Compare commits
22 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
| 2d0aaa9dd8 | |||
| 05d5af3cb7 | |||
| b2b8abf397 | |||
| 8548c8eef3 | |||
| 5ce8d717e7 | |||
| a86c6bc59a | |||
| 92786dbbf0 | |||
| 91b08e189a | |||
| 8ac49634d7 | |||
| a9cc99c1ef | |||
| 90f1e5235d | |||
| 07f985c393 | |||
| 637c4d8ae3 | |||
| cd8fc1ab2a | |||
| 8b367fa780 | |||
| c74698d75b | |||
| 60aeff3f2c | |||
| d735b3429f | |||
| 53c817f046 | |||
| e403754426 | |||
| 6ef6aaa352 | |||
| 947c3082ea |
-297
@@ -1,297 +0,0 @@
|
||||
# digital signal processing
|
||||
|
||||
A boring lecture about signals and systems, and its digital implementation in DSP.
|
||||
Borderline 'song' less 'poetic'.
|
||||
|
||||
The actual material (text that is read out by a vocal synthesiser) is in Japanese.
|
||||
|
||||
## Simple notes
|
||||
|
||||
For things that are not easily pronounced (e.g. formulae, expressions, symbols), inclusive either:
|
||||
- Mark how the thing should be pronounced in a pair of adjacent parentheses
|
||||
- e.g. $H(\omega)$ (エッチオメガ)
|
||||
- (inclusive) Or don't embed these elements too deep into the flow of sentences; use common, laymen's language instead
|
||||
- e.g. not `Fs/2(サンプリング周波数の半分)までの帯域を忠実に再構成できる` but `サンプリング周波数の半分までの帯域を忠実に再構成できる`
|
||||
|
||||
The general direction is to stay boring, but the deadpan kind of boring. Keep technical language appearing but explain things in more laymen's language.
|
||||
|
||||
## Outline
|
||||
|
||||
### 1. Signals and systems
|
||||
- What is a signal
|
||||
- A function of one or more variables carrying information; here, amplitude over time
|
||||
- What is a system
|
||||
- A system maps an input signal to an output signal
|
||||
- Characterized entirely by its behavior on inputs — the internals are not required
|
||||
- Most systems are nonlinear — linearity is a special property, not the default
|
||||
- Linearity: superposition holds — scaling and additivity
|
||||
- Most physical systems violate this under sufficient conditions (e.g. overdrive: linear below clip threshold, nonlinear at and above it)
|
||||
- Time-invariance
|
||||
- A time-invariant system responds identically regardless of when the input arrives
|
||||
- Violation: any system with time-varying parameters (e.g. modulated effects)
|
||||
- LTI as the restricted, studied case
|
||||
- Satisfies both simultaneously — analytically tractable
|
||||
- A deliberate simplification; not a description of typical reality
|
||||
- The impulse signal; impulse response
|
||||
- The impulse: unit energy concentrated at a single point in time
|
||||
- The impulse response: what an LTI system outputs when fed an impulse
|
||||
- Completely characterizes the system
|
||||
- The impulse response is a function — it exists in the same space as signals; systems are absorbed into signal space
|
||||
|
||||
### 2. LTI and convolution
|
||||
- An LTI system is fully characterized by its impulse response
|
||||
- Any input decomposes into scaled, shifted impulses
|
||||
- By linearity and time-invariance, the output is the corresponding sum of scaled, shifted impulse responses
|
||||
- Convolution as the operation
|
||||
- That sum is convolution: $y = x * h$
|
||||
- In continuous time, this is an integral: $y(t) = \int x(\tau) \, h(t - \tau) \, d\tau$
|
||||
- Symmetric in its arguments: $x * h = h * x$
|
||||
- Signal and system are interchangeable operands — both are functions; convolution makes no distinction between them
|
||||
- Commutativity: for LTI systems, order of composition is irrelevant
|
||||
- Two LTI systems in series have a combined impulse response equal to the convolution of each
|
||||
- $h_1 * h_2 = h_2 * h_1$ — order does not affect the result
|
||||
|
||||
### 3. Signal transformation
|
||||
- A signal can be represented equivalently in the frequency domain
|
||||
- The Fourier transform: $s(t) \leftrightarrow S(\omega)$; any signal is a sum of sinusoids at different frequencies
|
||||
- Each sinusoid has a frequency, amplitude, and phase
|
||||
- $S(\omega)$ is the spectrum: how much of each frequency is present
|
||||
- Analog signals carry components from DC to arbitrarily high frequencies
|
||||
|
||||
### 4. Filters
|
||||
- A filter is an LTI system — characterized by its frequency response $H(\omega)$
|
||||
- $H(\omega)$ is the Fourier transform of the impulse response $h(t)$
|
||||
- Filtering: shaping a signal's spectrum by selectively attenuating or preserving frequency bands
|
||||
- Basic types: lowpass, highpass, bandpass
|
||||
- Lowpass: passes low frequencies, attenuates high
|
||||
- Highpass: passes high frequencies, attenuates low
|
||||
- Bandpass: passes a target band, attenuates above and below
|
||||
- In discrete time: finite impulse response (FIR) and infinite impulse response (IIR) filters
|
||||
|
||||
### 5. Discretization
|
||||
- Moving from continuous to discrete: two axes require discretization
|
||||
- Amplitude: $N$ bits → $2^N$ discrete levels
|
||||
- Each sample is mapped to the nearest representable level; the error is quantization noise
|
||||
- The representable range is fixed by design and hardware (ADC/DAC reference voltage)
|
||||
- Hard clip at the boundary — analog degrades gradually toward its physical limits; digital does not
|
||||
- Time axis: sampling captures only finitely many values per second
|
||||
- Analog signals contain frequency components up to infinity — a discrete system cannot represent all of them
|
||||
- Discrete time is viable when the high-frequency components that would be lost are not needed
|
||||
- Nyquist: given sampling rate $F_s$, faithful reconstruction is guaranteed up to $F_s/2$
|
||||
- What $F_s/2$ should be is an applied engineering question — Nyquist does not answer it
|
||||
- Components above $F_s/2$ are lost: aliasing
|
||||
- LTI and convolution hold in discrete time — commutativity survives
|
||||
- Convolution becomes a sum (finite for FIR, infinite for IIR)
|
||||
|
||||
### 6. Nonlinearity and consequences
|
||||
- Real systems violate LTI: time-varying parameters, feedback, saturation
|
||||
- Each breaks a different condition: time-variance breaks TI, feedback and saturation break linearity
|
||||
- Examples
|
||||
- Modulated delay (chorus, flanger): delay time varies with time → time-invariance breaks; same input at different times yields different output
|
||||
- Overdrive: saturation above threshold → linearity breaks; EQ before overdrive ≠ overdrive before EQ
|
||||
- Commutativity breaks — order of composition is no longer irrelevant
|
||||
- $h_1$ followed by $h_2$ ≠ $h_2$ followed by $h_1$ in general
|
||||
- No universal rule; behavior must be examined per case
|
||||
|
||||
### 7. Summary
|
||||
- Most of the framework rests on the LTI assumption
|
||||
- Signal–system equivalence, convolution, commutativity, frequency-domain analysis — all require linearity and time-invariance
|
||||
- Real systems break the assumption routinely
|
||||
- Nonlinearity is not binary — it is a spectrum
|
||||
- Some systems are close enough to linear that the theory applies as a practical approximation
|
||||
- Others are not — linear theory offers no useful prediction
|
||||
- Judging where a system falls requires knowing what linearity is in the first place
|
||||
- The framework is studied not as a description of reality, but as a baseline against which deviations are measured
|
||||
|
||||
## Text (lyrics if you insist)
|
||||
|
||||
今度は、信号とシステム、そしてデジタル信号処理について説明する。
|
||||
|
||||
### 1. 信号とシステム
|
||||
|
||||
信号とシステムを知るには、信号とシステム両方を知る必要がある。
|
||||
|
||||
信号とは情報を運ぶものである。
|
||||
ここでは時間の経過とともに振幅がどう変化するか、その関数として扱う。
|
||||
|
||||
システムとは、入力信号を受け取り、出力信号を出す箱である。
|
||||
箱の中身は問わない。何を入れたら何が出るか、その関係だけがシステムを定義する。
|
||||
|
||||
システムにはいくつか性質がある。線形性。
|
||||
線形システムとは、重ね合わせの原理が成り立つシステムである。
|
||||
入力を2倍にすれば出力も2倍になる。二つの入力を足し合わせたときの出力は、それぞれ単独で入れたときの出力を足したものと等しい。
|
||||
この二つの条件が揃ったとき、そのシステムは線形である。
|
||||
|
||||
ただ、多くの物理システムはこの条件を満たさない。線形性は特殊であり、当たり前ではない。
|
||||
|
||||
そして時不変性。
|
||||
時不変システムとは、入力を時間軸上でずらしたとき、出力も同じだけずれるシステムである。
|
||||
今日と明日で同じ入力を与えれば、同じ出力が返ってくる。
|
||||
パラメータが時間とともに変わるシステムは、この条件を満たさない。
|
||||
|
||||
線形かつ時不変なシステムを、線形時不変システムと呼ぶ。ここから先は一旦これだけを扱う。
|
||||
これは解析的に扱いやすい、意図的に絞った特殊ケースである。現実の大半はこれに当てはまらない。けど、解析的に扱いやすいから線形時不変だけを考える。
|
||||
|
||||
ここで道具を一つ出す。インパルス信号である。
|
||||
単一の瞬間に単位エネルギーが集中した信号で、実際には存在しない理想だが、使える。ちなみに瞬間とは無限に短い間のことである。
|
||||
このインパルスをある線形時不変システムに通した出力を、そのシステムのインパルス応答と呼ぶ。
|
||||
|
||||
インパルス応答は、そのシステムを完全に表現できる。システムが確定したらインパルス応答も唯一になり、インパルス応答があれば唯一に対応するシステムも特定できる。
|
||||
任意の入力はインパルスの積み重ねとして分解できる。インパルス応答さえ分かれば、どんな入力に対する出力も計算できる。
|
||||
|
||||
それと、インパルス応答はシステムの出力でありながら、信号と同じ種類の関数である。
|
||||
つまりインパルス応答によって、システムは信号と同じ空間に置ける。操作の上では、システムと信号の区別は消える。
|
||||
|
||||
### 2. 線形時不変システムと畳み込み
|
||||
|
||||
インパルス応答がシステムを決定すると言った。なぜだろう。
|
||||
|
||||
任意の入力信号は、大きさの異なるインパルスを時間軸上に並べたものとして表せる。
|
||||
線形時不変システムはそれぞれのインパルスに独立に反応し、その応答を全て重ね合わせる。
|
||||
この重ね合わせ全体の操作を、畳み込みと呼ぶ。
|
||||
|
||||
連続時間では、畳み込みは積分になる。連続じゃない時間の場合は後ほど。
|
||||
時間軸に沿って入力を少しずつずらしながら、そのずれに対応するインパルス応答を強さに応じて足し合わせる。その積分の結果が出力である。
|
||||
|
||||
$y = x * h$ (y イコール x 畳み込み h)
|
||||
|
||||
畳み込みはその二つの引数に対して対称である。
|
||||
$x * h = h * x$ (x 畳み込み h イコール h 畳み込み x)
|
||||
これは計算上の偶然ではない。x と h が同じ種類の対象だからそうなる。前の節で言った、インパルス応答で、システムが信号空間に変換される話の続きである。
|
||||
|
||||
この対称性には実用上の意味がある。
|
||||
二つの線形時不変システムを直列につないだとき、合成されたシステムのインパルス応答は、各インパルス応答の畳み込みである。
|
||||
$h_1 * h_2 = h_2 * h_1$ (h1 畳み込み h2 イコール h2 畳み込み h1)
|
||||
どちらを先につないでも、結果は変わらない。
|
||||
|
||||
数学ではこれを可換性と呼ぶ。一応。
|
||||
|
||||
### 3. 信号変換
|
||||
|
||||
これまで信号を、時間に対する振幅の変化として扱ってきた。同じ信号を、周波数に対する振幅の分布として見ることもできる。
|
||||
厳密には各周波数成分に位相もあるが、ここでは振幅に注目する。
|
||||
|
||||
そのためにフーリエ変換がある。というか、フーリエがその変換の仕方を決めたのでフーリエ変換と呼ばれた。
|
||||
$s(t)$(エス ティー)と $S(\omega)$(エス オメガ)は同じ信号の二つの顔である。
|
||||
一方は時間の経過を、もう一方は周波数の分布を示す。
|
||||
|
||||
つまり、どんな信号も、周波数の異なる正弦波をいくつか重ね合わせたものとして表せる。
|
||||
楽器かエンジンの音がわかりやすい。どちらも複数の周波数成分が混ざっている。その配合が音色を決める。
|
||||
$S(\omega)$ はその内訳である。どの周波数の成分がどれだけの強さで含まれているか、それがスペクトルである。
|
||||
|
||||
アナログ信号は原理的には、ゼロから無限大まであらゆる周波数成分を持ちうる。
|
||||
これが後で問題になる。
|
||||
|
||||
### 4. フィルタ
|
||||
|
||||
信号は原理的に無限の周波数成分を持ちうると述べた。
|
||||
実用上、その全てが必要なことはまずない。例えば人間の聴覚はおよそ20kHzまでしか知覚できない。
|
||||
聞こえないので、人間に聞くための信号では要らない。必要な帯域だけを残し、不要な成分を落とす。それがフィルタの仕事である。
|
||||
フィルタは線形時不変システムであり、フーリエ変換の言葉で言えば、信号のスペクトルを形作るシステムである。
|
||||
|
||||
フィルタを特徴づけるのは周波数応答 $H(\omega)$(エイチ オメガ)である。
|
||||
これはインパルス応答 $h(t)$ をフーリエ変換したものである。
|
||||
インパルス応答が時間領域でシステムを表すように、周波数応答は周波数領域でそのシステムを表している。
|
||||
|
||||
フィルタの働きは、周波数によって信号をふるいにかけることである。
|
||||
基本的な分類としてローパス、ハイパス、バンドパスがある。
|
||||
ローパスフィルタは低い周波数だけを通し、高い周波数を削ぎ落とす。
|
||||
ハイパスフィルタはその逆で、高い周波数を通し、低い周波数を削ぐ。
|
||||
バンドパスフィルタは特定の帯域だけを通し、それより低い成分も高い成分も減衰させる。
|
||||
|
||||
バンドパスフィルタは普通にローパスとハイパスを組み合わせたものと見なせる。
|
||||
どちらを先に適用しても結果は変わらない。線形時不変システムだからの可換性です。
|
||||
|
||||
これらのフィルタは全て、カットオフ周波数という境界を持つ。
|
||||
理想的には、境界の手前までゲインが1、超えた瞬間にゲインが0になる。垂直な壁である。
|
||||
しかし、そのような理想は実現できない。
|
||||
|
||||
現実のフィルタは、カットオフ周波数の周辺に曲線を持つ。急な減衰ではなく、緩やかに下る斜面になる。
|
||||
理想の壁に近づけようとすると、設計が複雑になり、部品が増え、コストが上がる。
|
||||
時には、緩やかな斜面で十分である。
|
||||
|
||||
離散時間においては、有限インパルス応答フィルタと無限インパルス応答フィルタの区別がある。
|
||||
前者のインパルス応答はある時点で終わる。後者は終わらない。
|
||||
|
||||
### 5. 離散化
|
||||
|
||||
フィルタはアナログにもデジタルにも存在する。
|
||||
アナログ空間では、振幅も時間も連続である。最小単位というものがない。なめらかに、切れ目なく続く。
|
||||
デジタル空間では、切れ目がある。
|
||||
デジタルシステムが振幅を表現できる精度は、ビット深度によって決まる。
|
||||
$N$ ビットのシステムは $2^N$ 個の離散的な振幅レベルを持つ。別にこういう二進数の話はしなくていいはずですが。
|
||||
各サンプルの値は、最も近いレベルに丸められる。この丸め誤差を量子化ノイズと呼ぶ。
|
||||
表現できる振幅の上限と下限は、設計とハードウェアによって固定されている。ADCやDACの基準電圧がそれを決める。
|
||||
その境界を超えると、ハードクリップが起きる。
|
||||
アナログ回路は限界に近づくにつれて少しずつ歪む。デジタルにそういう曖昧さはない。境界は硬い。
|
||||
|
||||
次に時間軸について。
|
||||
デジタルシステムは、連続した時間の流れから、一定間隔で値を取り出す。これがサンプリングである。
|
||||
ここで問題がある。前の節で述べたように、アナログ信号は原理的には無限大まで周波数成分を持ちうる。
|
||||
しかし有限のサンプリング頻度では、表現できる周波数に上限がある。全てを保存することはできない。
|
||||
|
||||
その上限を超える成分がサンプリング時に残っていると、消えるわけではない。
|
||||
折り返される。上限を超えた分だけ、反対側に折り返して、低い周波数の成分として現れる。
|
||||
サンプリング周波数が40kHzのとき、25kHzの成分は15kHzとして記録される。存在しないはずの周波数が信号に混入する。
|
||||
これをエイリアシングと呼ぶ。
|
||||
対策は、サンプリングの前にローパスフィルタをかけておくことである。
|
||||
|
||||
ではそのフィルタのカットオフ周波数をどこに設定するか。
|
||||
それは用途次第である。例えば人間の聴覚はおよそ20kHzまでしか知覚できない。
|
||||
音声信号であれば、20kHz以上を保持する理由はない。
|
||||
|
||||
ただし、ここで単純に20kHzでローパスをかけて20kHzでサンプリングすると、情報は失われる。
|
||||
サンプリング周波数と保存したい帯域の上限が同じでは、再構成できない。余裕が要る。
|
||||
|
||||
ナイキストの定理が、その余裕を定量化する。
|
||||
サンプリング周波数を $F_s$ としたとき、$F_s/2$ より低い周波数成分は忠実に再構成できる。
|
||||
つまり保存したい帯域の上限の2倍でサンプリングすれば、再構成が保証される。
|
||||
人間の聴覚が20kHzまでなら、$F_s$ が40kHzを超えていれば聞こえる範囲はカバーできる。
|
||||
余裕を持たせて、一般的には44.1kHzや48kHzが使われる。192kHzは人間以外の何かのためにあるのだろう。
|
||||
定理が言うのはそれだけである。$F_s$ をどこに設定すべきかは用途が決める。
|
||||
$F_s/2$ を超える成分は再構成できない。それがエイリアシングである。
|
||||
|
||||
補足しておく。この離散化によって、線形時不変システムの性質が崩れるわけではない。
|
||||
連続時間において積分であった畳み込みは、離散時間においては和になる。有限インパルス応答なら有限和、無限インパルス応答なら無限和である。
|
||||
可換性を含む全ての性質は引き継がれる。
|
||||
|
||||
### 6. 非線形性とその影響
|
||||
|
||||
では、現実のシステムに戻る。
|
||||
先ほど線形時不変システムは特殊ケースであると述べた。
|
||||
現実のシステムは線形時不変システムの条件を様々な形で破る。
|
||||
時変パラメータは時不変性を破る。
|
||||
フィードバックと飽和は線形性を破る。
|
||||
|
||||
例を挙げる。
|
||||
モジュレーテッドディレイを考える。コーラスやフランジャーがその例である。
|
||||
ディレイタイムが時間とともに変化するとき、同じ入力でも入力するタイミングによって出力が異なる。
|
||||
これは時不変性の破れである。
|
||||
|
||||
次にオーバードライブを考える。
|
||||
入力が閾値を超えると飽和が生じ、出力は入力に比例しなくなる。
|
||||
これは線形性の破れである。
|
||||
そして順序が問題になる。EQの後にオーバードライブを適用した結果と、オーバードライブの後にEQを適用した結果は、一般に異なる。
|
||||
|
||||
このとき、畳み込みの可換性は成立しない。
|
||||
非線形システムでは、一般に
|
||||
$h_1$(システム1)の後に$h_2$(システム2)を適用した結果と、
|
||||
$h_2$(システム2)の後に$h_1$(システム1)を適用した結果は、等しくない。
|
||||
|
||||
普遍的な規則は存在しない。
|
||||
各ケースを個別に検討する必要がある。
|
||||
|
||||
### まとめ
|
||||
|
||||
以上がデジタル信号処理の基本的な枠組みである。
|
||||
|
||||
振り返ると、ここで述べたことの大半は線形時不変システムという前提の上に成り立っている。
|
||||
信号とシステムが同じ空間に住むこと。畳み込みによる合成。可換性。周波数領域での見通しのよさ。
|
||||
これらは全て、線形性と時不変性が揃っているからこそ成り立つ性質である。
|
||||
|
||||
そして最後で述べた通り、現実のシステムはその前提をしばしば破る。
|
||||
|
||||
では、なぜこの枠組みを学ぶのか、現実のシステムはほとんど非線形なのに。
|
||||
非線形性は白黒ではないから。線形と見なして良いシステムもあれば、線形理論が全く通用しないシステムもある。
|
||||
その見極めには、何が破れているのかを知る必要がある。線形の理論はそのためにある。
|
||||
線形時不変システムの理論は、現実そのものの記述ではない。現実を測るための基準線である。
|
||||
+158
-26
@@ -1,6 +1,46 @@
|
||||
# pulse under latent semantic envelopes
|
||||
|
||||
*(WIP. Description and pieces only now)*
|
||||
*Lyrics as well as the process of how it became like this*
|
||||
|
||||
## Lyrics, in full
|
||||
|
||||
第一部。
|
||||
|
||||
Surface.
|
||||
The equilibrium of silence between the static magnetic noise, the watery planes, and the optical waves.
|
||||
Blurred and mangled electronic signs over the outbound atmospheric leeway.
|
||||
Distant stations estimated in abnormal heat perform the inconsistent collimation rite.
|
||||
|
||||
Dive.
|
||||
The black screen of stalemate conceals the massive stars, the charred comets, and the meteorites.
|
||||
Their questionable glitters, once emissions of brave borderers and dark adventurers,
|
||||
now the only traces, obscured and indifferent to all but those with a special breed of faith.
|
||||
|
||||
The noise floor of a receiving environment is the aggregate of all background signal energy present across a given frequency band at any instant. In urban contexts, primary contributors include electromagnetic emissions from power distribution infrastructure, wireless communications networks, and consumer electronic devices. The noise floor is not centrally coordinated; no single source generates it. It emerges from the cumulative output of many independently operating transmitters, each functioning within its own parameters, whose combined effect is persistent and largely uniform. Under standard conditions, this background level is treated as a fixed baseline rather than a variable of interest. Signals whose amplitude falls at or below this baseline remain physically present in the medium but cannot be distinguished from the background by a receiving system, and are not considered recoverable.
|
||||
|
||||
An impulse signal, also referred to as a unit impulse, is defined as a signal with zero value at all points in time except a single instant, at which its value is notionally unbounded, subject to the constraint that its total integrated area equals unity. In practice, true impulse signals cannot be physically produced; they are approximated by short-duration, high-amplitude pulses of sufficient bandwidth. The impulse response of any linear time-invariant system — that is, the output produced when an impulse is presented as input — fully characterises that system's behaviour under any subsequent input. In the absence of any input, a system at rest produces no output and exhibits no properties that distinguish it from any other inactive system. From a single impulse and its measured response, the complete transfer function of the system can be derived.
|
||||
|
||||
第二部。
|
||||
|
||||
Subdued sonic scalars vectorise
|
||||
Acoustic aberrations dissipate
|
||||
|
||||
Isotropic funnel strikes a packet rain
|
||||
Fine-grained bags of acks ground astray
|
||||
|
||||
Ocean as a desert buries all of its life
|
||||
Deck the shell with a perfect thermocline
|
||||
|
||||
Curtain shut before the Fresnel bights
|
||||
Plasma buildup swells your rovers' thighs
|
||||
|
||||
第三部。
|
||||
|
||||
Particulate pressure across a landlocked island
|
||||
Ultraviolet utilitarian paints every third crag bland
|
||||
Lossy lies lament the ludicrous lawful liberty
|
||||
Siren sounds soaken in the sauna of substantiality
|
||||
Eccentric sediments launch unbidden perpetuity
|
||||
|
||||
## Planning area
|
||||
|
||||
@@ -26,7 +66,7 @@ Literally, there can be long, essay-like, paragraphs (not too many) with structu
|
||||
1. "Surface." - the breach, the fever-dream opener
|
||||
2. Instrumental - let the image settle
|
||||
3. Essay/documentary - the explanatory core, almost clinical, "semi-boring scientific" (this is where the concept gets stated plainly before the rest of the song complicates it)
|
||||
4. Instrumental + fragments - the logic starts dissolving
|
||||
4. Instrumental - going into the poemic verses
|
||||
5. Poem-ic verses - emotional density, maybe climax, maybe not
|
||||
6. P-U-L-S-E 5-liners - trademark closer, acrostic constraint
|
||||
7. Instrumental outro
|
||||
@@ -35,19 +75,19 @@ Literally, there can be long, essay-like, paragraphs (not too many) with structu
|
||||
|
||||
## Construction zone
|
||||
|
||||
The first locked in draft version of the song. Things get developed even after this version.
|
||||
|
||||
```cpp
|
||||
// Opener
|
||||
Surface.
|
||||
The equilibrium of silence between the static magnetic noise, the watery planes, and the optical waves
|
||||
blurring the electronic signs over the outbound informational motorway.
|
||||
Images imagined in the imaginary region of the protruding obelisks many kilometres away
|
||||
from this here expanding entropy killing various variables as abnormal heat usually may.
|
||||
The equilibrium of silence between the static magnetic noise, the watery planes, and the optical waves.
|
||||
Blurred and mangled electronic signs over the outbound informational motorway.
|
||||
Distant stations estimated in abnormal heat perform the collimation rite.
|
||||
// think a submarine surfacing at night to do some radio comms as the starting point, then infuse it with a fever dream
|
||||
// 'Images imagined in the imaginary region' is half wordplay, half multimedia transmission with a loss (signal-wise). the obelisks are basically antennae.
|
||||
|
||||
Dive.
|
||||
The black screen of stalemate conceals the black stars, the black comets, the black meteorites
|
||||
their questionable glitters and gleams, once emissions of brave borderers and dark adventurers
|
||||
their questionable glitters, once emissions of brave borderers and dark adventurers
|
||||
now the only traces, obscured and indifferent to all but those with a special breed of faith.
|
||||
// 'Dive' - an opposite to 'Surface'. This is on the receiving end.
|
||||
// Just see this as 'receiving other guys' pulses'.
|
||||
@@ -57,9 +97,9 @@ now the only traces, obscured and indifferent to all but those with a special br
|
||||
|
||||
// The pseudo documentary narration, half-blended into the background music
|
||||
// in a metalingual sense the music is the background noise and the narration is the signal with a low SNR. Or one could consider the roles swapped.
|
||||
The term "ambient frequency" refers to the aggregate of background signals present in a given environment. In urban contexts, this includes electromagnetic interference from power infrastructure, wireless communications networks, and consumer electronic devices. The ambient field is not centrally coordinated. No single source generates it. It emerges from the cumulative output of many transmitters operating within similar parameters. The resulting noise floor is persistent and of low significance under normal circumstances. Signals that fall below this threshold remain technically present but are not distinguishable from the noise itself, and are not considered practially receivable.
|
||||
The noise floor of a receiving environment is the aggregate of all background signal energy present across a given frequency band at any instant. In urban contexts, primary contributors include electromagnetic emissions from power distribution infrastructure, wireless communications networks, and consumer electronic devices. The noise floor is not centrally coordinated; no single source generates it. It emerges from the cumulative output of many independently operating transmitters, each functioning within its own parameters, whose combined effect is persistent and largely uniform. Under standard conditions, this background level is treated as a fixed baseline rather than a variable of interest. Signals whose amplitude falls at or below this baseline remain physically present in the medium but cannot be distinguished from the background by a receiving system, and are not considered recoverable.
|
||||
|
||||
An impulse signal, sometimes referred to as a unit impulse, is defined as a signal with zero amplitude at all points except for a single instant, at which point its amplitude approaches infinity while its total energy remains finite. In practical applications, true impulse signals cannot be generated; they are approximated by short-duration waveforms of high amplitude. The impulse response of any linear system - that is, the system's output when presented with an impulse input - fully characterises that system's behaviour. From a single impulse and its response, the complete identity of the system can be derived. This property is considered fundamental to signal analysis.
|
||||
An impulse signal, also referred to as a unit impulse, is defined as a signal with zero value at all points in time except a single instant, at which its value is notionally unbounded, subject to the constraint that its total integrated area equals unity. In practice, true impulse signals cannot be physically produced; they are approximated by short-duration, high-amplitude pulses of sufficient bandwidth. The impulse response of any linear time-invariant system — that is, the output produced when an impulse is presented as input — fully characterises that system's behaviour under any subsequent input. In the absence of any input, a system at rest produces no output and exhibits no properties that distinguish it from any other inactive system. From a single impulse and its measured response, the complete transfer function of the system can be derived. This property is considered foundational to systems analysis.
|
||||
|
||||
// build some tension with bass and drum and fiddle here
|
||||
|
||||
@@ -67,26 +107,118 @@ An impulse signal, sometimes referred to as a unit impulse, is defined as a sign
|
||||
// more 'rhythmic yelling' than singing, you get the idea
|
||||
Subdued sonic scalars vectorise
|
||||
Acoustic aberrations dissipate
|
||||
// sahb - duwd - sao - nihk - skay - lahrs - vehk - tahr - rayz,
|
||||
// (ah - ) kaws - tihk - ax - bahr - ray - shaxns - dihs - sah - payt
|
||||
|
||||
Spectral broadcast lossy with a narrow sight
|
||||
No perceived replies per design
|
||||
// needs further tweaking
|
||||
// --- stanza: the receiving end ---
|
||||
Isotropic funnel strikes a packet rain
|
||||
Fine-grained bags of acks ground astray
|
||||
|
||||
The ocean as a desert buries all of its life
|
||||
Ocean as a desert buries all of its life
|
||||
Deck the shell with a perfect thermocline
|
||||
// (thax - ) ao - sheh(n) - naes - ax - deh - zahrt - bah - riys - ao(l) - lahv - ihts - layf
|
||||
// (dehk - thax - ) shehl - wit - ax - pahr - fekt - thahr - mow - klayn
|
||||
// 'deck the shell': intentional vagueness between 'decorate the outer shell' and 'knock on the border surface'.
|
||||
|
||||
A dim light dies in the rover's eyes
|
||||
Or the next one could never come otherwise
|
||||
// Thoughts: 'the light dies' because a pulse is mostly dormant you wanker. You have to let the light die so it lights up again, simple as
|
||||
Curtain shut before the Fresnel bights
|
||||
Plasma buildup swells your rover's thighs
|
||||
|
||||
// ending verse(s) before the painfully long instrumental outro
|
||||
p.....
|
||||
u.....
|
||||
l.....
|
||||
s.....
|
||||
essential sentiments lay unconditionally permanent
|
||||
// P:
|
||||
particulate pressure plagues across a landlocked island
|
||||
// U:
|
||||
ultraviolet utilitarian paints every third crag bland
|
||||
// L:
|
||||
lossy lies lament the ludicrous lawful liberty
|
||||
// S:
|
||||
siren sounds soaken in the sauna of substantiality
|
||||
// E-S-L-U-P:
|
||||
essential sentiments leave unbidden perpetually
|
||||
```
|
||||
|
||||
## Notes about this song
|
||||
|
||||
I will be just writing about anything I think of related to this song. At least for the first go. It's not really creative when you set yourself an agenda and duly follow it.
|
||||
|
||||
### Why pulse
|
||||
|
||||
Some time after witnessing the rover song go locally viral at the second transparent fes I said to myself there will be a third. Almost immediately after that, I was sure the keyword would be 'pulse'. Then it took me several months to figure out the third recursive acronym composition. Maybe more than several months.
|
||||
The choice in 'pulse' is for a very apparent reason. It's literally the last kairo (original) song. That is, the last song in the last album. I know they had more 2hu released after that but that's a different business, isn't it.
|
||||
It also is a very good word in itself because there is vitality, there is machinery, there is electronics and engineering, all in one package. It's kairo plus Northwich committee of magic researches.
|
||||
|
||||
### Educational rock
|
||||
|
||||
It's now a traditional practice to sample random people speaking random stuff in a post-rock / prog rock / chill hip-hop / shoegaze / you name it song. There is a hip-hop remix of Niinistö's new year's day address but I can't really dig it out of the Internet rabbit hole. And to start with literally the first post-rock I ever listened to was the 2hu band 'the distant journey to you' by our Veto guy and there was an excerpt of Kon Satoshi's 'Paprika' in one of the songs.
|
||||
When I really started to shape this song I was listening to way more Snakes Don't Belong In Alaska than I should. Two of their later songs sampled someone presumably giving a lecture or speech or that sort of presentation. I very much liked it as I'm a big podcast person. It's like listening to your podcasts with psychedelica music accompanying.
|
||||
I also have been a firm believer that signal processing is something any music-aware people should have at least a slightest idea about. So here we are. It's only natural to make my own mini lecture about some signal basics and incorporate that audio signal into the song.
|
||||
The other parts of the song are intentionally cryptic, evasive, indirect, fuzzy, ambiguous but the educational bit should be factual enough it can be considered a decent crash course on what it talks about. Aside from being the catalyst for the song's underlying emotional progressions. We'll talk about that later or never.
|
||||
|
||||
|
||||
### Football manager ruined my life and LLM agents severed it into pieces
|
||||
|
||||
Back in your 2022s and 2023s, I wasn't aware of Anthropic and ChatGPT was basically chat powered by GPT. But this time I approach my songwriting as a relatively heavy Claude user. I never fully hand the lyric composition to LLM agents or raw LLMs as that's the fun bit, not the chore bit.
|
||||
What I personally do is usually tell the Claude 'hey I'm thinking about something like this' in either the plainest of languages or just a draft verse. 9 out of 10 times its response pisses me off. Which is consistent with the experience when you do these exchanges with any verified decent human counterpart. No disrespect to my verified 'I can talk poetic bullshit with them' friends but that just happens. And it happens for good.
|
||||
You see the key here is I have someone to talk to, 24/7, super fluent in languages I speak or not, only restricted by usage limits. For writing lyrics I never hit the limits, by the way. An adversary is an essential upgrade to a rubber duck as it gives you feedback. LLM at the core means this adversary is complex enough to be considered unpredictable and thus inspiring.
|
||||
Your Claudes and ChatGPTs and maybe Geminis out of the box loves proposing verses after verses which is less helpful. Fortunately, I've tweaked my instances to hold fire and communicate in a more atmospheric way. In a convo I get myself immersed in the atmosphere I initiated and the agent amplified. I guess that's where all the fever dreams root from.
|
||||
|
||||
|
||||
### On professionalism and why I consider myself pro enough even if I have no proper proof
|
||||
|
||||
In a way, I think it's probably true that flair and discipline has a dance of balance in many things, and I take music and association football as two seemingly irrelevant examples. Simply because I'm somewhat invested in those two.
|
||||
A classical piece in the broadest sense takes many, many players with their respective instruments. There would be a leader violinist or a pianist or similar role but let's face it, for like 99% of the time, orchestra players don't play flashy solo phrases. I feel it's like how you have no-nonsense centre backs and anchor defensive midfielders like Jordan Rossiter, that in a sense stand out even less than a goalkeeper.
|
||||
Your rock and pop bands of 3, 4, 5 people is something else. Basically everyone needs to be a star. Everyone will have flashy, tricky solo riffs.
|
||||
And don't get me wrong, I by no means am stating that playing simple notes correctly is inferior to smashing an outrageous riff, or that only getting the ball and passing to the nearest full back is inferior to dribbling over 5 defenders and driving home with a full bollocks swing. That is just not true. Doing the less fancy 'basic' but proper and well is just a slightly, or very different dimension of professionalism. I personally very like no-nonsense centre backs because I played as one because I couldn't do fancy stuff on the pitch.
|
||||
I don't think I can do fancy stuff in my FL studio either. The justification around all this equity of professionalism had a big portion of making myself feel less useless when I tried and mostly failed to learn to play the guitar then bass. Today I'm totally ok and free of complexes like that.
|
||||
|
||||
The financial stability is also very real. My plan about music has always been 'find a job so I can play with music' and I've finally done it. I mean I never got anywhere around trying to be self-sustainable with music alone or as the main means, but having that peace of mind of a roof above is certainly helpful even for a hobby. Having music to unleash my creativity buildup when there is any also helps me cope with the 'boring' day job so it's a win-win. Probably.
|
||||
|
||||
### Music, bloody hell
|
||||
|
||||
I don't think one should be expecting me to elaborate on how delicate my music craftsmanship is after the revelation and identification as a no-nonsense FL Studio user at best Conference National level. But you see, there are people who watch Conf Nat as devoted as they themselves watch the Prem, or even more. Absolute skill levels are not that much of a matter.
|
||||
|
||||
I opted to go for a classic dual-bass, dual-telecaster setup with various other helper instruments because that's simply my favourite. The second bass actually turned out to be utilised less than I originally planned based on how things went in 'rover over virtually eternal roundabouts'.
|
||||
|
||||
Also, the song started as my experiment of a more techno-heavy style. It didn't end well, but the techno leftovers might be (I'm doubtful) spotted around the place.
|
||||
Then I switched to the target of making this one the third piece in the recursive acronym trilogy, so essentially I'm doing humble plagiarism on the previous two songs. 'sauna as universal neural analeptic' and 'rover over virtually eternal roundabouts' if you're wondering (no one asked).
|
||||
Like stated elsewhere the psychedelica influence from SDBIA is faintly visible (audible). I also applied some of the minimalism principles to the guitar lines, they should be perceivably sparser than some contemporaries.
|
||||
|
||||
Drums and percussions have always been my passion but I think I just got too lazy to fine tune every loop pattern. A drum track of eternally-morphing patterns is not realistic in some sense after all.
|
||||
|
||||
### Vocal synthesis masterclass
|
||||
|
||||
Could be stating the obvious, but I'm much better at English than Japanese. When it comes to songwriting I might even argue English enjoys priority over my first language, which is Chinese. So it's a bit of a hassle when there is no good enough Yukari TTS on the shelf. On a side note, Kotonoha sisters got their official English / Chinese TTS but that's a shambles. To put it mildly.
|
||||
When I discovered GPT-SoVITS which in English is something that allows you to generate many languages in person A's voice and person B's accent (partly), it's only natural that I immediately went for it. It fit like a glove on the foot.
|
||||
You see, the problem with many voice changers, even in the l'époque de l'intelligence artificielle, is that they end up with English (simplified) pronunciations. Some give you RP but that's not what I want either. It's the models that are my problem and I needed something more weird.
|
||||
The goal was to mix rallienglanti, Bristolian, and some Aussie flavour to produce the random and weird English I usually speak. The end product is nowhere near that, but equally far from 'standard' English. Which is very acceptable.
|
||||
Also, this synthesised voice should not be considered truly Yukari in the copyright or trademarks sense. I credited Yukari because I included pieces of Japanese voice made with the authentic A.I.VOICE 2 editor with a valid license. Just saying.
|
||||
|
||||
For the sung parts, Synthesiser V Studio Pro 2 is the weapon of massive destruction of choice as always. The actual labour is rather boring so I'll leave it as 'I mix-matched mostly Spanish and English phonemes'.
|
||||
|
||||
### So what is this song about?
|
||||
|
||||
It's all in the title mate.
|
||||
|
||||
Keep reading if you think that alone is not meaningful enough.
|
||||
|
||||
I'm firmly against providing an answer as standard yet very happy to write my personal interpretations.
|
||||
On condition that no one promotes the following text as the only valid official explanation to the song, I can give this:
|
||||
|
||||
There is atmospheric pressure around every one of us. It is like the environmental noise that suffocates signals without malice. The noise itself also contains various meaningful, info-bearing signals but all piled up together as a big blob of something. This is the 'latent semantic envelopes' bit.
|
||||
The pulse is an object as well as an action. It's both the signal that gets sent (and drowned) and the action of sending such signals. Not to mention it also means one's heart beats which is a necessary condition for that guy to be alive.
|
||||
Pulsing, under 'latent semantic envelopes' or more straightforwardly, under overwhelming noise, is absurd. But if you subscribe to the absurd, it's not really a problem.
|
||||
That's it.
|
||||
|
||||
I probably should have required signing an EULA to have access to this.
|
||||
|
||||
### A personal crusade against claims of superhuman frequential cognitive abilities
|
||||
|
||||
The decision to use the 192 kbps MP3 as a source of truth, converting from that to nominally WAV and everything else was based on several grounds. But the dominant one is that I never liked high fidelity above a certain threshold of good enough. And that threshold for me is roughly 192 kbps of MP3. Another reason is that I would like even one more person to think outside of the box, into the channels, towards the corner flag, that this is a form of proper, intentional digital lofi rather than lossy media.
|
||||
It's totally fair to consider this shithousery because shithousery is what I see it as. I'm not doing this digital lo-fi for the sake of art, I do this as reactive banter invoked by phonophile nonsense. If 192 kbps MP3 is good enough to me, then it's good enough.
|
||||
I do think at 128 kbps I start to sense that something is missing and more information density would be appreciated. So the personal threshold is there or thereabouts.
|
||||
|
||||
|
||||
### Trivia
|
||||
- In case you didn't notice, the title forms a recursive acronym.
|
||||
- In case you didn't notice, the final verse is an acrostic of 'pulse'.
|
||||
- In case you didn't notice, the last line is a reverse acrostic of 'pulse'.
|
||||
- Brave Borderer and Dark Adventurer are RN fast patrol crafts, or in English, torpedo boats. And I believe both are playable in War Thunder which is why I know them.
|
||||
- 'Ocean as a desert buries all of its life' is a blatant steal from America's 'A horse with no name'. The original line is 'ocean is a desert with its life underground'.
|
||||
- I have to sneak in the exact words 'rover' and 'sauna', sorry. And in case you didn't notice, they are two related songs.
|
||||
- Many of the words could be understood with a leaning towards their optical engineering aspects. That said, I was more on the electronics end of my opto-electronic science and engineering syllabus.
|
||||
- Meta-trivia: this note is plagued by football cliches, and the song has contracted some.
|
||||
Reference in New Issue
Block a user