Soulframe Liang on the Work Behind Xia: Creative Choice, Studio Fit, and Staying in the Game

1. Guest Background

Soulframe Liang, identified in the guest background as the producer of Phantom Blade Zero, speaks with host Zhang Xiaojun in an interview indexed at more than six hours. The discussion moves between martial fiction and the organization required to produce it: what makes an action heroic, why a studio changes business models, and how a recognizable creative identity survives financing, management, and technological change.

Liang brings experience from an unusually gradual career transition. He recalls studying architecture at Tsinghua, losing an already secured postgraduate recommendation after failing Engineering Economic Analysis, and completing a fifth undergraduate year before studying architecture at Yale. He resists treating that interruption as the moment he discovered his vocation. Making games became a serious career possibility through roughly five years of changing expectations, encouraged by player feedback and several tens of thousands of dollars from student-era independent games. Although he had obtained an offer from architecture firm KPF, he chose to return to China in 2011 and start a games business.

His creative interests also explain the interview's emphasis on construction. Liang enjoys elaborate conspiracies and interlocking plots, but says he eventually understood that their structure must emerge from character motivation. He acknowledges the corresponding weakness: attention spent engineering a clever framework can leave characters and emotions underdeveloped. The episode is therefore valuable as both an account of artistic intention and a discussion of the limits of his preferred methods.

2. What the Episode Covers

The studio history challenges any simple equation between effort, quality, and commercial reward. Liang recalls Xu Xiaoping investing RMB1 million into his personal account before the company existed, and early contributors working for approximately RMB3,000 a month on Rain Blood: Mirage. He regards that game as an exhaustive use of the team's available money, people, energy, and creative ability. He nevertheless says it did not make money. In his account, completing a work to the best of one's ability and building a commercially successful product were separate achievements.

Mobile development produced the opposite lesson. Liang says NetEase's 2013 investment and publishing payment supported the first Phantom Blade mobile game, released in 2014. It reused Mirage assets at reduced specifications and simplified interaction; its combo-chain design answered the limitations of touch controls. He recalls initial monthly gross receipts in the tens of millions of renminbi, while elsewhere calling the RMB10–20 million received by the team each month profit. Gross receipts, revenue share, and profit remain distinct concepts despite their overlap in his spoken recollection.

He credits those years with teaching commercial judgment and project management. Yet around 2015–2017, expansion, roadshows, broader IP plans, and capital-market ambitions absorbed attention; he recalls sometimes doing development only one day a week. He describes Phantom Blade II's pursuit of mobile esports and anime-oriented styling as a departure from the team's strengths, and says it produced no profit. The third game reduced scope and returned to PVE and direct creative involvement. Liang reports first-month gross receipts of RMB150 million and says its profits funded early-shareholder buybacks.

Returning to a major single-player project required more than renewed enthusiasm. An attempt in 2017–2018 had been cancelled over costs, cash flow, and insufficient resolve. By late 2022, Liang judged that existing material, money, partnerships, and the momentum surrounding Black Myth: Wukong made Phantom Blade Zero feasible. The team still had only side-scrolling 2D experience, with no completed 3D project behind it. Project cancellations, staffing changes, recruitment, technical gaps, and uncertain financing had to be addressed. Trailers and hands-on events, he argues, helped stabilize morale through outside feedback.

Liang discloses an initial Tencent investment exceeding RMB400 million for a stake above 20%, approximately US$30 million in subsequent project financing, mainly convertible debt, and project spending of RMB400–500 million. He describes safeguards and buyback provisions without revealing complete terms. He also rejects the suggestion that Sony became cold toward the project: second-party discussions had progressed, and support continued after the studio chose self-publishing.

Cultural ambitions appear in production choices as well as financing. Liang describes recording southern lion-dance performances and drums at a Guangdong ancestral hall. Donnie Yen contributed advice on action and dramatic scenes, resisted an elderly, solemn master stereotype, and performed swordplay in modern leather clothing under red neon lighting for promotional material.

3. Core Views: Reasoning, Examples, and Limits

Liang's interpretation of xia, the martial hero, begins with a distinction between motivation and moral recognition. Hatred and injustice drive a story; chivalry becomes visible in what a person does under that pressure. His example is an older person collapsing amid traffic. Helping involves physical danger and the possibility of being blamed, yet someone acts anyway. The consequential element is a choice made with an understanding of the cost. Exposure to danger does not require death as proof of virtue. This is Liang's account of narrative ethics, rather than a general instruction to seek hazardous situations.

He connects that ethics to a particular fictional environment: ordinary order has failed, people on society's margins must act, and their personalities are often extreme. The combination explains wuxia's identity for him, although it cannot serve as a complete definition of every work in the genre. His aesthetic argument is similarly relational. In New Dragon Gate Inn, he remembers the magnificence of the desert alongside the inn's cruelty and a disturbing image of a hand stripped to bone. Beauty gains force through its distance from darkness. Since he also understands modern wuxia as having absorbed different literary, cinematic, and cultural methods, inheritance need not mean reproducing an established surface.

For Liang, contemporary renewal must also change why characters act. Comparing his 2006 Rain Blood work with current production, he sees a movement toward growth, survival, and inner consistency. His explanation of wuxia punk places personal expression inside inherited martial resources. A character need not set out to obtain the title of hero; other people can recognize heroism in the consequences of particular choices. His account of changing audience sentiment remains an interpretation, however, rather than a survey of an entire generation.

A trailer mistake gives that argument a practical anchor. Liang says the team inserted explicit declarations about protecting ordinary people and behaving chivalrously, hoping to create additional emotional resonance. Players instead perceived an attempt to claim moral prestige, and the language conflicted with a cold, dark assassin organization. He estimates that roughly 80% of the relevant players rejected it, but supplies no sampling method. The defensible lesson concerns coherence between characterization and expression; the percentage should not become a general audience statistic.

Soul and Zuoshang complicate moral labeling further. Liang describes Soul as intensely determined to survive despite having only 66 days left: he intends to save himself and ultimately saves the world. Zuoshang intends to save the world and brings destruction, yet Liang does not reduce him to a villain. That reversal separates declared intention from consequence. The conversation develops a possible reading about grand projects of salvation causing catastrophe. It remains one interpretation formed during the exchange, not the sole confirmed meaning of the story or a universal claim about political ambition.

The same attention to fit shapes Liang's business argument. He identifies free-to-play, continuously operated mobile games as difficult for his team because they demand sustained update rhythms, production control, and enthusiasm over years. He acknowledges creative work within cross-platform service games and rejects a hierarchy in which single-player games are morally superior. The question is whether a particular team can sustain the work. His preference for direct making, teaching craft, and project control follows from that judgment. Control also carries responsibility: the producer must answer for a collective work. Liang admits that anger, minute demands, and repeated changes can exhaust colleagues and create a toxic environment. His willingness to reverse bad requests is significant, but the interview does not establish that acknowledgment has resolved those organizational costs.

AI introduces another distinction: efficient production and willingness to pay. Liang says technical tasks already receive AI assistance, while Phantom Blade Zero's artistic creation had not used AI at the time of speaking. He believes purchases also express trust in people, their investment, and future work. No research in the episode quantifies that contribution. Nor does he rule out new AI-assisted authors: consistent personal style could allow audiences to attribute an experience to a creator rather than to a tool. His proposed signal—a widely recognized AI short-drama director—is a future observation, not an established development.

Market opportunity offers no final guarantee. Liang says the original IP still needs validation and refuses to call AAA development an uncomplicated blue ocean. His Steam comparison, Chinese players accounting for 25% of spending and Chinese companies receiving 5% of revenue, lacks a stated year, source, or methodology. More durable than those figures is his operating principle: money, experience, technology, and partnerships should improve the next game. That requires a continuing balance between creative work and external business activity.

4. Learning and Application

For writers, the episode suggests beginning a moral scene with a chain of decisions. Establish the pressure, the risk the character understands, the personal reason to act, and the effect on someone else. Then decide whether an explicit declaration adds anything. The trailer example is especially useful when dialogue sounds admirable but contradicts the institution or character delivering it. Liang's structural interests suggest a second pass: connect each major turn to a person's motivation. A plot can be intricate yet emotionally unearned. His idea of having AI map a novel's structure is an experiment to consider, not a demonstrated solution; a connected diagram would still not establish believable feelings or relationships.

Visual teams can ask equally specific questions of style. Liang explains that bright bamboo settings and dark environments serve different emotional functions within the whole. A review can therefore ask what a location should evoke—pressure, uncertainty, recollection, or release—before debating brightness in isolation. Weapons provide another practical example. Zuoshang's gold sword, strengthened by fictional killing energy, communicates luxury and confidence. Soul's Heishang draws on a Tang-style horizontal blade to express a rougher, grounded martial identity. The transferable method is to align silhouette, action, and personality. It does not require copying those designs or treating their fantasy logic as historical weapons evidence.

For production leaders, authorial control should be paired with visible correction. When requesting an expensive change, specify the intended player experience, the existing work affected, and the conditions under which the request should be withdrawn. Preserve a way back where an experiment is uncertain. These are practical deductions from Liang's account of repeatedly reversing demands, rather than a management system whose effectiveness the episode proves. Review attention alongside money: a growing portfolio of meetings, pitches, and incubated projects may leave the people responsible for craft with little time to practice it. Credits offer a complementary way to recognize contribution. Liang's account of former employees asking to be included shows why attribution matters, while his discussion of pay, career returns, and local alternatives prevents recognition from becoming a substitute for material conditions.

Publishing and financing require separate comparisons. A self-publishing assessment can list media access, event execution, overseas gaps, budget, and the creative value of controlling communication. Liang's estimate of a 20–30% overseas shortfall cannot be imported into another project's forecast. Likewise, company investment, project financing, convertible debt, and expenditure should occupy different lines in an analysis. Adding them together does not produce a development budget, and undisclosed safeguards cannot establish an independently verified absence of risk. Historical motives deserve similar care: Liang says NetEase's exit from early investments and an attractive price helped prompt the buyback. The eventual single-player direction should not be projected backward as an already settled rejection of capitalization.

AI decisions benefit from tracking production efficiency, recognizable expression, and purchasing trust separately. Better output or faster delivery does not by itself answer whether an audience values the author or anticipates another release. Teams can look for those distinctions in feedback and actual buying behavior without assuming either that handmade work necessarily commands a premium or that technically impressive generation necessarily sells. The longer-term test is what each commercial cycle leaves behind: reusable skills, stronger production capacity, and room to make another work. Liang's standard of sufficient expression sits alongside his warning that access to market resources does not establish world-class professional ability. For a team entering global competition, that means evaluating creative intent and concrete weaknesses together, rather than allowing financing scale or national identity to stand in for evidence of quality.

Source

• Original episode: "Xia" | A 6-Hour Interview with Soulframe Liang, Producer of Phantom Blade Zero