NIGHTCELL 7: When the Ravens Lied, a second Þrøngva album made in code
NIGHTCELL 7: When the Ravens Lied, a second Þrøngva album made in code
NIGHTCELL 7：《When the Ravens Lied》，一张由代码生成的 Þrøngva 第二张专辑
NIGHTCELL 7 has a second soundtrack album: When the Ravens Lied by Þrøngva, twelve new tracks in the same drop-D Viking metal and hip-hop style as After the Winter of Want. They are in the game now, next to the first album’s twenty. NIGHTCELL 7 推出了第二张原声专辑：由 Þrøngva 创作的《When the Ravens Lied》。这张专辑包含十二首新曲，延续了《After the Winter of Want》中 Drop-D 调弦维京金属与嘻哈融合的风格。目前这些曲目已加入游戏，与第一张专辑的二十首曲目并列。
Who this is for: Anyone playing the free sandbox, in the browser or on the desktop client. Nothing here costs money. 适用人群：任何在浏览器或桌面客户端游玩这款免费沙盒游戏的用户。所有内容均免费。
What the album is about: Odin sends two ravens out every morning and believes what they bring back. On this album somebody has taught them to lie. Forged smoke and forged voices tell two halls that the other one struck first, and a merchant sells spears to both sides. It is the game’s MIRAGE plot told as a saga. No real country is named, and neither side is the villain: the lie is, and so is whoever profits from it. 专辑主题：奥丁每天早晨都会派出两只乌鸦，并深信它们带回的消息。但在本专辑中，有人教会了它们撒谎。伪造的烟雾与声音告诉两个大厅，对方才是先动手的一方，而一名商人则向双方兜售长矛。这是游戏“MIRAGE”剧情的史诗化叙事。专辑中未提及任何现实国家，双方也并非反派：真正的反派是谎言，以及那些从中获利的人。
The last track, “True Dawn”, answers the episode title, FALSE DAWN. 最后一首曲目《True Dawn》回应了剧集标题《FALSE DAWN》。
Huginn Brings the Word, Who Fired First, Loki’s Mirror, Merchant of the Long Knife, Ledger of Ash, Heimdall Does Not Sleep, Same Rain on Both Shields, Lower the Spear, Muninn Remembers, Longship Bounce, Not Our Ragnarök, True Dawn. （曲目列表：Huginn Brings the Word, Who Fired First, Loki’s Mirror, Merchant of the Long Knife, Ledger of Ash, Heimdall Does Not Sleep, Same Rain on Both Shields, Lower the Spear, Muninn Remembers, Longship Bounce, Not Our Ragnarök, True Dawn。）
How it was made: Every note is synthesised by a Python script with numpy and ffmpeg. Nothing is sampled, and no AI audio model or music service was used. The script reads a lyric sheet that gives each section a tempo and a direction (“male baritone half-rap, drop-D riff with pockets of silence”) and arranges it. It plays a riff with the kick doubling the guitar, boom-bap or 808 drums on the hip-hop tracks, and choir, strings, tagelharpa, a throat-singing drone and horns. It also adds the wind, rain and ravens the sheet asks for. 制作方式：每一个音符均由 Python 脚本配合 numpy 和 ffmpeg 合成。没有任何采样，也没有使用任何 AI 音频模型或音乐服务。脚本会读取一份歌词表，其中为每个部分设定了节奏和指令（例如“男中音半说唱，带有静音间隙的 Drop-D 连复段”），并进行编排。它会演奏连复段，底鼓与吉他同步，嘻哈曲目中加入 Boom-bap 或 808 鼓点，并配以合唱、弦乐、塔格尔哈帕琴（tagelharpa）、喉音吟唱的持续音以及号角。它还会根据歌词表的要求添加风声、雨声和乌鸦声。
The voice is the honest limit. It is formant synthesis, one note per syllable on that syllable’s vowel, so it follows the rhythm of the lyrics without being words you can make out. The lyrics ship inside every MP3’s tags, and the repo has paste-ready sheets for a sung version later. 人声是其诚实的局限所在。它采用共振峰合成技术，每个音节对应一个音符，并落在该音节的元音上，因此它能跟随歌词的节奏，但无法听清具体词汇。歌词嵌入在每首 MP3 的标签中，代码库中也提供了可供后续演唱版本使用的歌词表。
Nobody listened to the first render, because the session that made it cannot hear. A spectrogram and an energy count showed 73% of the energy below 150 Hz, which is mud, so the mix was rebalanced before all twelve tracks were rendered. Each one is a 320 kbps, 48 kHz MP3 at about -10 LUFS. 没有人听过最初的渲染版本，因为生成它的会话本身无法“听”。频谱图和能量统计显示 73% 的能量集中在 150 Hz 以下，这会导致声音浑浊，因此在渲染全部十二首曲目之前，混音进行了重新平衡。每首曲目均为 320 kbps、48 kHz 的 MP3，响度约为 -10 LUFS。
How to use it: Open nightcell7.com/play and turn your volume up in the yard. Both albums play shuffled, one track after another, and stream as they are needed, so they add nothing to the download before your first match. The lyric sheet and the synthesiser are in the repo at github.com/profullstack/nightcell7. 如何使用：访问 nightcell7.com/play 并在庭院场景中调大音量。两张专辑会随机播放，曲目循环衔接，并按需流式传输，因此在进行第一场比赛前不会增加额外的下载负担。歌词表和合成器代码可在 github.com/profullstack/nightcell7 获取。