#tree explains things

(This transcript is approximately 75% accurate to the tutorial and 25% me condensing my waffle into more sensible narration.)

--

If you're not familiar with how to extract files from your video source, I have another tutorial about a program called SnapMotion (here) that explains some basics about frame rates and how you go about deciding what frame rates to use with particular sources. For this tutorial I'll be using caps extracted from a 24 fps clip at every second frame.

Since this is a pretty basic tutorial, we're not going to worry about things like the number of frames or the eventual size of the gif. However, I will briefly go through Tumblr's dimensions and maximum file size for images because ultimately that's what you'll be working towards when you're making your gifs.

When I started making gifs, Tumblr's maximum image file size was 3MB. It's now 5MB, which is a pretty significant increase. The Tumblr dashboard displays single images at a fixed width of 540 pixels. You can also post images in rows of two or three. If you have two images side by side, they will both have a width of 268 px; if you have three images side by side, two of them will have a width of 177 px and one will have a width of 178 (because 540 isn't evenly divisible by 3).

If you upload images that don't conform to this width, they'll be stretched or shrunk to fit. I recommend making your gifs at these exact display widths because I find gifs, in particular, don't shrink well, especially if they have text on them. They just don't look good at all. And nothing looks good when it's been stretched.

--

The Actual Tutorial Part of the Tutorial

Once you've opened Photoshop, go to File, Scripts, then Load Files into Stack. From the Load Layers screen, click Browse, and select all the files you want included in your gif. Then click Open. You can scroll down in the files box to see the names of the selected files. From here, click OK. Photoshop will open each image file and insert it into this newly created file as a layer so that each cap from your video source becomes a layer in a single Photoshop file.

My next step is to resize the images in the file. Since I'm making a single gif for this tutorial, I want to make it in Tumblr's large size. To resize an image, go to Image then Image Size. Enter 540 in the Width field. Make sure your dimensions are linked so that the height is changed proportionally to the width. Click OK.

To save the file, go to File then Save and choose a name and destination the way you would when saving any type of file.

Creating the Animation

Go to Window then Timeline. This is the area where you'll create your frame animation and then your video. To insert the layers you've just imported into this window, click Create Frame Animation. This creates a single frame populated by one layer in your stack. In order to insert the rest of the layers into frames, click on the timeline menu and select Make Frames From Layers. Each layer now populates its own frame and can be viewed by clicking on that frame.

At this point, the first layer in the stack appears in the last frame of the animation. When Photoshop inserts the image files in the Load Layers process, it does it backwards. This means that the last image (in this case, image 40) is loaded first and the first image is loaded last. In terms of viewing the layers panel, it makes sense, because it means that the images appear in order; however, the effect of this is that the frames are also backwards. In order to correct the order, go to the timeline menu and click Reverse Frames. Now the last frame corresponds with the last layer and the first frame corresponds with the first layer, which is what we want.

At the bottom left of the timeline window are the controls. In the centre is the Play/Stop button. The button to its right will take you to the next frame and the button to its left will take you to the previous frame. The left arrow|bar button takes you back to the first frame. Next to that is the Looping menu. Although it's possible to select other options, for our purposes looping should be set to Forever.

Playing through the animation at this point shows that the frames are in the correct order but it's too fast. That's because there's no delay between frames. To select the frame delay for a single frame, click on 0 sec. at the bottom of the frame. You can choose from the presets in the menu or click Other to enter the whole or fraction of a second of your choice. To select the frame delay for all frames at the same time, go to the timeline menu and click Select All Frames. Now when you click on the bottom of one of the frames and select a duration, all the frame delay durations are changed.

Playing through the animation after applying the frame delays shows that it's slower. Actually, it looks too slow now. For some reason, playback in the frame animation timeline appears slower than in the video timeline.

You can see what the frame delay speed really looks like by viewing it in the video timeline. Click on the icon at the bottom left of the window. The controls in the video timeline are basically the same as those in the frame animation timeline, plus a few others that we won't be needing right now. Each of the layers in the video timeline is one frame plus the frame delay we just selected. Playing the animation in the video timeline shows that it looks smooth and more like the normal speed that you'd see a video.

To return to the frame animation timeline, you can either undo the previous action (Edit, Undo) or click on the icon at the bottom left of the window, which will convert the video back into frames.

Ordinarily I like to crop the height of my images even after they've been resized. It's partly because I like the way they look with a larger width to height ratio, but also because reducing the dimensions of the images is going to reduce the size of the gif file and that's always a good thing when you're dealing with limits.

To do this I use the Crop tool. From the menu, I select W(idth) x H(eight) x Resolution. My last preset was 540px wide and 290px high. I'm going to stick with that because I don't want to cut off too much of Vic's head in this. The resolution is set to 72 px/in, which I check against the image files' resolution located at the bottom of the window next to the image dimensions. Clicking into the crop screen shows the crop grid, which by default is the centre of the image. I usually test that out first before making any adjustments. To approve the crop, I click the tick at the top right of the toolbar, then click the play button to determine if the cropped animation looks the way I want it to.

When I'm happy with my frame delay and crop, I go back to the video timeline. You can stretch or shrink the timeline by adjusting the slider at the bottom of the window. I usually run the image through one more time to make sure I'm happy with it before committing to the animation.

The next step is to create a Smart Object. Highlight all the layers in the layer panel and right click, then select Convert to Smart Object. Now there's only one item in the layer panel and one item in the video timeline.

The last step is to sharpen your video. Some people might say that that's actually an editing step but I absolutely disagree. I think sharpening is foundational and it should happen at this point. It should not be optional. Everyone should sharpen their damn gifs.

Go to Filter, Sharpen, then Smart Sharpen. The main setting to be concerned with is Radius. A smaller radius means less sharpening and a larger radius means more sharpening. When I'm working on larger images like this one, I tend to use 0.4 or 0.5 px sharpening; when I'm working on smaller images (e.g. 268 px wide), I tend to use 0.3 or 0.4 px. It depends on a lot of variables: the image quality, the amount of motion (sharpening is less effective when there's a lot of motion), what the light is like, etc. Work with the preview to determine what looks best. (Although what I just did is hideous. Don't do that.)

With this image I choose to go with 0.4 px because she's moving a lot and I don't think the extra point of a pixel is going to make a lot of difference. Click OK to confirm the selection. If you disable the filter visibility now, you can see that everything looks slightly blurry. When you enable the filter again, there's a lot more delineation in terms of the lighter areas and at the outer edges. I find that Smart Sharpen is particularly good at picking out the edges of light coloured things and also metal. Metal comes up really nicely in Smart Sharpen.

So that's Smart Sharpen. As I said, everyone should sharpen their gifs. This is a hill and I will die on it.

The very last step is to export your animation. Go to File, Export, then Save for Web (Legacy). The main thing to look at in this screen is in the lower left corner. (Obviously this file is too large for Tumblr.) The main contributor to your file size at this point is the percentage of Dither. I don't change any of the other settings. At the lower right of the window is a kind of final check area. It shows the dimensions of the animation and the looping. You can also play it back here.

When you're satisfied with all of these elements, click Save to save your gif file.

And here's my gif!

--

--

Create a New Document

The width is already set at 540 so we'll just go with that. I'll leave the height at 150 because the height really isn't that important. The important thing is that the Background Contents is set to Transparent. Click on Create.

Add Filler Text

With the Text tool selected, click anywhere inside the document. This will create your text layer. Then you can type in whatever you like. I'm going to be very unoriginal and type in what I have for the other ones.

Design Your Text Style

Since I've been using these presets recently they're already here in the Character panel. I also add text effects. All of these settings are subject to your preferences. You don't have to emulate these, although you're welcome to. I got some of them from a tutorial and I think I just played around with them until I liked the way it looked.

Convert Point Text to Paragraph Text

Go to the Type menu and select Convert to Paragraph Text. Now when you click into your layer, you have these little moveable sliders. (I don't know what to call them.) Bring the left and right sliders all the way to the edges of your document. Choose whatever height you like and when you're happy with the size, click the tick in the toolbar.

Now that you have your text set up, all you have to do is line it up with the borders of your gif and it will be perfectly in line every time. I've set the text alignment to centre in the Paragraph tab but, again, you can choose any alignment you like.

Choose Your Colours

I'm sure you're aware of the convention of using alternating colours for different characters. I start with white, obviously. My second colour is #FFFF33. That's the yellow that I use. My third colour, if I need it, is a turquoise that you probably saw (on the way to yellow, because I'm super lazy).

Add That Text! The 540px Version

Drag your text layer from the template to your gif and line it up. For some reason, text layers are always 5 seconds long. I don't know why; I don't know how to change it. If you do know how to change it, please tell me.

Now when I play this the text appears across the entirety of the gif. If I only wanted it to  appear while she moves her mouth, so while she's actually speaking, I find where I want it to end with my playhead and I drag the end of the layer to that point. Now when I play it, the text is not on the rest of the gif. It only appears where I've set it.

If she started speaking in the middle of the gif somewhere, I could move the text layer around along the timeline. You can basically have it appear anywhere and you can make it as short or as long as you like. Don't go past the boundary of the end of your gif, though, because if you do, you'll have a spot at the end of your gif that looks like this. I'll just show you. See, you get that little flicker.

Add That Text! The 268px Version

For this one, I only need two lines of text, so I'll just drag the two line layer over here. This is something that I forget all the time. If your playhead is not set at the beginning of the timeline, the text layer that you drag over will start where the playhead is positioned and you'll have to move it. I do that most of the time.

Now when I hit play, again, the text appears across the whole gif.

Multiple Characters! Option #1

I will just copy this text layer. Then I can move it and change it as I like. Let's say Walt responds to her. It being Walt, let's just say he says, "I don't know," because he says that a lot. So now Vic says, "What are you thinking?" and then Walt says "I don't know."

Multiple Characters! Option #2

Your other option for more than one person speaking in a gif is to have it all on the screen at the same time. In that situation, what I would do is go back into my original text layer and enter the other piece of dialogue, then change its colour to my alternate. Now both pieces of dialogue are going to show on the screen the whole time. That's the other way you can do it, if you prefer.

--

--

This transcript is about 90% accurate. I'd like to think it's because I'm getting better at making these videos, but it's more likely because most of what I said ended up getting deleted.

Part 1

Imagine you have a vault and you have a bunch of things to put in it. Before you put them in the vault, you take a photo of them. Once you've put them in the vault and shut it, you stick the photo on the outside so that you can always see what's inside it. It doesn't matter what you do to the photo. You can write on it or draw on it or erase parts of it; you can cut it up; you can make copies of it. But no matter what you do to it, the contents of the vault remain untouched. So if one day you go and open your vault, you'll still have all the things that you put in there exactly the way they were when you put them in there.

That's what a smart object is, essentially. When you save your layers into a smart object, you're left with the photo. So, like the photo, you can take the smart object and you can write on it, draw on it, cut it up, erase parts of it, duplicate it. You can treat it as you would treat any layer in Photoshop.

It took me a while to figure out why my file sizes were sometimes really large and sometimes not. That's when I started to figure out what smart objects actually were and did. I've got this gif that I've started working on. I've saved it with the dimensions of 540 x 304 pixels and as you can see there are 51 frames. In Finder, the file is 23.6 MB. I have another copy saved but in this one I've left the original dimensions of 1920 x 1080 pixels. In Finder, this file is 279.1 MB.

If I convert the layers in this file into a smart object and then resize them, the file is 280.2 MB in Finder, versus the 23.6 MB of the other file. This is because, as I said, your smart object retains your layers as they were. Inside the smart object my images are still 1920 x 1080 pixels. It doesn't matter that I've resized them to 540 x 304.

Right click on the smart object and go to Edit Contents. You can see the dimensions of these images are 1920 x 1080 pixels. Fifty-one images with those dimensions are inside this smart object that, when I look at it, is 540 x 304 pixels. That's why the file is so big. And that's really, for me, the only downside to them: if you forget to resize your image you're going to be lugging around ten times the size of a file than you otherwise would.

By the way, that doesn't affect the size of your gif. It only affects the size of your Photoshop file.

[I replay the pertinent section of the video from earlier.]

Once you have the .psb file, all you need to do is go to File, Save As. As you can see, a temporary copy is saved somewhere in the depths of your Library in this folder. Change the destination folder to wherever you want the new file to be saved. Then change the Format to Photoshop. Now you have a .psd file. When you save it, you have a file that's exactly what you had in the original file prior to converting the layers into a smart object.

Part 2A

When you're making larger size gifs that are 540 pixels wide, you'll often have to sacrifice portions of a shot in order to fit the eventual gif size under Tumblr's limits. The largest contributor, obviously, is the individual images themselves. The more frames you have and the larger the frames are, the more data is in the file. Fifty-one frames is too many for a gif of this size. It will definitely be over 5 MB. I usually try to stay under forty frames; that's still pushing it a lot of the time. That's one of the reasons why I crop my images, sometimes quite severely. It's because I don't want to sacrifice part of the shot.

Realistically, thirty to thirty-five frames is probably the best median to aim for when you're making a large gif like this. It gives you a little bit of leeway—it's going to be over 3 MB but probably not that much—especially if you don't want to sacrifice your height. For example, a gif of this size, at 304 pixels high, is going to have a significantly larger file size than if I cropped it down to my usual 250 pixels. Fifty-four pixels actually represents quite a significant amount of data. Obviously, the larger the image, the fewer the frames you want, in order to balance out your file size.

Normally what I do is take from the top or the bottom of the shot, or sometimes both.

I'll delete the first six frames. All you have to do it select them and then click on the bin icon. I still have 45 frames, which is still too many.

Now I want to turn it into a smart object except I have all these extra layers that I'm not going to be using. They're very simple to delete, the same way you would delete any other layer. Select them and click on the bin. (I said press. IDK why.) Then I scroll down to create my smart object. And finally I delete the layers at the end. So that's how you delete your frames prior to converting your layers into a smart object.

Part 2B

I saved a second copy of this gif with all fifty-one layers in my smart object and now I want to cut some of it to reduce the file size. To do that you'll be using the scissors icon: Split at Playhead (although I think of it as trimming). It allows you to cut your animation into multiple sections that you can then choose to delete.

When you have the playhead where you want it, click the scissors icon. (I said press again!) Now I have two animations — the original file name and the copy. Delete the extraneous part and move the remainder to the beginning of the timeline.

As you can see, if you look at the bottom of the window, this frame starts at 03:18. It goes through 19 and 20, and then at 03:21 it changes to a different frame. I'm going to cut there, which will remove from this frame onwards and retain the previous frame. Select the section to delete and click the bin icon. (I SAID PRESS AGAIN.)

In the previous tutorial, I mentioned that when you convert your frame animation into the video timeline, each layer in the timeline represents the frame plus the frame delay. Photoshop's animation speed is set to 30 fps and I've specified a 0.1 second frame delay, so that means every frame in this gif is visible for three frames in the video. At the bottom of the window you have the running time, which shows as seconds and frames. The 03:18 represents three seconds and eighteen frames into the animation. If I cut this at 03:19 or 03:20, the frame that spans 03:18 - 03:20 would appear in both cut sections, one at the beginning and one at the end, obviously.

The benefit of removing frames this way is that because it's a smart object and, as we know, a smart object is a vault, you will still have all fifty-one layers left in your smart object. So even though you've cut it up, it's all still in there.

Part 3

In this example I want to speed up a section in the middle of this shot. There's a bit in here between when he looks down and then when he looks back up that looks fine on screen but, watching it as a single gif, feels like it goes for too long. I don't want to cut the gif anywhere because it would be too obvious if I cut the middle out of it and it's a reaction shot and I want it in there.

I'm choosing frames 13 to 22 and I'm going to change the frame delay to 0.05 seconds, which is half the frame delay of the other frames.

It doesn't feel like it's dragging anymore, which it did before, in my opinion. His expression looks natural; there are no missing frames, so it doesn't look jerky; it's just that the little bit in the centre of the gif has been sped up so that the pause between him looking down and then looking back up has been shortened.

In different circumstances I might change every second frame or every third frame. It's really a case by case thing. There's no way to predict what you're going to want to do with any individual gif until you get it to this point and start looking at it as a single shot versus a shot in amongst a lot of other shots. There's a lot of difference between a series of shots that happen one after the other in continuity and those individual shots when you break them down and you're just looking at them one at a time.

And that's really all there is to it. It's just a matter of you deciding how you want your gif to look and then adjusting the frame delay accordingly. You just have to experiment and try things out and see what happens.

--

(baaaaa)

(This transcript is approximately 60% accurate to the tutorial and 40% me condensing my waffle into more sensible narration.)

--

SnapMotion is a Mac OS X program available to buy at the App Store. (As far as I'm aware, there are no releases for any other operating system.) It's fairly inexpensive: AUD $7.99 when I bought it two years ago. (In the video I said $12, from memory, but then I looked up the receipt email to confirm and found myself $4.01 wrong.) Considering that I've made thousands and thousands of caps in the couple of years I've been using it, I've definitely gotten my money's worth.

SnapMotion has two modes: what I call photo mode and then batch mode. Photo mode is good for one or two caps, but to extract caps in bulk you'll be using batch mode. (Something I forgot to mention in the video is that dragging your file to the SnapMotion icon will open it in photo mode by default. In order to access batch mode, you have to open the program and select it.)

Selecting Show Batch Mode takes you to the batch mode work space. Here you can drag and drop your clip or click the plus sign to add it. Once you've loaded a clip, select it in order to access the Add Batch Tasks menu.

One of the things that I particularly like about this program is that all the information you need about the file is clearly set out. In particular, it shows you how many frames per second you're working with. This is important because the fps rate determines how many frames you want to extract from a clip.

--

Why? Because math.

All the sources I've worked with have an fps rate of either 24 or 60. That number tells you how many individual images are contained in one second of video. Obviously, there's a considerable difference between 24 and 60, and that affects how many frames you choose to extract.

I've found I get the best results by extracting every other frame from 24 fps sources and every fifth frame from 60 fps sources. So:

24 fps = 2 (or 1/2 ← why does Tumblr do that?? I hate that!) 60 fps = 5 (or 1/5)

(That gives you roughly 12 caps per second for both sources. I say 'roughly' because, as you can see on my screen, the fps rate is rounded to the nearest whole number. 24 fps is actually 23.7 fps, so that missing .03 can have a cumulative effect.)

I find this is the best compromise between having smooth, natural looking motion in your gifs and keeping your file size within Tumblr's constraints. (I'll talk more about that in the next tutorial.)

--

Tools

The green bar is obviously your clip timeline. The cursor defaults to being in the centre for some reason. You have start and end caps (or whatever you call them) that define the start and end points of your extraction. The left and right arrow|bar buttons move your cursor to the start or end cap respectively, and between them is the play button. To the left are the transformation options. I've never used them, so I don't know what they do. (If you decide to experiment, please report back about your results!) To the right of that is the volume icon, which allows you to hear sound from your clip during playback. I don't find it useful, so I don't bother. The tortoise icon allows you to adjust the speed of your playback (left is slower, right is faster in case that’s not obvious). The arrowheads on the far right move your playback cursor to the start and end caps respectively.

Add Batch Tasks

This is where you specify the parameters for the batch.

Mode: there are four options. I primarily use Every X Frames. As this is a 24 fps source, I want to extract every 2 frames, so I use the value 2. When I type my value in and then press tab, the bottom of the panel shows the number of frames that will be generated from my selection. If I process this selection, SnapMotion will create twenty-five images. But because I moved the start and end caps earlier, and I actually want to extract frames from the whole clip, I need to adjust them. When I tab through my X value again, I see an updated number of generated frames.

I always include the Starting and Ending frames because I trim my clips to the exact frames I want (unless I'm being lazy).

Max size: This refers to the width of the image in pixels. (Technically, it refers to the maximum size of the largest dimension, but since video sources are universally landscape oriented, for our purposes this number will always refer to the width.) The default setting is 4,000 pixels and since the source clip is 1920 pixels wide my caps are going to remain 1920 pixels wide. However, if you know what width you want your eventual gif to be and you don't want to go through the extra step of resizing it in Photoshop, you can use this setting to do that for you. For example, if you want your gif to be 540 pixels wide you would enter 540 in this field. When you process this batch, SnapMotion will resize each frame it extracts to a width of 540 pixels with a proportional height. (Yeah, I said width at first and then I had trouble getting out the word height. Talking is hard!)

I don't use this feature simply because sometimes I change my mind, and I would have to rerun the batch if I did. (This isn't a big deal for a single shot like this, but if you're working with a whole scene that contains hundreds of frames, it's a hassle.) I also reuse caps sometimes. As you probably know, if you follow my Longmire gifs, sometimes I do a full scene using small gifs and then I go back and do large versions of some of the Vic shots. (Because I love her so.) And sometimes I make other graphics. In those situations, I want the image at its original size, so that I can resize or crop it the way I want.

Format: I prefer PNG.

Override: this tells the program what to do when a file with the same name already exists in the destination folder. The options are Replace existing files or Use alternative name. They're probably fairly self-explanatory (but just in case, I included a note).

Errors: I don't think I've ever had an error in this program. You can choose to generate a log file with each batch, but I kept getting logs that said "no errors" every time I ran one and it was annoying having to constantly delete them, so I switched to No error handling.

Headline/Credit: I believe this is like watermarking your images. If you want to watermark your gifs (I honestly don't see the point but some people do it), I think it'd be a lot easier to do it in Photoshop (or whatever program you use for image editing).

Export: this is where you specify the destination for your extracted image files. Click Choose to open the menu from here. Alternatively, when you're ready to process a batch, click the red button (aka the "Do the Thing" button according to me when I recorded this), and if there is no destination already specified, it will prompt you for a location.

Once you've clicked the Do the Thing button (yeah, I'm just leaning into it now), SnapMotion will display Task Added. If you look at the top of the window, you'll see a progress bar showing the percentage of completion as it processes.

When the images are fully exported, you'll see All Tasks Done and also an alert if you've selected that in your preferences (which I haven't gotten to yet but it's coming up). If you go to the destination folder, you'll find all the images that have been extracted. That's it!

--

I want to address the syntax for how files are named because I found the options very confusing to begin with. The default syntax made no sense to me and it took me a while to configure something that did. File names are defined in the program preferences, so go to the SnapMotion menu, select Preferences, then select the Batch tab.

Number of concurrent snaps: how many frames should be extracted at the same time. This has an impact on your RAM usage and also impacts how long it takes a batch to run.

Alert users with: this is where you choose whether or not you want a notification outside the program. It's useful if you're running a larger batch in the background and therefore don't see the All Tasks Done message.

Output name format:  these are the elements I use:

<movie_name> = the title of your clip file

hyphen

<time_seconds> = the actual second position of the frame within the clip; for example, a frame that was extracted at 14s would be numbered beginning with 14. This parameter converts durations of 1 minute+ into seconds, e.g. 60, 61, etc.

<frame _index> = the running number of the frames extracted in the batch, beginning at 1. The position of this number depends on the total number of frames you're extracting. From 1 to 99 frames, the number will be 01; from 100 to 999 frames, the number will be 001. I've never extracted more than 1,000 frames, but I assume in that situation you'd begin with 0001.

If I were extracting 40 frames from my clip beginning at 14s, then the numerical portion of my file names would start at 1401. If I were extracting 120 frames, then the numerical portion of my file names would start at 14001.

<image_extension> = the file format you've specified

So, my file names are constructed like this: 

<movie_name>-<time_seconds><frame _index>.<image_extension>

And the files from the batch I've just run look like this if you separate the elements: 

1x05_2206 - 0 01 . png

In this batch, the seconds run from 0 to 3 and there are 34 frames in total. The first number column represents <time_seconds>, while columns two and three combined represent <frame_index>. You can see that the numbers in columns two and three are sequential from 1-34, whereas the numbers in column one aren't. That's because I've extracted multiple frames per second. Caps 1-7 were extracted from the first second of the clip, caps 8-18 from the second, caps 19-31 from the third, and caps 32-34 from the fraction of the fourth second.

This naming system is what works for me, but you should experiment with the options so that you can construct file names that work best for you.

--

The last couple of things:

You can import multiple clips either individually or at the same time. 

You can delete a clip by selecting it and clicking the minus button.

--

And that's SnapMotion!