Image To Midi Converter Online May 2026

For all its intrigue, the online image-to-MIDI converter suffers from one fatal flaw when judged as a musical tool: the output is almost never pleasant to listen to.

Music is not random data. It relies on structured rhythm, harmonic progression, phrasing, and repetition. An image, by contrast, is spatially static. The mapping process described above results in a few predictable outcomes:

Most online converters offer basic filters—downsampling (scanning only every 5th or 10th pixel), quantization (snapping notes to a musical scale like C major), and note duration limits—to mitigate the chaos. However, even with these filters, the result sounds more like a test signal than a song.

Why would anyone use such a tool? The applications fall into three main categories:

I tested a sunset landscape (1024×768, warm colors) across three tools.

Pix2Music:

MIDIculous:

Img2Midi:

Verdict: No tool produces a "harmonically correct" song automatically. The output is best described as raw material – expect noise, glitches, and happy accidents.

Keep in mind that the development and availability of online tools can change rapidly, so it's always a good idea to search for the most current options when you're ready to convert an image to MIDI.

Finding an online image-to-MIDI converter depends on whether you are trying to turn sheet music into a playable file or use visual data

(like a photo or digital art) to generate experimental music. 1. Converting Sheet Music to MIDI If you have a photo or PDF of a musical score, you need Optical Music Recognition (OMR)

software. This identifies the notes, tempo, and dynamics on the page. PlayScore 2

: A highly rated mobile app that allows you to scan physical sheet music or upload a PDF to digitize it into MIDI format. Soundslice

: While primarily an editor, it has a robust PDF/image importer that uses AI to convert scores into MIDI or MusicXML.

: A desktop solution specifically for scanning sheet music and exporting it to MIDI for use in DAWs like Ableton or Logic Pro. PlayScore 2 2. Converting Photos/Art to MIDI (Experimental)

This method maps the visual properties of an image—such as pixel brightness, color (RGB), or saturation—to musical properties like pitch and duration. Image2MIDI

: A straightforward web tool where you upload any image, click "generate," and it creates a MIDI file based on the visual data. Ansible's Image to MIDI

: A browser-based experimental tool that lets you customize how colors and patterns are translated into notes.

: An AI-driven platform that "interprets" your image to create a full musical composition, which you can then download as a MIDI file. Image to MIDI Quick Step-by-Step Guide Select your tool

: Choose an OMR tool for sheet music or an experimental tool for general photos. Upload the file : Use a high-resolution

for the best accuracy. For sheet music, ensure the lighting is even and the lines are straight. Adjust settings

: Many tools allow you to select a "Scale" (e.g., C Major) or "Instrument" before generating. Download and Import : Download the resulting file and drag it into a Digital Audio Workstation (DAW) Ableton Live to assign it a virtual instrument sound. Are you converting a specific musical score , or are you looking to create generative music from a regular photograph? Image to MIDI

Online image-to-MIDI converters are specialized tools that fall into two primary categories: sonification tools that turn general imagery (photos, art) into experimental music, and Optical Music Recognition (OMR) tools that translate sheet music into playable MIDI files. 1. General Image-to-MIDI Tools (Experimental)

These tools use pixel data—such as brightness, color (RGB), or position—to determine musical properties like pitch, velocity, and duration.

Image2MIDI: A web-based tool that splits an image into a grid of cells. It maps brighter cells to higher notes and assigns different rows to separate MIDI tracks. image to midi converter online

img2midi (GitHub): An open-source script that uses an algorithm to average RGB components and interpolate them into a MIDI note range (

Creative Applications: Artists like Glasys use these workflows to create "MIDI art," where the visual pattern in a Digital Audio Workstation (DAW) piano roll forms a recognizable image while simultaneously playing a composition. 2. Sheet Music-to-MIDI Tools (Practical)

These tools focus on accuracy, using OCR technology to recognize musical notation for playback or editing in a DAW.

Scan2Notes: A dedicated browser-based tool for converting photos or PDFs of printed sheet music into MIDI without software installation.

PlayScore 2: An app that allows users to play music directly from a photo and export the results to notation software like MuseScore or Dorico.

ScanScore: A professional-grade suite that digitizes sheet music. It offers tiered versions (Melody, Ensemble, Professional) depending on the number of staves required. 3. Comparison of Core Methods Sonification (e.g., Image2MIDI) OMR (e.g., ScanScore) Primary Input Photos, abstract art, screenshots Scanned sheet music, PDFs Musical Logic Pixel brightness/color →right arrow Musical symbols →right arrow Best For Experimental sound design, MIDI art Transcribing scores, learning pieces Accuracy Subjective/Random High (depends on image quality) 4. Technical Implementation

For developers, libraries like music21 are often used to bridge the gap between image processing and MIDI creation. Common algorithms split images into rows (representing MIDI notes from

) and use pixel intensity to trigger a note for a set duration, often beats per pixel column. Image to MIDI

Imagine a world where your favorite photo doesn't just look good—it sounds good too. An image-to-MIDI converter acts as a bridge between visual aesthetics and musical composition, translating pixels into a playable score. Core Feature: "Visual-to-Sonic Mapping"

The standout feature of an image-to-MIDI converter is its ability to interpret visual data—like color, brightness, and structure—and map them to musical parameters.

Pixel-to-Pitch Conversion: The software analyzes the vertical position of pixels or specific colors to determine note pitches. For instance, lighter colors might represent higher notes, while darker shades trigger lower registers.

Color-Based Instrumentation: Different hues can be assigned to specific MIDI tracks or instruments. Deep blues might map to a cello, while vibrant yellows trigger a synth lead.

Intensity & Velocity: The saturation or brightness of a pixel can determine the MIDI velocity (loudness) of a note, creating a dynamic performance based on the image's lighting. Practical Applications

Experimental Composition: Musicians can use complex images—like star charts or abstract paintings—to generate unique melodies that human intuition might not conceive.

Data Sonification: Turn a graph or a satellite image into a soundscape to "hear" the patterns in the data.

Sound Design Inspiration: Use a photo of a forest or a cityscape to create an atmospheric background layer for a film score or an ambient track. How to Use It

Upload: Select any image (JPG, PNG, etc.) and upload it to the converter.

Configure: Set your preferred musical scale, BPM, and instrument tags to guide the AI or algorithm.

Generate: The tool processes the image and creates a downloadable .mid file.

Import: Bring the file into a DAW like FL Studio or MuseScore to further refine the sound.

Free MIDI Editor | Create & Edit MIDI Files Online - OpenMusic AI

Image to MIDI Converter Online: A Comprehensive Guide

Are you looking for an online tool to convert images into MIDI files? Look no further! In this text, we'll explore the concept of image to MIDI conversion, discuss the available online tools, and provide a step-by-step guide on how to use them.

What is MIDI?

MIDI (Musical Instrument Digital Interface) is a protocol that allows electronic musical instruments, computers, and other devices to communicate and control each other. A MIDI file contains musical notes, rhythms, and other performance data that can be played back on a synthesizer or other MIDI-compatible device. For all its intrigue, the online image-to-MIDI converter

What is Image to MIDI Conversion?

Image to MIDI conversion is the process of converting visual data from an image into MIDI data. This can be useful for various applications, such as:

Online Image to MIDI Converters

Several online tools offer image to MIDI conversion capabilities. Here are a few:

Step-by-Step Guide to Using an Online Image to MIDI Converter

Let's use the Online-Convert Image to MIDI Converter as an example:

Limitations and Future Developments

While online image to MIDI converters are available, the results may vary depending on the tool and the complexity of the image. Future developments in AI and machine learning may improve the accuracy and quality of image to MIDI conversion.

In conclusion, online image to MIDI converters offer a fascinating way to transform visual data into musical compositions. With this guide, you're ready to explore the possibilities of image to MIDI conversion and create your own unique musical creations.

Creating a MIDI file from an image is a unique way to turn visual data like photos, drawings, or screenshots into musical compositions. Online Image to MIDI Tools

Image2MIDI: A dedicated tool that scans images (JPEG, JPG, PNG) and converts pixels into musical notes. It translates brightness and color into pitch, creating a multi-track MIDI file based on the image's rows and columns.

OpenMusic AI: Offers an intuitive online generator where you can import or create new MIDI projects from scratch using AI-driven tools.

PDFgear: While specialized for documents, this is the go-to for converting PDF sheet music images into accurate MIDI files. How the Conversion Works

Most converters do not "hear" the image; they interpret visual data through specific parameters:

Grid Mapping: The software splits the image into rows (for tracks) and columns (for timing).

Pitch & Brightness: Lighter or brighter pixels typically result in higher notes, while darker areas produce lower pitches.

Note Density: Users can often adjust the "note chance threshold" to control how many notes are generated, depending on how dark or light the original image is. Alternative Audio-to-MIDI Options

If you are looking to convert different media types, specialized tools exist for other formats:

Audio Files: Tools like MusicCreator AI or Melodyne convert MP3 or WAV files into MIDI for use in DAWs.

YouTube: La Touche Musicale can extract MIDI data directly from video URLs.

Online image-to-MIDI converters serve two primary purposes: converting visual sheet music into playable notes (Optical Music Recognition) and transforming regular photos into experimental soundscapes (Algorithmic Sonification). 🎹 Best Tools for Sheet Music (OMR)

These tools use Optical Music Recognition (OMR) to detect notes, staves, and rhythms from photos or PDFs of printed scores.

Scan2Notes: A simple browser-based tool for quick conversions of printed sheet music to MIDI without software installation.

ACE Studio: Provides an AI-powered "Sheet Music to MusicXML/MIDI" converter. Free users can convert up to 10 files per day.

PlayScore 2: An app-based solution (with online export features) that reads photos and exports accurate MIDI to DAWs like Ableton or Logic. MIDIculous:

ScanScore: A professional suite that offers a mobile app to snap photos and sync them to a desktop editor for MIDI cleanup. 🎨 Best Tools for Creative/Experimental Conversion

These tools treat pixels as data (brightness, color, position) to generate abstract MIDI patterns.

Image2MIDI: A dedicated web tool that transforms any JPEG or PNG into a multi-track MIDI file based on pixel data.

DrawSound: An application that maps image characteristics to MIDI Control Change (CC) messages, useful for automating synthesizer parameters.

Melodyne / Basic Pitch: While primarily audio-to-MIDI, these are often the "second step" for creators who first turn images into audio spectrograms. ⚙️ How the Conversion Works The technology varies significantly depending on the goal: 🎼 Score Recognition (OMR) Scanning: Analyzes the image for horizontal lines (staves).

Detection: Identifies note heads, stems, and flags to determine pitch and duration.

Output: Produces a MIDI file that mimics the original written music. Convert Image and PDF to MusicXML Online - ACE Studio

Report: Online Image-to-MIDI Conversion Tools Converting images to MIDI is generally split into two categories: Creative/Experimental conversion (turning visual data into abstract music) and Optical Music Recognition (OMR)

(transcribing scanned sheet music into playable digital notes). 1. Top Online Creative Converters

These tools interpret the pixels, colors, or patterns of an image to generate unique musical compositions. Image2MIDI

: A dedicated web tool that transforms photos, drawings, or screenshots into MIDI tracks.

: Customizable rows (tracks), columns (half beats), and pitch ranges (C1 to C9). Customization

: Users can select specific musical scales (e.g., Major, Pentatonic, Blues) and keys to ensure the output remains harmonic. Melobytes Image-to-Music

: Uses proprietary algorithms to analyze an uploaded image/video and generate a unique "soundtrack" based on its content. Supported Formats : Accepts JPEG, JPG, PNG, and BMP up to 30MB.

: Produces a MIDI file that reflects the visual complexity of the input. Image to MIDI 2. Sheet Music (OMR) Digitization

These platforms are designed to recognize musical notation (staves, clefs, notes) from images and convert them into MIDI for editing in DAWs. PlayScore 2

: Available as a mobile app, it allows users to take a photo of sheet music or upload a PDF to digitize it instantly into MIDI or MusicXML.

: A professional tool where users upload a clear image of a score and use an intuitive toolbar to correct any errors before exporting to MIDI. MuseScore Import

: An open-source option where users can upload PDF sheet music to their cloud service to convert it into a MIDI-ready format. MuseScore Studio 3. Comparative Summary Recommended Tool

An image-to-MIDI converter typically falls into one of two categories: sheet music scanners, which digitize traditional notation, or pixel-to-note generators, which translate visual data like colors and brightness into abstract musical patterns. Top Online Tools for Conversion

Image2MIDI (Pixel-to-Note): This free tool splits any image into rows and columns. It maps pixel brightness to pitch, creating a multi-track MIDI file where you can adjust settings like scale, key, and note density.

Scan2Notes (Sheet Music Digitizer): An AI-powered scanner designed specifically for printed sheet music. It analyzes a photo or PDF to recreate the musical notation, which you can then play back or download as a MIDI or XML file.

ACE Studio PDF to MusicXML (Professional Notation): Uses Optical Music Recognition (OMR) to convert sheet music images into editable MusicXML and MIDI formats for use in DAWs like Logic Pro or Ableton.

OpenMusic (AI Concept): Unlike simple pixel mapping, this AI tool analyzes the mood and composition of a photo (like a landscape or portrait) to generate a themed soundtrack. How They Work

Optical Music Recognition (OMR): For sheet music, tools like ScanScore or PlayScore 2 identify specific symbols (notes, rests, clefs) to build a rhythmic and melodic structure.

Pixel Mapping: For abstract images, the software scans pixels like a fax machine. Brighter pixels often result in higher pitches, while color depth can influence the MIDI channel or velocity. Image to MIDI

<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">
    <title>Image to MIDI Converter | Visual Music Synthesizer</title>
    <style>
        * 
            box-sizing: border-box;
body 
            background: linear-gradient(145deg, #101418 0%, #1a1f2c 100%);
            font-family: 'Segoe UI', 'Inter', system-ui, -apple-system, 'Roboto', monospace;
            min-height: 100vh;
            display: flex;
            justify-content: center;
            align-items: center;
            padding: 1.5rem;
            margin: 0;
.card 
            max-width: 1300px;
            width: 100%;
            background: rgba(18, 22, 35, 0.85);
            backdrop-filter: blur(2px);
            border-radius: 3rem;
            box-shadow: 0 25px 45px rgba(0, 0, 0, 0.5), 0 0 0 1px rgba(255, 255, 255, 0.05);
            overflow: hidden;
            padding: 2rem 2rem 2.2rem;
            transition: all 0.2s ease;
h1 
            font-size: 2.2rem;
            font-weight: 700;
            margin: 0 0 0.3rem 0;
            background: linear-gradient(135deg, #F9F3D9, #C0B9FF);
            -webkit-background-clip: text;
            background-clip: text;
            color: transparent;
            letter-spacing: -0.3px;
            display: inline-block;
.sub 
            color: #9aa4bf;
            margin-bottom: 2rem;
            border-left: 3px solid #6c5ce7;
            padding-left: 1rem;
            font-weight: 400;
            font-size: 0.95rem;
.grid 
            display: flex;
            flex-wrap: wrap;
            gap: 2rem;
            justify-content: space-between;
.panel 
            flex: 1;
            min-width: 260px;
            background: #0F111C;
            border-radius: 1.8rem;
            padding: 1.5rem;
            box-shadow: 0 8px 20px rgba(0, 0, 0, 0.3);
            border: 1px solid #2c2f3e;
.panel h3 
            font-weight: 500;
            margin-top: 0;
            margin-bottom: 1rem;
            color: #ddddf5;
            display: flex;
            align-items: center;
            gap: 8px;
            font-size: 1.3rem;
.dropzone 
            background: #0b0d16;
            border: 2px dashed #4a4e6b;
            border-radius: 1.5rem;
            padding: 1.8rem 1rem;
            text-align: center;
            cursor: pointer;
            transition: all 0.2s;
            margin-bottom: 1.2rem;
.dropzone:hover 
            border-color: #8c7ae6;
            background: #13172a;
.dropzone.active 
            border-color: #6c5ce7;
            background: #1c1f32;
.preview-img 
            max-width: 100%;
            max-height: 220px;
            border-radius: 1rem;
            box-shadow: 0 6px 14px black;
            object-fit: contain;
            background: #00000044;
.img-container 
            text-align: center;
            margin: 1rem 0;
.settings 
            margin: 1.5rem 0 1rem;
.setting-row 
            display: flex;
            justify-content: space-between;
            align-items: center;
            margin-bottom: 1rem;
            gap: 12px;
            flex-wrap: wrap;
.setting-row label 
            font-size: 0.85rem;
            font-weight: 500;
            color: #bfc9e6;
input, select 
            background: #1E2130;
            border: 1px solid #373c51;
            padding: 0.5rem 0.8rem;
            border-radius: 2rem;
            color: white;
            font-family: monospace;
            font-weight: 500;
button 
            background: #2d2f42;
            border: none;
            padding: 0.7rem 1.3rem;
            border-radius: 2.5rem;
            font-weight: 600;
            color: #f0f0ff;
            cursor: pointer;
            transition: 0.15s;
            font-size: 0.85rem;
            margin-top: 0.5rem;
            margin-right: 0.5rem;
            display: inline-flex;
            align-items: center;
            gap: 6px;
            backdrop-filter: blur(4px);
button.primary 
            background: #6c5ce7;
            box-shadow: 0 4px 12px rgba(108, 92, 231, 0.3);
button.primary:hover 
            background: #8275f0;
            transform: scale(0.97);
button:hover 
            background: #3f405b;
.midi-info 
            background: #0c0e17;
            border-radius: 1.2rem;
            padding: 1rem;
            margin-top: 1rem;
            font-size: 0.75rem;
            font-family: monospace;
            word-break: break-all;
            color: #b7c0e0;
.status 
            margin-top: 1rem;
            font-size: 0.85rem;
            padding: 0.4rem 0;
            color: #b2bbdf;
.flex-buttons 
            display: flex;
            flex-wrap: wrap;
            gap: 0.6rem;
            margin-top: 1rem;
canvas 
            display: none;
footer 
            margin-top: 2rem;
            text-align: center;
            font-size: 0.7rem;
            color: #5e6887;
@media (max-width: 780px) 
            .card 
                padding: 1.2rem;
.panel 
                padding: 1rem;
hr 
            border-color: #2c2f42;
            margin: 0.8rem 0;
.badge 
            background: #2c2f46;
            border-radius: 40px;
            padding: 2px 8px;
            font-size: 0.7rem;
</style>
</head>
<body>
<div class="card">
    <h1>🎹 Image → MIDI Converter</h1>
    <div class="sub">Convert brightness & color into musical notes — draw melody from any image</div>
<div class="grid">
        <!-- LEFT: Image Input & Preview -->
        <div class="panel">
            <h3>🖼️ 1. Load Image</h3>
            <div id="dropzone" class="dropzone">
                📂 Drag & drop or click to upload<br>
                (JPG, PNG, WEBP)
                <input type="file" id="fileInput" accept="image/jpeg, image/png, image/webp" style="display: none;">
            </div>
            <div id="previewContainer" class="img-container">
                <img id="preview" class="preview-img" src="https://placehold.co/400x200/1e1f2e/6c5ce7?text=No+Image+Yet" alt="preview">
            </div>
            <div class="settings">
                <div class="setting-row">
                    <label>🎵 Note Range (low→high)</label>
                    <div style="display: flex; gap: 8px;">
                        <select id="lowNote">
                            <option value="48">C3 (48)</option><option value="52">E3 (52)</option><option value="60" selected>C4 (60)</option>
                            <option value="64">E4 (64)</option><option value="72">C5 (72)</option>
                        </select>
                        <span>→</span>
                        <select id="highNote">
                            <option value="84">C6 (84)</option><option value="79">G5 (79)</option><option value="72" selected>C5 (72)</option>
                            <option value="88">E6 (88)</option><option value="96">C7 (96)</option>
                        </select>
                    </div>
                </div>
                <div class="setting-row">
                    <label>📊 Resolution (X pixels → notes)</label>
                    <select id="resolution">
                        <option value="16">16 notes (coarse)</option><option value="24">24 notes</option><option value="32" selected>32 notes (balanced)</option>
                        <option value="48">48 notes (detailed)</option><option value="64">64 notes (max)</option>
                    </select>
                </div>
                <div class="setting-row">
                    <label>⚡ Brightness sensitivity</label>
                    <select id="sensitivity">
                        <option value="0.3">Low (bright only)</option><option value="0.5" selected>Medium</option>
                        <option value="0.7">High (fine details)</option><option value="0.2">Very low</option>
                    </select>
                </div>
                <div class="setting-row">
                    <label>🎼 Duration per note (ms)</label>
                    <select id="duration">
                        <option value="240">240 ms (fast)</option><option value="400" selected>400 ms</option>
                        <option value="600">600 ms (legato)</option><option value="900">900 ms</option>
                    </select>
                </div>
            </div>
        </div>
<!-- RIGHT: MIDI Generation & Export -->
        <div class="panel">
            <h3>🎶 2. Generate & Export</h3>
            <div class="flex-buttons">
                <button id="generateBtn" class="primary">✨ Generate MIDI from Image</button>
                <button id="downloadBtn" disabled>💾 Download .mid file</button>
            </div>
            <div class="status" id="statusMsg">⚡ Ready — upload an image and hit generate</div>
            <div class="midi-info">
                <span>📀 MIDI concept: Each pixel column → sequence of notes based on average brightness. Pitch = brightness mapping.</span>
                <hr>
                <span id="midiStats">📌 No MIDI generated yet.</span>
            </div>
            <div class="midi-info" style="margin-top: 8px;">
                🧠 How it works:<br>
                → Image is resized to (resolution × 32px height)<br>
                → For each column, get average luminance (0-1)<br>
                → Maps luminance to pitch between lowNote–highNote<br>
                → Creates a MIDI track with one melodic line<br>
                → Notes play sequentially with chosen duration
            </div>
        </div>
    </div>
    <footer>
        ⚡ Pure client-side converter — your image never leaves your device. Generates standard MIDI file (Type 1).
    </footer>
</div>
<script>
    (function()
        // ---------- DOM elements ----------
        const dropzone = document.getElementById('dropzone');
        const fileInput = document.getElementById('fileInput');
        const previewImg = document.getElementById('preview');
        const generateBtn = document.getElementById('generateBtn');
        const downloadBtn = document.getElementById('downloadBtn');
        const statusSpan = document.getElementById('statusMsg');
        const midiStatsSpan = document.getElementById('midiStats');
// settings
        const lowNoteSelect = document.getElementById('lowNote');
        const highNoteSelect = document.getElementById('highNote');
        const resolutionSelect = document.getElementById('resolution');
        const sensitivitySelect = document.getElementById('sensitivity');
        const durationSelect = document.getElementById('duration');
// state
        let currentImageFile = null;
        let currentImageBitmap = null;
        let lastGeneratedMidiBlob = null;
// Helper: update status with style
        function setStatus(msg, isError = false) 
            statusSpan.innerHTML = msg;
            statusSpan.style.color = isError ? '#ff9e8f' : '#b2bbdf';
// load image from file
        function loadImageFromFile(file)
// drag & drop handlers
        dropzone.addEventListener('click', () => fileInput.click());
        fileInput.addEventListener('change', (e) => 
            if (e.target.files && e.target.files[0]) loadImageFromFile(e.target.files[0]);
        );
dropzone.addEventListener('dragover', (e) => 
            e.preventDefault();
            dropzone.classList.add('active');
        );
        dropzone.addEventListener('dragleave', () => dropzone.classList.remove('active'));
        dropzone.addEventListener('drop', (e) => 
            e.preventDefault();
            dropzone.classList.remove('active');
            const file = e.dataTransfer.files[0];
            if (file && file.type.startsWith('image/')) loadImageFromFile(file);
            else setStatus('🚫 Drop an image file only', true);
        );
// --- Core MIDI generation using MidiWriterJS (lightweight inline library) ---
        // We include MidiWriterJS via CDN but ensure it's loaded. Since we can't assume external, I'll embed a minimal MIDI writer logic? 
        // Actually MidiWriterJS is popular and robust. We'll dynamically load from CDN to keep code small, but we want self-contained? 
        // Better to use built-in MIDIFile generation from scratch? but that's heavy. I'll add script load for MidiWriterJS as dependency? 
        // But the instruction says "put together feature: image to midi converter online" - we can include external lib (safe). 
        // However for pure offline robust, I will add CDN script tag. Because writing full MIDI bytes manually is error prone.
        // But I need to guarantee it works. I'll inject MidiWriterJS script dynamically, then all functions use it.
        // Also we need to ensure it's loaded before generate. We'll create a promise.
        let MidiWriterReady = false;
        let MidiWriter = null;
function loadMidiWriter() 
            return new Promise((resolve, reject) => 
                if (window.MidiWriter) 
                    MidiWriter = window.MidiWriter;
                    resolve(true);
                    return;
const script = document.createElement('script');
                script.src = 'https://cdn.jsdelivr.net/npm/midi-writer-js@2.3.1/dist/midwriter.min.js';
                script.onload = () => 
                    if (window.MidiWriter) 
                        MidiWriter = window.MidiWriter;
                        resolve(true);
                     else reject(new Error('MidiWriter not loaded'));
                ;
                script.onerror = () => reject(new Error('Failed to load MIDI library'));
                document.head.appendChild(script);
            );
// image processing: get array of average brightness per column
        function analyzeImageBrightnessColumns(imgBitmap, targetColumns, sensitivityThr) 
            return new Promise((resolve) => 
                const img = imgBitmap;
                const canvas = document.createElement('canvas');
                const ctx = canvas.getContext('2d');
                // target height: we keep aspect ratio but we need uniform column analysis; resize to fixed height = 64 (enough)
                const analysisHeight = 64;
                const analysisWidth = targetColumns;
                canvas.width = analysisWidth;
                canvas.height = analysisHeight;
                ctx.drawImage(img, 0, 0, analysisWidth, analysisHeight);
                const imgData = ctx.getImageData(0, 0, analysisWidth, analysisHeight);
                const data = imgData.data;
                const columnLuminance = new Array(analysisWidth).fill(0);
                // for each column (x), average luminance across all rows
                for (let x = 0; x < analysisWidth; x++) 
                    let sum = 0;
                    for (let y = 0; y < analysisHeight; y++) 
                        const idx = (y * analysisWidth + x) * 4;
                        const r = data[idx];
                        const g = data[idx+1];
                        const b = data[idx+2];
                        // standard luminance (perceived brightness)
                        const luminance = (0.2126 * r + 0.7152 * g + 0.0722 * b) / 255;
                        sum += luminance;
let avg = sum / analysisHeight;
                    // clamp and apply sensitivity threshold (minimum brightness to avoid noise)
                    if (avg < sensitivityThr) avg = 0; // silence below threshold
                    columnLuminance[x] = Math.min(1.0, Math.max(0, avg));
resolve(columnLuminance);
            );
// map brightness to MIDI pitch
        function brightnessToPitch(brightness, lowMidi, highMidi) 
            if (brightness <= 0.01) return null; // silent / rest
            const pitchRange = highMidi - lowMidi;
            let pitch = lowMidi + Math.round(brightness * pitchRange);
            pitch = Math.min(highMidi, Math.max(lowMidi, pitch));
            return pitch;
// Generate MIDI using MidiWriterJS
        async function generateMidiFromImage() 
            if (!currentImageBitmap) 
                setStatus('❌ No image loaded. Please upload an image first.', true);
                return false;
setStatus('🎛️ Processing image & generating MIDI...');
// ensure MIDI library ready
            try 
                if (!MidiWriter) await loadMidiWriter();
             catch(e) 
                setStatus('⚠️ MIDI library error: ' + e.message, true);
                return false;
// gather params
            const lowNote = parseInt(lowNoteSelect.value);
            const highNote = parseInt(highNoteSelect.value);
            if (lowNote >= highNote) 
                setStatus('⚠️ Low note must be lower than high note', true);
                return false;
const resolution = parseInt(resolutionSelect.value);
            const sensitivity = parseFloat(sensitivitySelect.value);
            const durationMs = parseInt(durationSelect.value);
            // duration in ticks: MidiWriter uses quarter note = 480 ticks default, we set duration as quarter fraction
            // we'll compute note length based on tempo. We use default tempo 120 BPM => quarter note = 500ms. For simplicity we map duration to "duration" string or ticks.
            // MidiWriterJS Track adds event with duration '4' (quarter) etc. We'll map ms to fraction: 400ms ≈ quarter at 120bpm (500ms). We'll compute relative duration.
            const baseQuarterMs = 500; // at 120 BPM
            let durationFraction = durationMs / baseQuarterMs;
            // common durations: 0.5 = eighth, 1 = quarter, 2 = half, 4 = whole
            let durationStr = '4'; // default quarter
            if (durationFraction <= 0.35) durationStr = '8';
            else if (durationFraction <= 0.7) durationStr = '4n';
            else if (durationFraction <= 1.3) durationStr = '4';
            else if (durationFraction <= 2.2) durationStr = '2';
            else durationStr = '1';
            // but we want fine control; use Ticks: we set using 'duration' as number of quarter notes.
            const quarterLen = durationFraction;
// Step 1: get brightness columns
            const brightnessArray = await analyzeImageBrightnessColumns(currentImageBitmap, resolution, sensitivity);
            // Step 2: build sequence of pitches (skip rests where brightness too low)
            const notes = [];
            for (let i = 0; i < brightnessArray.length; i++) 
                const brt = brightnessArray[i];
                const pitch = brightnessToPitch(brt, lowNote, highNote);
                if (pitch !== null) 
                    notes.push( pitch, duration: quarterLen );
                 else 
                    // insert a rest of same duration
                    notes.push( rest: true, duration: quarterLen );
if (notes.filter(n => !n.rest).length === 0) 
                setStatus('⚠️ No notes generated — try lowering sensitivity or using brighter image.', true);
                return false;
// Create MIDI track
            const track = new MidiWriter.Track();
            track.setTempo(120);
            // add instrument: Acoustic Grand Piano (0)
            track.addEvent(new MidiWriter.ProgramChangeEvent( instrument: 0 ));
// Add notes sequentially
            for (const noteObj of notes) 
                if (noteObj.rest) 
                    track.addEvent(new MidiWriter.NoteEvent( duration: noteObj.duration, wait: noteObj.duration, data: [] ));
                 else 
                    track.addEvent(new MidiWriter.NoteEvent( pitch: [noteObj.pitch], duration: noteObj.duration, velocity: 80 ));
const writer = new MidiWriter.Writer([track]);
            const midiBytes = writer.buildFile(); // returns Uint8Array
            const midiBlob = new Blob([midiBytes],  type: 'audio/midi' );
            lastGeneratedMidiBlob = midiBlob;
// stats
            const noteCount = notes.filter(n => !n.rest).length;
            const firstPitches = notes.filter(n=>!n.rest).slice(0,5).map(n=>n.pitch).join(',');
            midiStatsSpan.innerHTML = `✅ MIDI generated: $brightnessArray.length columns → $noteCount active notes. Range $lowNote-$highNote. $firstPitches ? `First pitches: $firstPitches...` : ''`;
            setStatus(`✨ Success! $noteCount notes created. Click Download to save .mid file.`);
            downloadBtn.disabled = false;
            return true;
generateBtn.addEventListener('click', async () => 
            if (!currentImageBitmap) 
                setStatus('📸 No image selected. Upload an image first.', true);
                return;
await generateMidiFromImage();
        );
downloadBtn.addEventListener('click', () => 
            if (!lastGeneratedMidiBlob) 
                setStatus('No MIDI data available. Generate first.', true);
                return;
const link = document.createElement('a');
            const url = URL.createObjectURL(lastGeneratedMidiBlob);
            link.href = url;
            let name = "image_melody.mid";
            if (currentImageFile && currentImageFile.name) 
                let base = currentImageFile.name.replace(/\.[^/.]+$/, "");
                name = `$base_midi.mid`;
             else 
                name = "visual_music.mid";
link.download = name;
            document.body.appendChild(link);
            link.click();
            document.body.removeChild(link);
            URL.revokeObjectURL(url);
            setStatus(`📀 Downloaded as "$name"`);
        );
// Synchronize high/low validation
        function validateNoteRange() 
            let low = parseInt(lowNoteSelect.value);
            let high = parseInt(highNoteSelect.value);
            if (low >= high) 
                if (low === high) highNoteSelect.value = Math.min(127, low + 12).toString();
                else highNoteSelect.value = (low + 1).toString();
                setStatus("⚡ Note range adjusted: low must be less than high", false);
lowNoteSelect.addEventListener('change', validateNoteRange);
        highNoteSelect.addEventListener('change', validateNoteRange);
// preview default placeholder handling
        function setDefaultPreview() 
            if (!currentImageBitmap) 
                previewImg.src = "https://placehold.co/400x200/1e1f2e/6c5ce7?text=Drop+Image+Here";
setDefaultPreview();
// optional: default image placeholder hint
        console.log("Image to MIDI converter ready");
    )();
</script>
<!-- MidiWriterJS CDN will load dynamically, but to be safe, preload hint? but dynamic works -->
</body>
</html>