Three Years After Graduating from Communication University of China: "I Want to Push Myself to the Limits of Animation

Gorgeous, cold, and stunning…… At the end of 2025, the Chinese animated short film Fanfan Xiangsheng was officially released online, immediately drawing attention within the industry.

​Replay Share Like

Animation AcademicBaseFollowing

/

​0.5x 0.75x 1.0x 1.5x 2.0x

​HD Smooth

Three years after graduating from Communication University of China: “I wanted to push myself to the limits of animation”

,Three years after graduating from Communication University of China: “I wanted to push myself to the limits of animation”

​Animation AcademicBaseFollowing

This is an animated short that Qingyu made independently after graduating from the Animation program at Communication University of China, taking on freelance commissions while independently producing it.

Fanfan Xiangsheng draws inspiration from the history of the ancient Dian Kingdom. With richly detailed lines and colors, a cold and restrained visual language, and an ethereal, solemn musical atmosphere, it creates an unforgettable viewing experience in just three minutes and twenty seconds.

Its psychedelic and subtle narrative takes the viewer across a thousand years of time and space — from a modern museum to the pastoral scenes of an ancient kingdom, from rituals and war to rebirth and transcendence — completing a poetic expression with a long-lasting resonance.

​Its recommendation in AcademicBase's earlier 2025 Animation of the Year

With Qingyu's consent, the following shares a full behind-the-scenes retrospective of the making of Fanfan Xiangsheng.

In the article, the author candidly and rationally walks through the entire process — from creative motivation and pre-production preparation to production planning, shot design, inking and coloring, and post-processing — and even discusses how to adjust one's personal state when facing fatigue and anxiety.

In addition, Qingyu has documented more specific behind-the-scenes details about the coloring, layouts, and complex shots on their Weibo page. Interested readers can search Weibo for @A青宇A after reading this article to learn more.

00 Preface by the Creator of Fanfan Xiangsheng​

The meaning of retrospection: The benefits of retrospection aren't immediately apparent; it mainly helps clarify one's thinking, gives a rough overall picture of the project, and serves as a reference for future projects.

Many problems are solved through "iteration," and the prerequisite for iteration is improving upon an existing plan. If you cross the river by feeling for stones every time, muddle your way through and finish without any concept of how it was done, let alone optimization, that's a problem. So you need to promptly record each step of your actions, then organize them afterward.

However, this article can only offer ideas for "independent creation." I have relatively little experience with team-based creation; I hope to find suitable collaborators in the future and fill this gap.

Why I worked independently:

First, I wanted to be involved in every step — if it weren't something I led myself, I probably wouldn't have the chance. Second, the workload was enormous and I couldn't afford to pay anyone; even if people helped, it would be at most a shot or two — a real drop in the bucket. Even if someone were willing to take on a lot of work, first, I'd worry about making them sick from exhaustion, and second, their skills might not fit the needs. Plus, independent production has its own appeal, so I just exploited myself instead.

Why I chose the MV format: First, the purpose of this piece was to try a new art style, but I also wanted to add some narrative. An MV has more storytelling than a single shot, and it's shorter than a feature film, with less sunk cost.

01 The Full Production Process​

(1) Motivation for making the film

1 — To make up for regrets about the art style of my previous film, hoping to continue earlier experiments and at least produce an art style that looks different.

2 — To see where my limits in art lie.

The meaning of clarifying motivation: when you're lost or considering a change of direction, you can compare your original motivation with your current trajectory, avoiding trying to have it both ways, and reclarify your goals.

(2) Set a target audience: animation creators and art enthusiasts.

The meaning of setting a target audience: it draws a boundary that lets you filter feedback, block purely venting comments, and receive external criticism more precisely.

However, note that once you've defined your audience, you can't only accept positive praise — even people who identify with you will raise objections. Their criticism is well-intentioned and should be taken into consideration.

(3) Research and find inspiration

This film began with deciding to make a music MV, and only then did I look for a suitable subject — less certain than starting from a theme. During the inspiration-finding period, I generally watched documentaries and visited museums without a clear purpose, gradually arriving at a theme. Since these activities are fragmentary, they can be interspersed among many important tasks — like moving house, doing commissions, etc. — which saves time.

(3.5) Setting the schedule

Since the purpose of this film was to push myself to the limit, the schedule was not strictly controlled.

The initial expectation was about a year — my previous ten-minute short took ten months, so I figured a three-minute short could surely be done in a year. The result: two delays, taking 20 months in total.

Although it looks like setting a schedule was pointless, it still gave me constraints and prevented endless procrastination. It was delayed twice even with deadlines; if I had let myself work indefinitely, who knows how long it would have taken.

(4) Text-based storyboard, rough animatic

Settings made at this stage will most likely be overturned. But if you act out of fear of revision, wanting perfection from the start, it becomes very hard to begin.

Making a pure failure piece right away, then continuously adjusting and revising, actually lets the plan proceed smoothly. So at this stage, I generally don't consider especially polished logic — I just stuff everything I want into the film, remove what doesn't fit, patch up logical holes, and don't fear starting.

But revision has costs. To reduce losses, I usually go with a text-based storyboard: a specific description for each shot, with a rough estimated duration. Then, based on the text storyboard, I draw an extremely rough animatic directly, synced against the likewise rough melody designed earlier. Combining sound and picture earlier helps discover problems earlier — like whether the visuals can sustain the duration, whether the shots flow together logically, and so on.

Following this process, like washing wheat gluten, you wash away the unneeded parts from the initial mountain of piled-up ideas; repeating this, the story's final form gradually takes shape.

Since my creative method is to gradually comb a pile of scattered ideas into a thread, during later revisions I often need to swap the positions of certain shots to make the logic flow.

Compared to pushing forward linearly from the beginning, step by step, the advantage is easier revision and easier incorporation of new ideas that pop up suddenly; the downside is that it takes more effort to straighten out the logic, and occasionally you suddenly discover bugs and have to grind away at fixes.

(5) More precise research

After the story content is officially decided, there's no longer a need to watch all kinds of material in a broad sweep. Taking this film as an example, since the time focus was on “the Han army entering Dian and the Dian people's exile,” research could be guided by descriptions of that period.

Of course, as I'm not a professional and my channels for obtaining information are limited, it was enough that the material sufficed to fill this segment of animation.

(1) The formal storyboard

Once the content is decided, of course comes the storyboarding stage. The reason I went straight to an animatic rather than traditional storyboards is precisely the creative method mentioned above.

Because the “point-to-point” creative method frequently involves reordering, with traditional storyboards there would always be pages whose images need to be swapped back and forth — move this one forward a page, move that one back — such a hassle! So I just drew an animatic directly, where I could move things freely along the timeline without worrying about page swaps.

And an animatic can be synced directly with the audio; while revising the melody, I could also verify whether the rhythm felt comfortable, adjusting the duration in time.

(2) Refining the storyboard

After determining the rough duration and content of each shot, refinement can begin. The refinement mainly covers: the main characters' actions (stick figures), supporting characters' actions, scene framing, and specific content.

During refinement, I also design the elements involved. That's right — character designs, scene and map design, and other pre-production designs only began after the storyboard.

The reason I didn't follow the traditional flow of doing designs before storyboarding is as said before: the “point-to-point” creative mode caused massive revisions. If character designs were finalized at the start, they'd inevitably have to be redone after the story was settled. To reduce revision costs, doing designs after the storyboard is settled greatly reduces the number of revisions.

Also, designing characters based on the storyboard avoids redundant elements in designs/scenes. If a character never shows their back in the whole film, designing only their front halves the workload.

Therefore, prioritizing work by urgency and refining in order can reduce rework costs. Under this logic, I prefer to prioritize refining the parts that are [hardest to finalize] and [hardest to revise]: 1 — The storyboard has a huge volume, is hard to revise, and rough animation doesn't need accurate character designs or scenes, so it comes first in refinement.

(3) Character, scene, and pattern design, plus simple mood paintings

As mentioned above, in order of revision difficulty, various designs went on the agenda during storyboard refinement. Designs were done according to each element's share in the storyboard: the two protagonists first, then supporting characters with action design, then scenes.

(3.5) The problem

After the various designs are done, a problem usually arises at this point: everything's progress is roughly equal — where should refinement continue from?

This is where the importance of “re-tallying tasks” shows. A convenient method is to make a spreadsheet, fill in the current progress of each part, then arrange the order of mutual influence: for example, refining key animation requires accurate character designs as a basis, so character design ranks before key animation. Then refine in that order.

Of course, there's another possibility for being stuck: not enough reference material, or not knowing what you want to do. In that case, set the pen down temporarily, revisit the material you gathered before, or, based on existing designs, look at more existing examples to fill the gap. Once the direction is clear, it's easier to start drawing.

This method of re-tallying the workload will be used many times in the future when you don't know where to start. Often, being stuck is just going in circles in place; once you clarify your thinking, you'll find many entry points.

(4) Wrapping up pre-production

After re-tallying the workload, in my understanding, the things that need to be finalized are: character designs and the storyboard. Scene designs, being easy to revise, can be set aside for now — a color sketch suffices. The music just needs a reasonable melody; enrichment can wait until mid-to-late-stage compositing.

During this stage of refinement, I prefer alternating between tasks: on one hand, alternating avoids getting stuck on one item; on the other, after going through the “don't know what to refine” state, I really lost patience and could only chip away here and there.

Once the items above that need finalizing are done, key animation can begin — officially entering mid-production.

(1) Rough animation / first key animation

I don't really know what to call this part — it's basically a “nude body” animation, roughly looking like the image below. Sometimes I call it rough animation, sometimes first key, sometimes rough keys. Just know they're all the same thing.

The actual process is standard animation workflow — nothing worth writing about.

At this stage, you'll again hit points where refinement stalls. Then, pull up the spreadsheet again, write evaluations for each shot, and see where refinement is possible.

Then, when the rough animation is mostly refined, be sure to stop and review — at this stage, re-examine whether the direction has drifted: has complexity met requirements? Are some shots too ordinary? Since the rough animation exists, reviewing the whole reveals more problems.

At this stage, I found many shots too ordinary (camera angles too normal, character movements lacking purpose, art designs too plain and uncreative, etc.). I sent shots I considered inadequate back for redoing; since body sketch revisions are cheap, the revision cost here was acceptable.

(2) Test shots

Since one purpose of this film was to create a different art style, from the second key stage onward the process deviates somewhat from the standard workflow (commonly known as: I don't know what I'm doing). To proceed efficiently later, I entered the test shot phase at this point to find a solution first.

My criteria for choosing a test shot were: 1 — The action isn't especially complex (i.e., easy to revise)

​3 — The shot I chose was B3-2: a close-up can better showcase skin texture, and the shot also includes effects, with low revision difficulty.

After the test shot, the new inking approach, colored-line technique, and compositing method were established, and second key animation could proceed using this as the standard.

(3) Second key animation

Again not much to say about second key — it's just dressing the body sketches, adding follow-through, and such.

For the breakdown of difficult shots, see the Weibo post “Fanfan Xiangsheng Behind the Scenes (5) Color Design” — kindly search for it yourself.

(4) Color script

The reason for doing color scripts at this point is likewise to reduce revision costs.

Anyone who refines artwork often knows that color sketches and actual coloring are two different things. Often, as elements are deleted, the number and position of color blocks during actual coloring differ greatly from the sketch — frequently leading to “the color sketch looked nice, but once colored it turned ugly.” To avoid this, doing the color script after all elements' positions are settled (i.e., after second key is done) avoids relying on redundant color blocks for effect.

Since I needed to create a slightly different art style, the colored lines were handled differently from tradition: this film's colored lines were designed individually per color script, with accurate color modeling in place, rather than shadows imagined out of thin air in the key animator's head. This approach yields great results, but the drawback is that the process isn't universal and can't be put into production. Distilling a systematizable colored-line approach is one of my future goals.

(5) Inking and coloring

Not much to say — this part is the most mindless work in the whole process, a relaxing phase. But is it really?

First, what relaxes is the brain, not the hand. During second key, some near-identical lines could be skipped if I didn't feel like inking them, but during inking, except for fully overlapping parts, everything else still has to be inked. Since I don't like using stabilization aids and have a heavy hand, my hand often hurt while drawing.

Second, the brain relaxed — but a bit too much. Without suitable background audio, it's very easy to go numb. Mechanical work plus an enormous workload lasting months — you become like someone tightening screws in a factory, feeling no value at all. As the timeline shows, this period is highly prone to hand problems and mental health issues. [The boxed periods]

One solution is preparing plenty of background audio during breaks — don't watch interesting content in advance; save it to fill the void while drawing. Another is interleaving other work: as the table shows, adding art work as a diversion during the inking process can reduce boredom-induced mental problems. After this adjustment, there were almost no health issues in the later inking stage.

(6) Scene test shots

To reiterate, one purpose of this film was to push my art to its limit. So to first see where that limit is, I generally pick the background of one shot and refine it until I no longer know what to refine.

Note that the “limit” here is not the limit of the final film: with high-intensity drawing, skills improve on a monthly basis — often what I drew at the start of a month looked unacceptable by its end. So this was only the first verification of my limit; there would be continued progress, and only in the last month or two could I truly know where the ceiling is.

The criteria for choosing this test shot were: 1 — The scene contains frequently appearing objects (for reuse and reference)

(6.1) Scenes

The intent of scene rendering was to see where my ceiling was, yet the first version of the scene could only be described as a child's drawing. It was a splash of cold water; I was utterly disheartened.

Creation humbles you — that's the truth. Only after facing my actual level did I suddenly realize that those animation scenes I once “looked down on” were rendered with such mastery that the skill within them was invisible.

I could no longer use “not enough time” as an excuse, because that was genuinely my highest level at the time. I felt ashamed of my arrogance.

What followed was putting my paintings side by side with the masters', constantly finding problems, adjusting and revising. Though still far from the masters, I'd improved a lot from the start. Thinking back, being able to see the gap is a good thing — at least you know where to strive. The road ahead is long; keep going!

(7) Bottleneck again

In mid-to-late production, most shots had reached the “watchable” stage, and it's at this point that you may hit difficulty pushing forward — not knowing where to start, similar to the early-stage state of being unable to begin.

Because most shots only have refinement left, and refinement is work that takes many days to show the slightest effect, it's hard to see a qualitative leap at once:

Often, you render for a whole day, compare before and after, see no difference — argh! Breakdown! Then anxiety sets in: Am I actually making progress? Why does it feel like everything's frozen? Should I try something else? Then you randomly open other files, refine a few strokes, switch, draw a few more strokes. Invisibly, time is wasted on anxiety and canvas-switching.

So when you don't know where to start, I suggest pulling up the spreadsheet again and assigning priorities to the scenes. For example, the opening scene (A0-1) is the first impression, very important, ranked first; (A4-2), the Twin Mountains wide shot, is a test shot and also important, also ranked near the front; (B6-1), the overhead battlefield ground, not yet figured out how to draw, ranked last. Ordering shots by this standard and refining by importance avoids “worrying about wasting energy on unimportant shots.”

Then just start pushing forward. Note: once you've chosen, don't change — wavering wastes enormous time. And finally, trust that [as long as you keep drawing, you keep making progress]; don't worry too much about time.

(1) Post compositing/camera work

For this part, following the earlier test-shot method, just add effects to all shots.

Of course, this only applies to shots with characters. Some shots need motion effects added separately in post — like B4-1's falling water surface and drifting leaves. All these dynamics had to be figured out in post by myself, so they're harder than shots whose realization method was already established.

So such shots have far higher priority than the rest; the other shots are just diversions between work.

(1.1) Compositing philosophy

Note: This is merely an introduction; the main purpose is a retrospective to clarify thinking, not a tutorial. The workflow is also not standard — apologies.

One, two “nevers”: In this film, to ensure the detail of key animation and scenes is presented fully, there are two things I absolutely would not do. First, no blurring — foreground, midground, and background are distinguished only by light and dark, which sets the imagery apart from the strong-depth-of-field films on the market. (Of course, the drawback is a weaker sense of spatial depth)

Second, no solid-color overlays above the line art. Specifically, the black contour lines stay the darkest, never lightened by overlaid light and shadow, so character silhouettes remain clear, avoiding so-called “light pollution”;

Purely in terms of color, the top one is indeed more “airy.” But compositing doesn't serve a single shot — such floaty colors feel somewhat “ungrounded,” mismatched with the whole film's tone; what was actually needed was richer color. On the other hand, this black-furred character is backlit, and stacking too much “glow” would contradict the storyboard design.

That said, not overlaying bright colors doesn't mean the frame can't be brightened at all. The light and dark in a frame are relative: besides simply increasing brightness to make whites whiter, you can also deepen the grays to make whites seem whiter — the latter is what I use now.

Of course, the two methods aren't superior or inferior to each other; each has pros and cons. Pick one or mix them — whatever works.

Two, revisions for 3D layers

For compositing containing 3D layers, the process is generally: 1 — rough-version assets; do the first camera move according to the storyboard.

2 — On screenshots of the rough composite, paint a correction color sketch.

3 — Based on the corrections, readjust positions and motion, repaint the correction color sketch.

4 — Continue adjusting assets and refining, then composite and export normally.

This process was inspired by some 3D films, where many models, after their motion is keyed, are handed back to key animators for correction, and then keyframes are revised accordingly.

After all, there's a big difference between 2D design layouts and three-dimensional motion — it's hard to predict the result before placing the assets. This process eases the compositing burden; with a physical reference, you don't have to stare at assets and imagine what the effect should be.

Three, layer adjustments

Since this film doesn't use “depth-of-field blur” but distinguishes foreground, midground, and background by luminance, after finishing normal compositing I generally re-check the layers.

The thinking here is the same as in illustration: convert the frame to black and white, checking purely for light and dark.

· Single-layer revision

Taking this shot as an example: the leftmost is the first version, conceived as far-to-near, light-to-dark — the distant woods brightest, the foreground trunks darkest. Sounded nice, but it ignored the existing light source: the foreground trunks are side-lit, so they can't possibly be the darkest. First version rejected.

The middle is the second version, revised to the idea of darkest in the distance, brightest in the foreground. But in practice, the foreground trunks blended into the person in the middle; the foreground trunks needed further brightening.

Yet during revision I found that if I brightened the foreground trunks overall, their shadow side became too bright, blending with the midground person. Finally I chose to increase contrast: ensuring the foreground trunks' highlights were the brightest, and their shadow side darker than the midground's shadow side — that finally stopped the blending.

Though foreground and midground were now separated, the layers still weren't very distinct: on one hand, the foreground shadows, the midground horses, and the distant shadows had too-similar luminance; on the other, the left and right sides of the frame were too evenly lit.

· Overall revision

The revision approach is somewhat similar to current illustration techniques: add “buffer zones” at the contact points between different depth planes, lightening or darkening around contours to make the foreground silhouettes clearer.

For example, the horses' legs are closest to the foreground bushes, and their luminance clashed with the foreground shadows — so a light color was masked over the horse legs, and the luminance difference separated the two.

Above the legs, the horse's upper body's luminance was actually unchanged, so this fine adjustment doesn't affect the midground's overall luminance.

Overlaying bright colors on the right side of the frame follows the same logic: by design, the storyboard already had the light coming from the right, so lightening the right side is logical.

In practice, two solid-color layers were overlaid, one below the foreground layer and one below the midground layer.

Originally, “foreground shadows,” “midground horses,” and “distant shadows” had similar luminance; after overlaying the solid colors, the midground was lightened once and the distance twice, numerically pulling apart the three layers' luminance differences.

Four, adding brushstrokes in post

Strictly speaking, this isn't really about compositing — I truly had no other option; failing to achieve this effect in post was my own shortcoming.

This method means taking the already-composited footage, importing it into Photoshop frame by frame to repaint, then exporting the PNG sequence back into After Effects.

The treasure, the feathers, the bird on the hand, the explosion, the crowd transition, and the falling objects — these six parts all used this kind of repainting.

(2) Benchmarking shots

During the compositing stage, most shots' effects had basically come together, then I hit a bottleneck again. Every shot's progress was roughly equal — where to start?

Naturally, by benchmarking against shots from other excellent films. With a rough effect in place, the gap becomes intuitively visible.

This is much like drawing from life: you draw for hours thinking it looks great, take a photo with your phone, and immediately think it's ugly. Focusing your gaze too much on yourself easily becomes limiting; judging an imperfect work alongside excellent works can ruthlessly reveal the direction of progress.

Take B3-1, for example — viewed alone, I quite liked its effect. But once animated, as a wide shot it felt rather empty. So I found real underwater photos, checked what the frame lacked, and adjusted the assets accordingly, revising continuously.

A more detailed introduction to the frame adjustments was covered in my earlier Weibo post “Fanfan Xiangsheng Behind the Scenes (4 supplement) — Failure and Improvement” — kindly head over there.

(3) Compositing checks

This content is also covered in the Weibo post “Fanfan Xiangsheng Behind the Scenes (4) — From Layout to Compositing.”

Light pollution in compositing is something I strove to avoid. To prevent it, the usual practice is to keep the light-source layer from being placed above the character layer; overlaid light sources blur the original boundary between characters and background, destroying silhouettes and turning everything into mush.

If a light source must be overlaid on characters, there's a way to check whether silhouettes are destroyed: “black and white”:

​(Set saturation to 0, or just apply a black-and-white effect) Checking the pure value relationships directly makes it easier to see whether the composite has muddied the assets, and after desaturating you can re-check the shot's layers, adjusting the foreground, midground, and background luminance according to the value structure.

(4) Music harmonies / re-syncing

After benchmarking against excellent shots, many shots gained new animation and transitions, so some shots' durations changed.

For the music, first adjust the rhythm — adjusting the pauses according to shot lengths — then add sound effects matching the newly added visual effects.

Once the rhythm was roughly adjusted, harmonies could be added for effect. Mainly, the storyboard work had hit a bottleneck and I didn't know what to refine — so I worked on music to switch my brain.

During this period, shot durations and pause positions were continuously adjusted — refining while adjusting — and before I knew it, progress moved forward.

(1) Music

[2023.8–2023.11] Early stage

At first, mainly确定 the melody — humming or a keyboard both work — and settle the rhythm to sync with the storyboard. Once the rhythm is roughly set, embedding into the storyboard can begin. In the first four months, i.e., August–November 2023, the melody came intermittently, interspersed among many other tasks — marginal work not occupying dedicated time blocks.

[2024.12–2025.1] Mid stage

On one hand, writing lyrics — since the rhythm after adding lyrics differs somewhat from the hummed melody version, the storyboard lengths were revised simultaneously. On the other hand, enriching the melody and adding sound effects that matched the visuals; the harmony melodies were roughly settled around this time.

However, due to my rather thin music theory knowledge, about half the harmonies sounded terrible — if they could be fixed, fix them; if not, leave them out. That's why some places in the short are harmonized and others left empty.

There were originally drums in the draft, but since I kept missing the beat, I cut the drums too.

[2025.5–2025.7] Late stage

This part was mainly adding lots of unison vocals so the melody wasn't so monotonous. The rhythm kept being adjusted, usually together with the storyboard. Occasionally frames needed adding — just extend the vocal section a bit; anyway, the music follows the shots.

There's also the fact that in July, crying, I added the bass drum back — otherwise it was too empty. And so, adding the bass drum changed the rhythm again, so I had to grit my teeth and adjust the storyboard durations and rhythm sync, then squeeze in some sound effects wherever possible.

(2) 3D modeling

The reason for incorporating 3D is that in some respects it's more efficient. As mentioned in Behind the Scenes (1): objects with many angle changes and no deformation are chosen for modeling.

In other words, for parts with deformation — character motion, effects — modeling isn't necessarily more efficient. Even if a character is fairly realistic and the design is solid, looking like a model, that doesn't mean 3D could easily achieve the same effect. 3D isn't easier than 2D for adjusting subtle movements; sometimes it even requires repainting over the render by hand to calibrate the motion and make it more natural.

Some people think rotoscoping, 3D-rendered-2D, and the like are “faster, better, stronger,” far superior to hand-drawing. Indeed, for fast-food animation, 3D has the upper hand; but if you want to make something good, there are no shortcuts on any path.

A somewhat inapt example:

Generally, in modeling software, a shot's focal length is uniform. If you want a motion of “a character facing the camera, both arms extended forward, with big distortion on the foreground arm,” you'd usually keep the current focal length and enlarge the foreground arm — looking like Popeye from the side — so that the character's body has no distortion while the foreground shows strong perspective.

Whereas 2D animation just requires drawing what's in your head — quick, intuitive adjustment is 2D's advantage, though it correspondingly demands stronger drawing fundamentals.

For creators, more skills never hurt: absorbing the essence of each technique without infringing on others' rights, reducing redundant work — that's how you can better devote yourself to creation.

From the charts, apart from the final shot, each shot took an average of about 11 days of work. This work time only covers the animation portion, not backgrounds, color scripts, or early conception.

The final shot (B8-1_10-1) took longer because it had an idea that absolutely had to be realized; most of the time went to inking and coloring. Outsourcing the inking and coloring would have saved a lot of time, though the final result would have been somewhat compromised.

The work times also show that shots have primary and secondary roles: relatively normal shots used to connect the story take less time (A0-3, A4-3 — walking, standing, etc.), while shots intended to add more animation take longer (A5-1_7-2, A12-2 — blooming flowers, multi-layer transition camera moves).

Since this film's purpose was art and animation, purely connective shots were sketched through quickly, concentrating more energy on individual shots for greater effect. Though that's the theory, the effect wasn't obvious — my control of pacing still isn't quite there. I'll do better next time.

02 Supporting Work​

1 — How to summarize? When to summarize? What to write in the summary?

—— Work done that month, work rate, estimated deadlines. Monthly feelings and main problems.

2 — The meaning of end-of-month summaries and daily work records

—— The feedback from these behaviors isn't immediately available. The purpose is to make later retrospectives easier — meaning it only helps if you'll make animation / long-term projects again.

Daily work records make it easier to have an accurate picture of past work. Without records, recent emotional swings distort your impression of past months.

For example, if you were highly efficient for 28 days of a month but got seriously ill in the last 3 days, without records you'd likely feel at month's end that you did nothing all month, crushing morale. But with records of the month's work, you can objectively see the actual number of highly productive days, reducing the influence of momentary emotions.

The end-of-month summary is a more macro statistic compared to daily summaries — essentially no difference; it's mainly an overall picture of the whole month. When the project ends and you do the final retrospective, just read the monthly summaries — no need to flip through every day's table. It's for convenience.

Record at all times, and consolidate everything into one document; during retrospectives you can fill problems in accordingly (What a hollow answer — but that's just how it is)

For estimating schedules — first fully refine every part of one shot, tally the shot's content and required time, assign difficulty values to each part and calculate a drawing rate (see speed records), then multiply this rate by the rough workload of other shots to get an estimated time per shot. Calculating in turn yields an estimated time for the whole stage.

Rate estimation and per-part time tallies are both relatively tedious; besides, generally after calculating a few shots you have a feel for it, so in theory you don't need to calculate every shot precisely — once you have a sense of your drawing speed, direct estimation works.

For how to choose shots to estimate — pick one shot each of different difficulties and different types. Then, once you estimate that shot's time, you can estimate the time needed for shots of the same type.

But people aren't machines — simply calculating drawing speed isn't enough; rest time after high-intensity work also needs to be factored in.

03 Mental Preparation​

First, let's clear up a misconception: that if a person has enough passion, they can create month after month as if on a caffeine high.

Completing a long-term work doesn't rely on a single continuous motivation. So-called motivation works like a relay baton — it plays its role in a certain phase; when it no longer applies, find another motivation to replace it, and sometimes you even survive on muscle memory alone.

To explain more specifically: motivation divides into [long-term motivation], [short-term motivation], and [snack-like motivation].

[Long-term motivation] is something like “finish this work.” Its sense of achievement is strong, but for a long-term work it's too grand, and it basically only stands out at the very beginning and end of the project. Along the long road in between, it's just a white dot in the distance.

That's when you need the boost of [short-term motivation]: cut a grand goal into countless small goals, and the motivation to complete each small goal counts as short-term motivation — like “finish this shot's storyboard,” “finish inking this shot.” Short-term motivation is indeed more immediate, but the problem is that because goals are cut so fine, the sense of achievement after completing them is greatly discounted. Sometimes, though you complete a small goal, because too many small goals still await me, motivation is actually lost.

Simply setting goals doesn't seem able to sustain steady creation. So to keep going, either increase the satisfaction of completing each task, or reduce the pain of working. I chose the latter, and the method is simple: [form habits]. Make completing fixed daily work the norm — like clocking in at a job, completing the things that need doing one by one. Though there's almost no satisfaction in this, the pain is also much less.

How exactly to form habits varies from person to person. I generally choose “get the wheel turning first, then speed it up”: start with small things so simple they can't fail, claim the time first, then as the habit forms, gradually replace the simple small tasks with real work.

▼ An inapt example (rough order of motivations): [Long-term motivation] At first I wanted to make an animated short — I fantasized about the finished short, about my improvement, about everyone's praise.

[Short-term motivation] So I started drawing — some sketches and storyboards. Hey, feels promising.

[Daily habit] Then refined the storyboard per the process, mechanically completing repetitive work.

[Snack motivation] Early designs were mostly done; drawing felt a bit tiresome — watch some excellent works for a jolt of energy.

[Short-term motivation] To switch my brain, I began experimenting with shots; the results were decent — motivation renewed.

[Daily habit] Continued mechanically refining shots.

[Snack motivation] Mechanical refinement was so boring — watch some excellent works for a jolt of energy.

[Daily habit] Continued numbly refining.

Roughly that kind of cycle. Since I don't show anyone the work until it's finished, I may lack externally supported motivation; if you share as you go and have others' support, there may be far more motivation than in the process above.

(1) Anxiety / numbness

One's emotional state always comes in waves — sometimes extremely sensitive and anxious, sometimes numb and lazy, not wanting to do anything. For these two states, I generally have two opposite solutions.

When emotionally sensitive, I choose to be a turtle hiding in its shell: block keywords, uninstall apps, deliberately filter what I consume. Anxiety-peddling content is strictly off-limits; instead, search for stories of people who sharpened a sword for ten years or achieved greatness late in life.

When numb or sluggish, I take out the anxiety-peddling content I previously dared not look at to provoke myself: whatever you fear, look at that; whatever you feel inferior about, look at that. Pressuring yourself can actually fill you with motivation.

This doesn't mean becoming closed off, only seeing what you want to see — it's just to let yourself focus more on production. Being able to choose what suits you in different states is, in itself, a good habit of getting along with yourself.

Also, when gazing at the future, don't forget the road you came by. Why did I originally choose to make this film? Why this format? These were written down at the start. But a year blurred the direction; when lost again, I re-read the original outlook, which eases the confusion a little.

There's also [writing out the things that make you anxious, then reading them aloud] — don't just circle them in your head. When problems are written on paper one by one, the main contradictions become clear. Sometimes, the moment you open your mouth to read aloud, you find some problems utterly childish and laugh yourself angry. (For fighting emotions only)

I know we still have many shortcomings — lacking scripts, funding, talent. But I believe that when all aspects accumulate to a certain point, good works will sprout like bamboo shoots after the rain. All I can do now is make works well, success or failure.

Besides, troubles related to creation are happy troubles; life's trivia is the real mess. On that front, I have no solution — just push through.

(2) Epilogue

During production I always said, “It'll be fine once the film is done.” It seemed that once the film was finished, all other problems would solve themselves and good things would follow. I thought that after finishing I'd feel relief, at least level up a step. But once it really ended — nothing. Not happier, no leap in progress.

In fact, the end of production broke a rhythm maintained for over a year, replaced by intense maladjustment. With no fixed work each day, I constantly felt I was wasting time. Driven by unease, I urgently wanted to find the next job, but panic-driven choices are rarely reliable — throwing myself in headlong, only halfway discovering it didn't suit me at all.

Also, don't keep thinking that focusing on one thing justifies postponing everything else. Delays breed changes; many things, if not solved, won't solve themselves. Otherwise, when the main work is done and there's no excuse left to postpone, you'll have plenty of messes to clean up. (Slap my own face — smack) (Ouch — covering face)

The point of saying all this is to remind myself not to exalt “completing something” too much — the greater the hope, the greater the disappointment. Adjust your mindset after finishing, and only then can you set off better.

A task's rewards don't necessarily pop out suddenly like a boss rewarding you upon completion; they arrive in a way that moistens things silently, imperceptibly transforming you. If you take each tiny reward as a unit, each time feels no different from before. Only when the time scale is stretched long enough, and you look back, will you marvel at how far you've come.

Having said all this — what comes next? Hmm… I don't know either. The things I want to do are just those; I'll do whichever fits. So I'll just walk vaguely in the direction I like, one step at a time.

(3) The freedom to give up

Though I've spouted a lot of platitudes above, making it seem everything must be seen through to the end, in truth only the kind of hardship where [it's painful or confusing, but this way is probably fine?] is worth tasting. If the pain reaches “feeling nothing” or “losing the ability to judge,” giving up at the right time is better.

Works that make it to audiences' eyes all carry survivorship bias. As for “giving up,” even I personally have quite a few: some concepts progressed even into mid-production but were ultimately not pursued. Some couldn't proceed due to technical limits; some ended half-finished due to insufficient experience; and others, when I repeatedly asked myself “is this meaningful?”, received a negative answer.

In short, not everything requires grinding to the end; only things that speak to your heart deserve persistence — and of course, the premise is knowing what you want. (Won't expand on this… though I really don't have much more to write anyway… haha…)

04 Conclusion​

This rambling account ends here. With my shallow experience, I can't present anything more profound. This article set out from the perspective of independent production, hoping to offer creative ideas.

Creation is far less mysterious than imagined. The best time to plant a tree was ten years ago; the second best time is now.

Just start creating! And creation will begin…

Wishing everyone happy creating…