fix&feat: hiển thị tools chỉnh sửa audioclip

This commit is contained in:
2026-07-19 10:19:25 +07:00
parent 615e0e8530
commit 95dc9346ef
6 changed files with 5146 additions and 36 deletions
+283
View File
@@ -0,0 +1,283 @@
# Technical Specification: Advanced Editing Toolset & Graph-Based Continuous Waveform Painting on Sub-Tab
This document defines the interactive layout design, the configuration of the toolbar button arrays, and the signal processing routines for compiling a Graph-based Continuous Waveform graph optimized for the microscopic viewports inside the isolated temporary document workspace (Sub-tab), referencing the structural paradigms of `image_5ec2e5.png` and `image_5ec363.png`.
---
## 1. Target Selection Scope
The toolset within the Sub-tab environment supports two target operational boundaries:
* **Global Clip:** When no specific timeline selection highlighted mask is present, all active DSP effects apply uniformly across the entire length of the extracted Audio Clip.
* **Selected Range:** When an explicit timeline segment $[T_{\text{start}}, T_{\text{end}}]$ is highlighted by the user, DSP routines calculate changes exclusively inside those boundaries. Splice junctions automatically compute crossfades to mitigate transient click/pop anomalies.
---
## 2. Ruler-Based Tools
These utilities display as intuitive, linear slider scales (Sliders/Rulers) embedded in the top toolbar row:
```text
[ Normalize: |======o======| 0 dB ] [ Gain: |====o====| +3 dB ] [ Pitch: |==o==| -2 Semi ]
```
### 2.1. Peak Normalization
* **UI Layout:** A slide scale control allowing users to configure target amplitude thresholds variable from $-12\text{ dBFS}$ down to $0\text{ dBFS}$.
* **DSP Math Algorithm:** Locate the maximum absolute peak amplitude value $A_{\text{max}}$ within the targeted area, then multiply all active samples by a static scalar gain multiplier $G$:
$$G = \frac{10^{\frac{\text{Target\_dB}}{20}}}{A_{\text{max}}}$$
### 2.2. Volume Up / Down (Quick Gain)
* **UI Layout:** A linear sliding ruler modulating the overall absolute gain structure of the focused segment.
* **Operational Range:** Adjustable from $-\infty\text{ dB}$ (complete mute attenuation) up to $+12\text{ dB}$ of linear amplification.
### 2.3. Pitch Shifting
* **UI Layout:** A calibrated slider modifying the project's fundamental frequencies discrete in semitones or cents.
* **Operational Range:** Boundaries map from $-12\text{ semitones}$ (one octave down) to $+12\text{ semitones}$ (one octave up).
* **DSP Engine Routine:** Employs a spectral Phase Vocoder to shift frequencies without affecting the physical, real-time duration layout of the segment.
---
## 3. Graph-Based Fades
Fading curves overlay graphically directly onto the highlighted waveform canvas region, enabling precise boundary amplitude adjustments:
```text
Linear Fade-In Exponential Fade-Out
+───────────────────────────+ +───────────────────────────+
| /███████████████| |███████████\ |
| / ███████████████| |███████████ \ |
| / ███████████████| |███████████ \___ |
| / ███████████████| |███████████ \______|
+───────────────────────────+ +───────────────────────────+
|<──────── Fade-In ────────>| |<─────── Fade-Out ────────>|
```
* **Fade-In:** Multiplies an ascending amplitude ramp from $0.0$ to $1.0$ at the starting index profile of the selection region. Users can toggle seamlessly between **Linear** or **Exponential** curves to achieve a smoother, more psychoacoustically natural volume build-up.
* **Fade-Out:** Multiplies a descending amplitude decay ramp from $1.0$ down to $0.0$ at the trailing boundary edge of the selection range.
---
## 4. Ruler Percentage Stretch Tool
A dedicated percentage metric scale control (`Ruler %`) sitting on the control toolbar dictates time-stretching and playback velocity parameters:
```text
[ Speed Stretch %: |========o========| 100% (Native) ] -> Range: 50% - 200%
```
* **Interaction Mapping:** Users drag the percentage slider node or hold down the `Alt` key and drag the rightmost boundary edge of the clip along the horizontal axis to change this scale metric.
* **Sync Formula:** Let $D$ map to the unscaled native duration value, and $D'$ map to the target modified duration footprint. The resulting structural playback speed ratio percentage ($S$) is given by:
$$S = \frac{D}{D'} \times 100\%$$
---
## 5. Continuous Graph-Based Waveform Painting & Microscopic Viewports
The waveform graph inside the Sub-tab is compiled as a unified, continuous line vector (Continuous Line Graph) that flows seamlessly along the timeline axis, mapping the literal physical phase displacements of the underlying audio signal.
### 5.1. Logarithmic Amplitude Axis Grid Layout
Following the professional paradigm established in `image_5ec2e5.png`, the waveform painting canvas is divided by a symmetrical layout grid reflecting both positive and negative polarity limits of the central horizontal axis:
```text
+6.0 dB ───────────────────────────────────────────────────────────────────
~ ~ ~ ~ ~ ~ ~ ~ (Sub-division Grid Line) ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
-6.0 dB ───────────────────────────────────────────────────────────────────
\ / \ / \
-Inf dB ─○───────────/───────────────○───────────/───────────────○───────── (Zero-Line Axis)
\ / \ / \
-6.0 dB ───────────────────────────────────────────────────────────────────
+6.0 dB ───────────────────────────────────────────────────────────────────
```
* **Visual Bounding Thresholds:**
* **Central Zero Axis (-Inf. dB):** Maps the absolute baseline $0\text{V}$ electrical reference (complete absence of audio signal / absolute silence).
* **Symmetrical Decibel Grids:** Project accurate scale metrics tracking normalized peak levels (the inner $-6.0\text{ dB}$ sub-grid marks a $50.1\%$ amplitude ceiling, while the outermost physical frame boundary aligns to $+6.0\text{ dB}$ or $0\text{ dBFS}$).
### 5.2. Standard Workspace View vs. Ultra Zoom Viewport Scaling
The drawing engine dynamically hot-swaps its rendering calculations (Rendering Routine) depending on the active pixel compression metric $Z$ (pixels/second):
* **Standard View Mode ($Z < 500\text{ pixels/second}$):** The system deploys a structural peak compression layout algorithm (**Peak Waveform**—as referenced in `image_5ec2e5.png`). It connects the maximum absolute upper peak bounding indices (Max) with the lower minimum value ranges (Min) passing through a common pixel column into a unified vector line, generating an organic, aliases-free continuous waveform silhouette.
* **Micro Viewport Zoom-In ($Z \ge 500\text{ pixels/second}$—as referenced in `image_5ec363.png`):** Once viewport stretching scales past this threshold, the framework transitions into a **Single Continuous Sine Polyline** loop. Chronologically sequential acoustic sample addresses ($x[i]$, $x[i+1]$) map as discrete vector coordinate indices bound together by thin lines (using sharp smooth polyline vectors or linear/cubic spline interpolation loops), charting pristine, individual sinusoidal phases explicitly.
### 5.3. Zero-Crossing Alignment within Ultra Zoom Viewports
When performing rapid cursor tracking edits (Scrub/Drag Selection), the alignment routine locks the selection boundary marker coordinates onto the nearest baseline sample offset exhibiting a complete algebraic phase conversion (sign inversion):
$$x[i] \cdot x[i+1] \le 0$$
---
## 6. Top Duration Timeline
Directly above the isolated sub-tab waveform canvas lane, a dedicated horizontal measuring ruler tracks clip timing data:
```text
| 0:00.000 | 0:01.000 | 0:02.000 | 0:03.000 | 0:04.000 (Duration: 4.152s)
+───────────────────────────────────────────────────────────────────────────────────────+
| [==================== VÙNG QUÉT CHỌN (RANGE SELECTION) ====================] |
+───────────────────────────────────────────────────────────────────────────────────────+
```
* **Total Duration Monitoring:** Renders the absolute, precise time extent of the isolated audio block in the right-hand corner of the timeline ruler layout (e.g., `Duration: 12.450s`).
* **Duration Selection Drag:** Left-clicking and dragging horizontally inside this top duration bar defines a highlighted selection overlay window. This range indicator automatically projects down into the waveform lane underneath.
---
## 7. Bottom Transport Panel & Master Tools
A comprehensive control framework containing expanded navigation buttons and deep session processing controls anchors the bottom row of the sub-tab environment, matching the layout structure in `image_5ec2e5.png`:
```text
+─────────────────────────────────────────────────────────────────────────────────────────────+
| [● Rec] [◀◀ Back] [▶ Play] [|| Pause] [■ Stop] | Rate: |====o====| 0.00 | Loop: [X] |
|---------------------------------------------------------------------------------------------|
| [Volume Pencil Tool] [AI Analysis Tool] | Active Asset: linh_ngua_powerup.wav |
+─────────────────────────────────────────────────────────────────────────────────────────────+
```
### 7.1. Functional Mapping Matrix:
* **Record (● Red Indicator):** Drives live microphone capture sequences targeted straight into the isolated sub-tab data matrix.
* **Back (◀◀ Rewind):** Resets the timeline playhead position index back to the absolute starting point ($t = 0.0\text{ s}$).
* **Play / Pause / Stop:** Coordinates low-latency runtime audio execution tracking locked onto the sub-tab's RAM cache blocks.
* **Rate Slider:** Adjusts the global monitoring playback pitch speed metrics in real time without overwriting source asset length (calibrated step ranges variable from `-1.00` scaling up to `+1.00`).
* **Loop Toggle:** Toggles continuous cycle loops over the highlighted section or the whole clip.
* **Volume Pencil Tool:** Engages the drawing framework to map point nodes for automated amplitude envelopes.
* **AI Analysis Tool:** Instructs the dockerized engine to evaluate rhythmic transient markers and pitch tracking grids.
---
## 8. Volume Automation Envelope (Pen Tool)
This advanced timeline automation layer allows audio designers to draw custom gain curves over the background waveform graphics.
```text
VOLUME AUTOMATION ENVELOPE (PEN TOOL)
+3 dB ──────────────────────────────────────────────────────────────
\ Node 1 Node 3
\ ○ ○
0 dB ───\────/─\─────────────────────────────────────/─\─────────── (0 dB Unity Gain Axis)
\ / \ / \
\/ \ / \
○ \_______________________________/ \________
Node 2 Node 4
-30 dB ──────────────────────────────────────────────────────────────
|<─────────────────── Horizontal Axis (Time) ─────────────────────>|
```
### 8.1. Pen Tool Interaction Mechanics
* **Activation:** Clicking the designated Pencil Tool icon in the control panel modifies the pointer device presentation into a pencil graphic.
* **Envelope Initialization:** Activating the Pen Tool generates a solid horizontal neon green line representing $0\text{ dB}$ (Unity Gain) across the track workspace, acting as the baseline master axis.
* **Drawing Automation Curves:**
* Left-clicking anywhere along this line creates an adjustable anchor point (**Control Node**).
* Dragging an initialized control node upward increases signal amplitude (up to a maximal ceiling boundary of $+3\text{ dB}$).
* Dragging a control node downward reduces signal amplitude (down to a lower attenuation floor of $-30\text{ dB}$).
* The graphics framework automatically updates straight vector paths between sequential nodes utilizing simple linear interpolation.
### 8.2. DSP Volume Envelope Math
Given two chronologically adjacent drawn points $P_1(t_1, V_1)$ and $P_2(t_2, V_2)$, the targeted instantaneous decibel gain variable $V_{\text{dB}}(t)$ at an arbitrary time index $t$ ($t_1 \le t \le t_2$) matches the following linear equation:
$$V_{\text{dB}}(t) = V_1 + (t - t_1) \cdot \frac{V_2 - V_1}{t_2 - t_1}$$
This decibel value must be translated into a standard linear gain scalar coefficient $G_{\text{linear}}(t)$ to multiply it into the core audio sample stream values:
$$G_{\text{linear}}(t) = 10^{\frac{\text{V}_{\text{dB}}(t)}{20}}$$
$$x_{\text{automation}}[n] = x[n] \cdot G_{\text{linear}}\left( \frac{n}{\text{Sample Rate}} \right)$$
---
## 9. Porting Guidelines for Python Desktop Layouts (PyQt6 QPainter Context)
When translating the polyline vector engine and the symmetrical decibel gridding lines into a containerized desktop application using the native `QPainter` canvas inside PyQt6, leveraging a structured `QPainterPath` prevents rendering lag when mapping high-density signal segments:
```python
# [PYTHON PORTING BLUEPRINT] - Continuous Polyline Waveform Rendering via QPainterPath
from PyQt6.QtGui import QPainter, QPainterPath, QPen, QColor
from PyQt6.QtCore import QPointF, Qt
import numpy as np
def paint_continuous_waveform_path(painter: QPainter, rect_width: int, rect_height: int, y: np.ndarray, zoom_level: float):
"""
Renders a unified continuous single polyline path tracing absolute physical signal transitions.
y: A 1D NumPy float32 array tracking raw sample amplitudes bounded within [-1.0, 1.0].
zoom_level: The scale allocation mapping physical drawing pixels per second of audio data.
"""
if len(y) == 0:
return
painter.setRenderHint(QPainter.RenderHint.Antialiasing, True)
mid_y = rect_height / 2.0
# 1. Compile background Decibel reference grids (-6.0 dB, -Inf. dB, -6.0 dB)
grid_pen = QPen(QColor(45, 45, 45), 1, Qt.PenStyle.DashLine)
painter.setPen(grid_pen)
# A threshold of -6.0 dB maps approximately to an absolute scalar amplitude index of 0.501
y_6db_top = mid_y - (0.501 * (rect_height * 0.42))
y_6db_bottom = mid_y + (0.501 * (rect_height * 0.42))
painter.drawLine(0, int(y_6db_top), rect_width, int(y_6db_top))
painter.drawLine(0, int(y_6db_bottom), rect_width, int(y_6db_bottom))
# Paint the absolute Zero-Line horizontal center axis (-Inf. dB)
center_pen = QPen(QColor(60, 60, 60), 1, Qt.PenStyle.SolidLine)
painter.setPen(center_pen)
painter.drawLine(0, int(mid_y), rect_width, int(mid_y))
# 2. Initialize the Continuous Vector Polyline Route Layout Block
wave_path = QPainterPath()
wave_pen = QPen(QColor(100, 149, 237), 1.2, Qt.PenStyle.SolidLine) # Professional Cornflower Blue
painter.setPen(wave_pen)
# Map raw buffer indexes into structural coordinate pixels
start_point_set = False
for x_pixel in range(rect_width):
# Translate current canvas pixel offset back to timeline seconds metrics
time_at_pixel = x_pixel / zoom_level
# Calculate target array element offset
sample_index = int(time_at_pixel * 44100) # Assuming project sample rate baseline at 44.1kHz
if sample_index >= len(y):
break
amplitude = y[sample_index]
y_pixel = mid_y + (amplitude * (rect_height * 0.42))
if not start_point_set:
wave_path.moveTo(float(x_pixel), y_pixel)
start_point_set = True
else:
wave_path.lineTo(float(x_pixel), y_pixel)
# Draw the continuous vector polyline overlay onto the viewport canvas
painter.drawPath(wave_path)
```
+51
View File
@@ -31,6 +31,57 @@ class SubTabDSPEngine:
gain = target_amplitude / max_amplitude gain = target_amplitude / max_amplitude
return y * gain return y * gain
@staticmethod
def apply_volume_automation_envelope(y: np.ndarray, sr: int, nodes: list) -> np.ndarray:
"""
Applies a user-drawn volume automation envelope onto an acoustic signal NumPy array.
nodes: A list of point dictionaries, e.g., [{"time": 0.0, "db": 0.0}, {"time": 2.5, "db": -12.0}, ...]
"""
if not nodes:
return y
# Sort envelope nodes chronologically by time axis
nodes = sorted(nodes, key=lambda x: x["time"])
# 1. Map node variables into distinct coordinates arrays
node_times = np.array([node["time"] for node in nodes])
node_dbs = np.array([node["db"] for node in nodes])
# Hard-clamp boundary constraints matching the operational floor [-30.0dB, +3.0dB]
node_dbs = np.clip(node_dbs, -30.0, 3.0)
# 2. Evaluate absolute timeline timestamps for every index position inside the signal array
total_samples = len(y)
sample_times = np.arange(total_samples) / sr
# 3. Linearly interpolate localized decibel thresholds across every single sample step
# Handle edge cases for interpolation: if sample_times is outside node_times range,
# np.interp uses the first/last value of node_dbs.
interpolated_dbs = np.interp(sample_times, node_times, node_dbs, left=node_dbs[0], right=node_dbs[-1])
# 4. Map logarithmic values into standard linear gain scale arrays
linear_gains = 10.0 ** (interpolated_dbs / 20.0)
# 5. Multiply the raw amplitude vector array by the linear gain modifier mask
return y * linear_gains
@staticmethod
def pitch_shift(y: np.ndarray, sr: int, n_steps: float) -> np.ndarray:
"""
Shift the pitch of an audio signal by a specified number of semitones.
Args:
y: Input audio signal
sr: Sample rate
n_steps: Number of semitones to shift (positive = higher pitch, negative = lower pitch)
Returns:
Pitch-shifted audio signal
"""
if n_steps == 0:
return y
return librosa.effects.pitch_shift(y, sr=sr, n_steps=n_steps)
@staticmethod @staticmethod
def merge_back_to_parent( def merge_back_to_parent(
parent_track_audio: np.ndarray, parent_track_audio: np.ndarray,
+135 -18
View File
@@ -779,6 +779,41 @@
); );
}; };
const SubTabToolbar = ({ st, activeTool, setActiveTool, handleSubTabNormalizeWithValue, handleSubTabGainWithValue, handleSubTabPitch, handleSubTabStretch, handleSubTabFade }) => (
<div className="daw-header flex h-16 items-center px-4 border-b daw-border gap-3 bg-zinc-800">
<div className="flex items-center space-x-1 border-r border-zinc-700 pr-3">
<button className={`p-1.5 rounded ${activeTool === 'select' ? 'bg-cyan-700' : 'bg-zinc-700'}`} onClick={() => setActiveTool('select')} title="Select"><i data-lucide="mouse-pointer" className="w-4 h-4"></i></button>
<button className={`p-1.5 rounded ${activeTool === 'grab' ? 'bg-cyan-700' : 'bg-zinc-700'}`} onClick={() => setActiveTool('grab')} title="Grab"><i data-lucide="hand" className="w-4 h-4"></i></button>
<button className={`p-1.5 rounded ${activeTool === 'razor' ? 'bg-cyan-700' : 'bg-zinc-700'}`} onClick={() => setActiveTool('razor')} title="Razor"><i data-lucide="scissors" className="w-4 h-4"></i></button>
<button className={`p-1.5 rounded ${activeTool === 'pen' ? 'bg-cyan-700' : 'bg-zinc-700'}`} onClick={() => setActiveTool('pen')} title="Pen Tool"><i data-lucide="pen-tool" className="w-4 h-4"></i></button>
</div>
<div className="flex items-center space-x-3 flex-1 overflow-x-auto no-scrollbar">
<div className="flex items-center gap-1">
<span className="text-[10px] text-zinc-400">Normalize:</span>
<input type="range" min="-12" max="0" value={st.effects?.normalizeDb || 0} step="0.1" onChange={(e) => handleSubTabNormalizeWithValue(st.id, parseFloat(e.target.value))} className="w-20 h-1.5"/>
</div>
<div className="flex items-center gap-1">
<span className="text-[10px] text-zinc-400">Gain:</span>
<input type="range" min="-40" max="24" value={st.effects?.gainDb || 0} step="0.1" onChange={(e) => handleSubTabGainWithValue(st.id, parseFloat(e.target.value))} className="w-20 h-1.5"/>
</div>
<div className="flex items-center gap-1">
<span className="text-[10px] text-zinc-400">Pitch:</span>
<input type="range" min="-12" max="12" value={st.effects?.pitch || 0} step="0.1" onChange={(e) => handleSubTabPitch(st.id, parseFloat(e.target.value))} className="w-20 h-1.5"/>
</div>
<div className="flex items-center gap-1">
<span className="text-[10px] text-zinc-400">Stretch:</span>
<input type="range" min="50" max="200" value={st.effects?.speedStretch || 100} step="1" onChange={(e) => handleSubTabStretch(st.id, parseInt(e.target.value))} className="w-20 h-1.5"/>
</div>
</div>
<div className="flex items-center space-x-1 pl-3 border-l border-zinc-700">
<button onClick={() => handleSubTabFade(st.id, 'in')} className="px-2 py-1 text-[10px] rounded hover:bg-zinc-700">Fade In</button>
<button onClick={() => handleSubTabFade(st.id, 'out')} className="px-2 py-1 text-[10px] rounded hover:bg-zinc-700">Fade Out</button>
</div>
</div>
);
const App = () => { const App = () => {
// ── State Definitions ── // ── State Definitions ──
const [tracks, setTracks] = useState([ const [tracks, setTracks] = useState([
@@ -810,14 +845,6 @@
snapValueRef.current = snapValue; snapValueRef.current = snapValue;
const bpmRef = useRef(bpm); const bpmRef = useRef(bpm);
bpmRef.current = bpm; bpmRef.current = bpm;
useEffect(() => {
setTimeout(() => {
if (window.lucide) {
window.lucide.createIcons();
}
}, 50);
}, [activeTool]);
// BMP for Tempo Track - LOOP_EDITOR_2.md §6 // BMP for Tempo Track - LOOP_EDITOR_2.md §6
const [selectedTrackId, setSelectedTrackId] = useState('1'); const [selectedTrackId, setSelectedTrackId] = useState('1');
const [currentTime, setCurrentTime] = useState(0); const [currentTime, setCurrentTime] = useState(0);
@@ -945,6 +972,15 @@
fadeOutMs: 0, fadeOutMs: 0,
}); });
// Lucide icons initialization
useEffect(() => {
setTimeout(() => {
if (window.lucide) {
window.lucide.createIcons();
}
}, 50);
}, [activeTool, activeTab]);
const timelineWrapperRef = useRef(null); const timelineWrapperRef = useRef(null);
const tcpContainerRef = useRef(null); const tcpContainerRef = useRef(null);
@@ -1380,7 +1416,7 @@
startTime: selLeft, startTime: selLeft,
endTime: selRight, endTime: selRight,
buffer: subBuffer, buffer: subBuffer,
effects: { reverse: false, gainDb: 0, fadeInMs: 0, fadeOutMs: 0 }, effects: { reverse: false, gainDb: 0, fadeInMs: 0, fadeOutMs: 0, normalizeDb: 0, pitch: 0, speedStretch: 100 },
currentTime: 0, currentTime: 0,
selectionStart: null, selectionStart: null,
selectionEnd: null, selectionEnd: null,
@@ -1421,7 +1457,7 @@
startTime: clip.startTime, startTime: clip.startTime,
endTime: clip.startTime + clip.buffer.duration, endTime: clip.startTime + clip.buffer.duration,
buffer: subBuffer, buffer: subBuffer,
effects: { reverse: false, gainDb: 0, fadeInMs: 0, fadeOutMs: 0 }, effects: { reverse: false, gainDb: 0, fadeInMs: 0, fadeOutMs: 0, normalizeDb: 0, pitch: 0, speedStretch: 100 },
currentTime: 0, currentTime: 0,
selectionStart: null, selectionStart: null,
selectionEnd: null, selectionEnd: null,
@@ -1466,14 +1502,58 @@
const fadeSamples = Math.min(edData.length, Math.floor(fx.fadeInMs / 1000 * sr)); const fadeSamples = Math.min(edData.length, Math.floor(fx.fadeInMs / 1000 * sr));
for (let i = 0; i < fadeSamples; i++) edData[i] = edData[i] * (i / fadeSamples); for (let i = 0; i < fadeSamples; i++) edData[i] = edData[i] * (i / fadeSamples);
} }
// Fade out // Fade out
if (fx.fadeOutMs > 0) { if (fx.fadeOutMs > 0) {
const sr = edBuffer.sampleRate; const sr = edBuffer.sampleRate;
const fadeSamples = Math.min(edData.length, Math.floor(fx.fadeOutMs / 1000 * sr)); const fadeSamples = Math.min(edData.length, Math.floor(fx.fadeOutMs / 1000 * sr));
for (let i = edData.length - fadeSamples; i < edData.length; i++) { for (let i = edData.length - fadeSamples; i < edData.length; i++) {
edData[i] = edData[i] * ((edData.length - 1 - i) / fadeSamples); edData[i] = edData[i] * ((edData.length - 1 - i) / fadeSamples);
} }
} }
// Normalize (apply to selection or entire buffer)
if (fx.normalizeDb !== 0) {
const targetDb = fx.normalizeDb;
let maxVal = 0;
for (let i = 0; i < edData.length; i++) {
const abs = Math.abs(edData[i]);
if (abs > maxVal) maxVal = abs;
}
if (maxVal > 0) {
const targetAmp = Math.pow(10, targetDb / 20);
const scale = targetAmp / maxVal;
for (let i = 0; i < edData.length; i++) {
edData[i] = Math.max(-1, Math.min(1, edData[i] * scale));
}
}
}
// Pitch shift (simple resample)
if (fx.pitch !== 0) {
const ratio = Math.pow(2, fx.pitch / 12);
const newLength = Math.round(edData.length / ratio);
const newData = new Float32Array(newLength);
for (let i = 0; i < newLength; i++) {
const srcIdx = i * ratio;
const idx0 = Math.floor(srcIdx);
const idx1 = Math.min(idx0 + 1, edData.length - 1);
const frac = srcIdx - idx0;
newData[i] = edData[idx0] * (1 - frac) + edData[idx1] * frac;
}
edData.set(newData);
}
// Speed stretch
if (fx.speedStretch !== 100) {
const ratio = fx.speedStretch / 100;
const newLength = Math.round(edData.length * ratio);
const newData = new Float32Array(newLength);
for (let i = 0; i < newLength; i++) {
const srcIdx = i / ratio;
const idx0 = Math.floor(srcIdx);
const idx1 = Math.min(idx0 + 1, edData.length - 1);
const frac = srcIdx - idx0;
newData[i] = edData[idx0] * (1 - frac) + edData[idx1] * frac;
}
edData.set(newData);
}
if (subTab.clipId) { if (subTab.clipId) {
// Clip-based merge // Clip-based merge
@@ -1549,6 +1629,33 @@
setSubTabs(prev => prev.map(s => s.id === tabId ? { ...s, effects: { ...s.effects, ...effects } } : s)); setSubTabs(prev => prev.map(s => s.id === tabId ? { ...s, effects: { ...s.effects, ...effects } } : s));
}; };
const handleSubTabNormalizeWithValue = (tabId, normalizeDb) => {
updateSubTabEffects(tabId, { normalizeDb });
};
const handleSubTabGainWithValue = (tabId, gainDb) => {
updateSubTabEffects(tabId, { gainDb });
};
const handleSubTabPitch = (tabId, pitch) => {
updateSubTabEffects(tabId, { pitch });
};
const handleSubTabStretch = (tabId, speedStretch) => {
updateSubTabEffects(tabId, { speedStretch });
};
const handleSubTabFadeState = (tabId, type) => {
const st = subTabs.find(s => s.id === tabId);
if (!st) return;
const fx = st.effects || {};
if (type === 'in') {
updateSubTabEffects(tabId, { fadeInMs: (fx.fadeInMs || 0) + 10 });
} else if (type === 'out') {
updateSubTabEffects(tabId, { fadeOutMs: (fx.fadeOutMs || 0) + 10 });
}
};
// ── Context Menu Handlers ── // ── Context Menu Handlers ──
const handleContextMenu = (e, trackId, clickTime) => { const handleContextMenu = (e, trackId, clickTime) => {
e.preventDefault(); e.preventDefault();
@@ -4318,6 +4425,16 @@
</button> </button>
</div> </div>
<div className="flex-1 flex flex-col p-3 gap-3 overflow-y-auto"> <div className="flex-1 flex flex-col p-3 gap-3 overflow-y-auto">
<SubTabToolbar
st={st}
activeTool={activeTool}
setActiveTool={setActiveTool}
handleSubTabNormalizeWithValue={handleSubTabNormalizeWithValue}
handleSubTabGainWithValue={handleSubTabGainWithValue}
handleSubTabPitch={handleSubTabPitch}
handleSubTabStretch={handleSubTabStretch}
handleSubTabFade={handleSubTabFadeState}
/>
<SubTabWaveform <SubTabWaveform
buffer={st.buffer} buffer={st.buffer}
subTabId={st.id} subTabId={st.id}
+135 -18
View File
@@ -779,6 +779,41 @@
); );
}; };
const SubTabToolbar = ({ st, activeTool, setActiveTool, handleSubTabNormalizeWithValue, handleSubTabGainWithValue, handleSubTabPitch, handleSubTabStretch, handleSubTabFade }) => (
<div className="daw-header flex h-16 items-center px-4 border-b daw-border gap-3 bg-zinc-800">
<div className="flex items-center space-x-1 border-r border-zinc-700 pr-3">
<button className={`p-1.5 rounded ${activeTool === 'select' ? 'bg-cyan-700' : 'bg-zinc-700'}`} onClick={() => setActiveTool('select')} title="Select"><i data-lucide="mouse-pointer" className="w-4 h-4"></i></button>
<button className={`p-1.5 rounded ${activeTool === 'grab' ? 'bg-cyan-700' : 'bg-zinc-700'}`} onClick={() => setActiveTool('grab')} title="Grab"><i data-lucide="hand" className="w-4 h-4"></i></button>
<button className={`p-1.5 rounded ${activeTool === 'razor' ? 'bg-cyan-700' : 'bg-zinc-700'}`} onClick={() => setActiveTool('razor')} title="Razor"><i data-lucide="scissors" className="w-4 h-4"></i></button>
<button className={`p-1.5 rounded ${activeTool === 'pen' ? 'bg-cyan-700' : 'bg-zinc-700'}`} onClick={() => setActiveTool('pen')} title="Pen Tool"><i data-lucide="pen-tool" className="w-4 h-4"></i></button>
</div>
<div className="flex items-center space-x-3 flex-1 overflow-x-auto no-scrollbar">
<div className="flex items-center gap-1">
<span className="text-[10px] text-zinc-400">Normalize:</span>
<input type="range" min="-12" max="0" value={st.effects?.normalizeDb || 0} step="0.1" onChange={(e) => handleSubTabNormalizeWithValue(st.id, parseFloat(e.target.value))} className="w-20 h-1.5"/>
</div>
<div className="flex items-center gap-1">
<span className="text-[10px] text-zinc-400">Gain:</span>
<input type="range" min="-40" max="24" value={st.effects?.gainDb || 0} step="0.1" onChange={(e) => handleSubTabGainWithValue(st.id, parseFloat(e.target.value))} className="w-20 h-1.5"/>
</div>
<div className="flex items-center gap-1">
<span className="text-[10px] text-zinc-400">Pitch:</span>
<input type="range" min="-12" max="12" value={st.effects?.pitch || 0} step="0.1" onChange={(e) => handleSubTabPitch(st.id, parseFloat(e.target.value))} className="w-20 h-1.5"/>
</div>
<div className="flex items-center gap-1">
<span className="text-[10px] text-zinc-400">Stretch:</span>
<input type="range" min="50" max="200" value={st.effects?.speedStretch || 100} step="1" onChange={(e) => handleSubTabStretch(st.id, parseInt(e.target.value))} className="w-20 h-1.5"/>
</div>
</div>
<div className="flex items-center space-x-1 pl-3 border-l border-zinc-700">
<button onClick={() => handleSubTabFade(st.id, 'in')} className="px-2 py-1 text-[10px] rounded hover:bg-zinc-700">Fade In</button>
<button onClick={() => handleSubTabFade(st.id, 'out')} className="px-2 py-1 text-[10px] rounded hover:bg-zinc-700">Fade Out</button>
</div>
</div>
);
const App = () => { const App = () => {
// ── State Definitions ── // ── State Definitions ──
const [tracks, setTracks] = useState([ const [tracks, setTracks] = useState([
@@ -810,14 +845,6 @@
snapValueRef.current = snapValue; snapValueRef.current = snapValue;
const bpmRef = useRef(bpm); const bpmRef = useRef(bpm);
bpmRef.current = bpm; bpmRef.current = bpm;
useEffect(() => {
setTimeout(() => {
if (window.lucide) {
window.lucide.createIcons();
}
}, 50);
}, [activeTool]);
// BMP for Tempo Track - LOOP_EDITOR_2.md §6 // BMP for Tempo Track - LOOP_EDITOR_2.md §6
const [selectedTrackId, setSelectedTrackId] = useState('1'); const [selectedTrackId, setSelectedTrackId] = useState('1');
const [currentTime, setCurrentTime] = useState(0); const [currentTime, setCurrentTime] = useState(0);
@@ -945,6 +972,15 @@
fadeOutMs: 0, fadeOutMs: 0,
}); });
// Lucide icons initialization
useEffect(() => {
setTimeout(() => {
if (window.lucide) {
window.lucide.createIcons();
}
}, 50);
}, [activeTool, activeTab]);
const timelineWrapperRef = useRef(null); const timelineWrapperRef = useRef(null);
const tcpContainerRef = useRef(null); const tcpContainerRef = useRef(null);
@@ -1380,7 +1416,7 @@
startTime: selLeft, startTime: selLeft,
endTime: selRight, endTime: selRight,
buffer: subBuffer, buffer: subBuffer,
effects: { reverse: false, gainDb: 0, fadeInMs: 0, fadeOutMs: 0 }, effects: { reverse: false, gainDb: 0, fadeInMs: 0, fadeOutMs: 0, normalizeDb: 0, pitch: 0, speedStretch: 100 },
currentTime: 0, currentTime: 0,
selectionStart: null, selectionStart: null,
selectionEnd: null, selectionEnd: null,
@@ -1421,7 +1457,7 @@
startTime: clip.startTime, startTime: clip.startTime,
endTime: clip.startTime + clip.buffer.duration, endTime: clip.startTime + clip.buffer.duration,
buffer: subBuffer, buffer: subBuffer,
effects: { reverse: false, gainDb: 0, fadeInMs: 0, fadeOutMs: 0 }, effects: { reverse: false, gainDb: 0, fadeInMs: 0, fadeOutMs: 0, normalizeDb: 0, pitch: 0, speedStretch: 100 },
currentTime: 0, currentTime: 0,
selectionStart: null, selectionStart: null,
selectionEnd: null, selectionEnd: null,
@@ -1466,14 +1502,58 @@
const fadeSamples = Math.min(edData.length, Math.floor(fx.fadeInMs / 1000 * sr)); const fadeSamples = Math.min(edData.length, Math.floor(fx.fadeInMs / 1000 * sr));
for (let i = 0; i < fadeSamples; i++) edData[i] = edData[i] * (i / fadeSamples); for (let i = 0; i < fadeSamples; i++) edData[i] = edData[i] * (i / fadeSamples);
} }
// Fade out // Fade out
if (fx.fadeOutMs > 0) { if (fx.fadeOutMs > 0) {
const sr = edBuffer.sampleRate; const sr = edBuffer.sampleRate;
const fadeSamples = Math.min(edData.length, Math.floor(fx.fadeOutMs / 1000 * sr)); const fadeSamples = Math.min(edData.length, Math.floor(fx.fadeOutMs / 1000 * sr));
for (let i = edData.length - fadeSamples; i < edData.length; i++) { for (let i = edData.length - fadeSamples; i < edData.length; i++) {
edData[i] = edData[i] * ((edData.length - 1 - i) / fadeSamples); edData[i] = edData[i] * ((edData.length - 1 - i) / fadeSamples);
} }
} }
// Normalize (apply to selection or entire buffer)
if (fx.normalizeDb !== 0) {
const targetDb = fx.normalizeDb;
let maxVal = 0;
for (let i = 0; i < edData.length; i++) {
const abs = Math.abs(edData[i]);
if (abs > maxVal) maxVal = abs;
}
if (maxVal > 0) {
const targetAmp = Math.pow(10, targetDb / 20);
const scale = targetAmp / maxVal;
for (let i = 0; i < edData.length; i++) {
edData[i] = Math.max(-1, Math.min(1, edData[i] * scale));
}
}
}
// Pitch shift (simple resample)
if (fx.pitch !== 0) {
const ratio = Math.pow(2, fx.pitch / 12);
const newLength = Math.round(edData.length / ratio);
const newData = new Float32Array(newLength);
for (let i = 0; i < newLength; i++) {
const srcIdx = i * ratio;
const idx0 = Math.floor(srcIdx);
const idx1 = Math.min(idx0 + 1, edData.length - 1);
const frac = srcIdx - idx0;
newData[i] = edData[idx0] * (1 - frac) + edData[idx1] * frac;
}
edData.set(newData);
}
// Speed stretch
if (fx.speedStretch !== 100) {
const ratio = fx.speedStretch / 100;
const newLength = Math.round(edData.length * ratio);
const newData = new Float32Array(newLength);
for (let i = 0; i < newLength; i++) {
const srcIdx = i / ratio;
const idx0 = Math.floor(srcIdx);
const idx1 = Math.min(idx0 + 1, edData.length - 1);
const frac = srcIdx - idx0;
newData[i] = edData[idx0] * (1 - frac) + edData[idx1] * frac;
}
edData.set(newData);
}
if (subTab.clipId) { if (subTab.clipId) {
// Clip-based merge // Clip-based merge
@@ -1549,6 +1629,33 @@
setSubTabs(prev => prev.map(s => s.id === tabId ? { ...s, effects: { ...s.effects, ...effects } } : s)); setSubTabs(prev => prev.map(s => s.id === tabId ? { ...s, effects: { ...s.effects, ...effects } } : s));
}; };
const handleSubTabNormalizeWithValue = (tabId, normalizeDb) => {
updateSubTabEffects(tabId, { normalizeDb });
};
const handleSubTabGainWithValue = (tabId, gainDb) => {
updateSubTabEffects(tabId, { gainDb });
};
const handleSubTabPitch = (tabId, pitch) => {
updateSubTabEffects(tabId, { pitch });
};
const handleSubTabStretch = (tabId, speedStretch) => {
updateSubTabEffects(tabId, { speedStretch });
};
const handleSubTabFadeState = (tabId, type) => {
const st = subTabs.find(s => s.id === tabId);
if (!st) return;
const fx = st.effects || {};
if (type === 'in') {
updateSubTabEffects(tabId, { fadeInMs: (fx.fadeInMs || 0) + 10 });
} else if (type === 'out') {
updateSubTabEffects(tabId, { fadeOutMs: (fx.fadeOutMs || 0) + 10 });
}
};
// ── Context Menu Handlers ── // ── Context Menu Handlers ──
const handleContextMenu = (e, trackId, clickTime) => { const handleContextMenu = (e, trackId, clickTime) => {
e.preventDefault(); e.preventDefault();
@@ -4318,6 +4425,16 @@
</button> </button>
</div> </div>
<div className="flex-1 flex flex-col p-3 gap-3 overflow-y-auto"> <div className="flex-1 flex flex-col p-3 gap-3 overflow-y-auto">
<SubTabToolbar
st={st}
activeTool={activeTool}
setActiveTool={setActiveTool}
handleSubTabNormalizeWithValue={handleSubTabNormalizeWithValue}
handleSubTabGainWithValue={handleSubTabGainWithValue}
handleSubTabPitch={handleSubTabPitch}
handleSubTabStretch={handleSubTabStretch}
handleSubTabFade={handleSubTabFadeState}
/>
<SubTabWaveform <SubTabWaveform
buffer={st.buffer} buffer={st.buffer}
subTabId={st.id} subTabId={st.id}
+4506
View File
File diff suppressed because it is too large Load Diff
+36
View File
@@ -42,3 +42,39 @@ def test_merge_back_to_parent():
assert np.allclose(res[:900], 1.0) assert np.allclose(res[:900], 1.0)
# Inside the edited range should be zero (except crossfades) # Inside the edited range should be zero (except crossfades)
assert np.allclose(res[1100:2900], 0.0) assert np.allclose(res[1100:2900], 0.0)
def test_apply_volume_automation_envelope():
sr = 44100
duration = 2.0
y = np.ones(int(sr * duration), dtype=np.float32) * 0.5 # Constant signal at -6dB
# Simple fade in from -inf to 0dB over 1 second
nodes = [
{"time": 0.0, "db": -60.0}, # Effectively -inf
{"time": 1.0, "db": 0.0},
{"time": 2.0, "db": 0.0},
]
y_automated = SubTabDSPEngine.apply_volume_automation_envelope(y, sr, nodes)
y_automated = SubTabDSPEngine.apply_volume_automation_envelope(y, sr, nodes)
# Check start: should be 0.5 * 10**(-30/20) due to clipping
expected_start_val = 0.5 * (10**(-30/20.0))
assert np.isclose(y_automated[0], expected_start_val, atol=1e-5)
# Check at 0.5 seconds: interpolated to -15dB (halfway between -30dB and 0dB)
# y * (10 ** (-15 / 20))
expected_mid_val = 0.5 * (10**(-15/20.0))
assert np.isclose(y_automated[int(0.5 * sr)], expected_mid_val, atol=1e-5)
# Check at 1.0 seconds: should be 0.5 * (10**(0/20)) = 0.5
assert np.isclose(y_automated[int(1.0 * sr)], 0.5, atol=1e-5)
# Check at end: should be 0.5
assert np.isclose(y_automated[-1], 0.5, atol=1e-5)
# Test with empty nodes
y_no_nodes = SubTabDSPEngine.apply_volume_automation_envelope(y, sr, [])
assert np.array_equal(y_no_nodes, y)