some are worse than others. Unifying the User Interfaces and adding missing functions would be a much needed update:
The dialogs above are pretty fine as they are. They give you a small 100% window into your document so you get a quick idea of what it’s doing, and hitting the “Preview” checkbox will show you the big picture in the background. Alas, the Diffuse dialog is different — you have no control over the amount. The only true issue here is the fact that you can’t resize any window, so you’re stuck with the 206 × 206 pixel image.
Onwards…
Here’s where it gets weird. The sliders have changed their look and you get this odd 256 × 256 pixel preview window (with scrollbars..?) with a re-done zoom menu. It’s like a whole different world.
The preview-checkbox isn’t there at all and the sliders give no units, except %. Your options are dumbed down, too. For example, in the Wind menu you have radio buttons only for three steps in the Method and two in the Direction. Why isn’t the Direction a dial, like in Motion Blur? Why isn’t the Method a slider, like in everything else? None of these dialogs are resizable either.
The drop-down menus in Smart Blur, Ripple and Mezzotint make little sense too. Sliders would give much more power to the user.
And like I mentioned earlier, Polar Coordinates has no slider to control how much you want to warp it. Can’t set the reference point either.
Which brings us to…
I already made a chapter regarding the Displace dialog, but these other five are the worst offenders among it. There is no thumbnail preview nor full preview. You get very little context. Your only controls are radio buttons, check boxes and text fields with no hints of implied maximum or minimum values. These dialogs are the very definition of “I have no idea what I’m doing.”
And finally…
I don’t know for certain but I have a feeling that these (besides Offset) are remnants from the old, old versions of Photoshop that the folks at Adobe never bothered to remove. Like Radial Blur.
For Spherize, Pinch and ZigZag you have a 128 × 128 pixel net that gives you some idea of what the distortion is going to look like. You can’t set the reference point in any of them and your options are very limited. The preview window is the same strange box and you don’t have a Preview checkbox.
You get no slider or numerical field in Shear at all, and Wave only gives you three options for the Wave type but doesn’t let you define them yourself.
None of the Filter dialogs can be resized nor do they offer Interpolation options.
The whole Filter Gallery is the same story, too. You get a bunch of ambiguous sliders and that’s it. You can’t change your foreground/background colors afterwards either.
Not even the new-ish tabbed panels are safe from the User Interface problems.
Going through every resizable dialog one by one and re-designing them to fit every possible size and orientation would probably be a good idea.
Adding support for operators to all numerical fields would be a much-needed addition as well.
Illustrator and InDesign let you do this already and After Effects can even do advanced math, but Photoshop throws an error immediately if I type an asterisk in there.
Besides Blending Modes and math in numerical fields, After Effects also has a lot more Effects than Photoshop, despite the formats are somewhat related. There’s no denying that unifying some of those as well would make a gigantic difference.
3. This isn’t all.
I was greatly inspired to write this article after reading Adobe Gripes, Bjango’s blog and 4chan’s /gd/. While I’ve done my best to list the things that make me mad/sad/irritated, this wall of text certainly isn’t everything that’s wrong with Photoshop. There are probably a hundred other points I missed and some of the things I just listed can probably be fixed with a simple trick I just had no idea of.
There are also many smaller things I didn’t bother to elaborate on. You can’t place multiple files at the same time. You can’t have Pointer Locking with the Hand Tool. You can’t have a true full screen mode with Menu Bar. The default repeat setting for GIFs is “Once.” But these are minor gripes.
It’s just that if you have a problem or a suggestion you really don’t have a choice but to go through the bureaucracy of feature request forms or visit the forums. Even if you were right all along, there’s no guarantee anything will happen. I hope this lament of mine has some effect at least.
I know Adobe is a giant company with thousands of employees, but even a small investment in transparency and embracing the community would probably pay itself back tenfold.
Guillaume Stordeur made a fantastic program that gives Photoshop (and Illustrator/Flash) users extremely diverse Lazy Mouse features. It’s Windows only and has a price tag, despite being a fairly simple thing to implement straight into Photoshop.
Although I’d imagine Lazy Mouse is a fairly often requested feature, Adobe has so far done nothing. Maybe someone even in the Photoshop development team coincidentally uses that very tool right now yet wonders why don’t we do just the same — pay the price, get the features and support the developer while you’re at it.
But that brings us back to Trevor Morris’ scripts. Being able to distribute space properly in Photoshop is a very useful feature but we already have the scripts so many might not care. It’s the same issue with Lazy Mouse. And probably many, many others too. What Adobe needs to give us users are the things we need, big or small. Not features nobody asked for.
That approach simply doesn’t cut it anymore. To stay relevant, re-examination of every feature is needed. Re-factoring of the code is needed. Fixing old problems before creating new ones is needed.
One might argue that Photoshop is old and making radical changes would break compatibility.
But that’s precisely the problem here.
If the final arguments are “it will break the format” or “it’s too processor intensive for Photoshop” or “it would require a total rewrite from the ground up” or, God forbid, “this functionality requires additional work to the new architecture” then Photoshop is fucked. As of writing this article Affinity Photo opens PSD files faster than Photoshop itself.
Computers get better and better every day and it breaks my heart to see Photoshop still as clunky as it was years ago, getting “performance updates”, “stability fixes” and practically useless or broken features with the price of irritated customers and hundreds of dollars. But hey, at least the investors are happy.
In closing
Here’s some figures.
If I were to buy Affinity Photo and Sketch 3 now, that would cost me a total of 165 €. Once. Even if I never used them at all.
If I were to buy Adobe Photoshop, that would cost me around 20 € per month (well, depending on what browser I use and a handful of other factors) for the rest of my days. Sure, I’d get all the cool new features with the updates, but I’d also sign up for all the risks, too.
Is it worth it? Well, there’s no guarantee of what Photoshop’s future is like, and once you dwell into the Cloud, it can be tough getting out.
That smart someone I was talking about in the preface was Jon Hicks.
He also said “I don’t want to rent the tools I use for work.”
And he has a valid point.
In the end, I’m not the right person to give answers to all these questions and I have no idea what the future of Photoshop or Adobe in general holds, but I know for certain that if this is the direction they’re headed, obsolescence might be just around the corner.
The last tips I can give for Adobe is to embrace the community and always give your absolute best, even if that means major re-working under the hood.
Adobe Illustrator, another love of mine, has some problems of its own, too.
But that’s a story for another time.
