feature: comment on annotations

app/www/annotate.html

@@ -98,6 +98,7 @@
             position: absolute;
             left: 100%;
             width: 30px;
+            height:30px;
         }
 
         .controls button.paused::before {
@@ -108,6 +109,22 @@
             content: '⏸';
         }
 
+
+        .buffering .controls button:is(.playing,.paused)::before {
+            content: '↺';
+            display:inline-block;
+            animation: rotate 1s infinite;
+        }
+
+        @keyframes rotate {
+          0% {
+            transform: rotate(359deg)
+          }
+          100% {
+            transform: rotate(0deg)
+          }
+        }
+
         .controls {
             position: absolute !important;
             z-index: 100;
@@ -182,6 +199,14 @@
             pointer-events: all;
         }
 
+        .controls .annotation-comment{
+            width: 100%;
+            visibility: hidden;
+        }
+        .selected-annotation .controls .annotation-comment{
+            visibility: visible;
+        }
+
         .noUi-handle:focus {
             /* background: red;; */
             border: solid 2px #601be0;
@@ -376,6 +401,7 @@
         navigator.mediaSession.setActionHandler('seekforward', function () { /* Code excerpted. */ });
         navigator.mediaSession.setActionHandler('previoustrack', function () { /* Code excerpted. */ });
         navigator.mediaSession.setActionHandler('nexttrack', function () { /* Code excerpted. */ });
+        navigator.mediaSession.setActionHandler('playpause', function () { /* Code excerpted. */ });
     </script>
 </body>
 

app/www/annotate.js

@@ -1,8 +1,9 @@
 class Annotation {
-    constructor(tag, t_in, t_out) {
+    constructor(tag, t_in, t_out, comment) {
         this.tag = tag;
         this.t_in = Number.parseFloat(t_in);
         this.t_out = Number.parseFloat(t_out);
+        this.comment = comment;
     }
 }
 
@@ -110,9 +111,14 @@ class StrokeSlice {
 }
 
 class Annotator extends EventTarget {
-    constructor(wrapperEl, tagFile, fileurl) {
+    constructor(wrapperEl, tagFile, fileurl, config) {
         super();
 
+        this.config = {
+            is_player: config && config.hasOwnProperty('is_player') ? config.is_player : false, // in player mode annotations are not loaded, nor is the annotator shown
+            crop_to_fit: config && config.hasOwnProperty('crop_to_fit') ? config.crop_to_fit : false, // don't animate viewport, but show the whole drawing
+        }
+
         this.formatter = wNumb({
             decimals: 2,
             edit: (time) => {
@@ -142,6 +148,7 @@ class Annotator extends EventTarget {
         this.wrapperEl = wrapperEl;
         this.svgEl = document.createElementNS('http://www.w3.org/2000/svg', 'svg');
         this.wrapperEl.appendChild(this.svgEl);
+        this.wrapperEl.classList.add(this.config.is_player ? "svganim_player" : "svganim_annotator");
 
 
         this.controlsEl = document.createElement('div');
@@ -161,7 +168,7 @@ class Annotator extends EventTarget {
         this.playheadEl.addEventListener("input", (ev) => {
             this.scrubTo(ev.target.value);
         });
-        this.playheadEl.addEventListener('keydown',(ev) => {
+        this.playheadEl.addEventListener('keydown', (ev) => {
             ev.preventDefault(); // we don't want to use arrow keys, as these are captured in the overall keydown event
         })
 
@@ -178,7 +185,7 @@ class Annotator extends EventTarget {
         this.playPauseEl.addEventListener("click", (ev) => {
             this.playPause()
         })
-        this.playPauseEl.addEventListener('keydown',(ev) => {
+        this.playPauseEl.addEventListener('keydown', (ev) => {
             ev.preventDefault(); // we don't want to spacebar, as this is captured in the overall keydown event
         })
 
@@ -191,77 +198,101 @@ class Annotator extends EventTarget {
         this.annotationsEl.classList.add('annotations')
         this.controlsEl.appendChild(this.annotationsEl);
 
-        this.loadTags(tagFile).then(() => {
-            this.tagsEl = document.createElement('ul');
-            this.tagsEl.classList.add('tags');
-            const addTags = (tags, tagsEl) => {
-                Object.entries(tags).forEach(([tag, tagData]) => {
-                    let tagLiEl = document.createElement('li');
-                    let tagEl = document.createElement('div');
 
-                    tagEl.classList.add('tag');
-                    tagEl.dataset.tag = tag;
-                    tagEl.innerText = tagData.hasOwnProperty('fullname') ? tagData.fullname : tag;
-                    tagEl.addEventListener('click', (e) => {
-                        this.addTag(tag, this.inPointPosition, this.outPointPosition);
-                    });
+        this.inPointPosition = [0, 0];
+        this.inPointTimeMs = null;
+        this.outPointPosition = null;
+        this.outPointTimeMs = null;
+        this._currentTimeMs = 0;
+        this.videoIsPlaying = false;
 
-                    tagEl.title = tagData.hasOwnProperty('description') ? tagData.description : "";
-
-                    let signEl = document.createElement('span');
-                    signEl.classList.add('annotation-' + tag);
-                    signEl.style.backgroundColor = this.getColorForTag(tag);
-                    tagEl.prepend(signEl);
-
-                    tagLiEl.appendChild(tagEl);
-
-                    if (tagData.hasOwnProperty('sub')) {
-                        const subEl = document.createElement('ul');
-                        subEl.classList.add('subtags');
-                        addTags(tagData.sub, subEl);
-                        tagLiEl.appendChild(subEl);
-                    }
-
-                    tagsEl.appendChild(tagLiEl);
-                });
-            };
-            addTags(this.tags, this.tagsEl);
-
-            let tagEl = document.createElement('li');
-            tagEl.classList.add('tag');
-            tagEl.classList.add('annotation-rm');
-            tagEl.dataset.tag = 'rm';
-            tagEl.title = "Remove annotation";
-            tagEl.innerHTML = "🚮"; // ×
-            tagEl.addEventListener('click', (e) => {
-                if (this.selectedAnnotation) {
-                    this.removeAnnotation(this.selectedAnnotationI);
-                }
-            });
-            this.tagsEl.appendChild(tagEl);
-
-            this.controlsEl.appendChild(this.tagsEl);
-
-            this.inPointPosition = [0, 0];
-            this.inPointTimeMs = null;
-            this.outPointPosition = null;
-            this.outPointTimeMs = null;
-            this._currentTimeMs = 0;
-            this.videoIsPlaying = false;
-
-            const groups = ['before', 'annotation', 'after']
-            this.strokeGroups = {};
-            groups.forEach(group => {
-                let groupEl = document.createElementNS('http://www.w3.org/2000/svg', 'g');
-                groupEl.classList.add(group)
-                this.svgEl.appendChild(groupEl);
-                this.strokeGroups[group] = new StrokeGroup(groupEl, this);
-            });
-
-            this.annotations = [];
-
-            this.load(fileurl);
+        const groups = ['before', 'annotation', 'after']
+        this.strokeGroups = {};
+        groups.forEach(group => {
+            let groupEl = document.createElementNS('http://www.w3.org/2000/svg', 'g');
+            groupEl.classList.add(group)
+            this.svgEl.appendChild(groupEl);
+            this.strokeGroups[group] = new StrokeGroup(groupEl, this);
         });
+
+        this.annotations = [];
+
+        if (this.config.is_player) {
+            this.load(fileurl);
+        } else {
+
+            this.loadTags(tagFile).then(() => {
+                this.tagsEl = document.createElement('ul');
+                this.tagsEl.classList.add('tags');
+                const addTags = (tags, tagsEl) => {
+                    Object.entries(tags).forEach(([tag, tagData]) => {
+                        let tagLiEl = document.createElement('li');
+                        let tagEl = document.createElement('div');
+
+                        tagEl.classList.add('tag');
+                        tagEl.dataset.tag = tag;
+                        tagEl.innerText = tagData.hasOwnProperty('fullname') ? tagData.fullname : tag;
+                        tagEl.addEventListener('click', (e) => {
+                            this.addTag(tag, this.inPointPosition, this.outPointPosition);
+                        });
+
+                        tagEl.title = tagData.hasOwnProperty('description') ? tagData.description : "";
+
+                        let signEl = document.createElement('span');
+                        signEl.classList.add('annotation-' + tag);
+                        signEl.style.backgroundColor = this.getColorForTag(tag);
+                        tagEl.prepend(signEl);
+
+                        tagLiEl.appendChild(tagEl);
+
+                        if (tagData.hasOwnProperty('sub')) {
+                            const subEl = document.createElement('ul');
+                            subEl.classList.add('subtags');
+                            addTags(tagData.sub, subEl);
+                            tagLiEl.appendChild(subEl);
+                        }
+
+                        tagsEl.appendChild(tagLiEl);
+                    });
+                };
+                addTags(this.tags, this.tagsEl);
+
+                let tagEl = document.createElement('li');
+                tagEl.classList.add('tag');
+                tagEl.classList.add('annotation-rm');
+                tagEl.dataset.tag = 'rm';
+                tagEl.title = "Remove annotation";
+                tagEl.innerHTML = "🚮"; // ×
+                tagEl.addEventListener('click', (e) => {
+                    if (this.selectedAnnotation) {
+                        this.removeAnnotation(this.selectedAnnotationI);
+                    }
+                });
+                this.tagsEl.appendChild(tagEl);
+
+                this.controlsEl.appendChild(this.tagsEl);
+
+                this.commentEl = document.createElement('input');
+                this.commentEl.type = 'text';
+                this.commentEl.classList.add('annotation-comment');
+                this.commentEl.title = "Add comment to annotation";
+                this.commentEl.placeholder = "comment";
+                this.commentEl.value = "";
+                this.commentEl.addEventListener('keyup', (e) => {
+                    e.stopPropagation(); // prevent keyup event to propagate and set i/o points
+                });
+                this.commentEl.addEventListener('input', (e) => {
+                    e.stopPropagation(); // prevent keyup event 
+                    if (this.selectedAnnotation) {
+                        this.selectedAnnotation.comment = this.commentEl.value;
+                        this.updateAnnotations(true)
+                    }
+                });
+                this.controlsEl.appendChild(this.commentEl);
+
+                this.load(fileurl);
+            });
+        }
     }
 
     getColorForTag(tag) {
@@ -278,6 +309,10 @@ class Annotator extends EventTarget {
 
     updateAnnotations(save) {
 
+        if (this.config.is_player) {
+            return false;
+        }
+
         this.annotationsEl.innerHTML = "";
         for (let annotation_i in this.annotations) {
             const annotation = this.annotations[annotation_i];
@@ -295,7 +330,7 @@ class Annotator extends EventTarget {
             if (this.selectedAnnotationI == annotation_i) {
                 this.annotationEl.classList.add('selected');
             }
-            this.annotationEl.title = annotation.tag;
+            this.annotationEl.title = `[${annotation.tag}] ${annotation.comment}`;
 
             this.annotationEl.addEventListener('mouseover', (e) => {
 
@@ -344,6 +379,7 @@ class Annotator extends EventTarget {
         this.updateAnnotations(false); //selects the right tag & highlights the annotation
 
         this.wrapperEl.classList.add('selected-annotation');
+        this.commentEl.value = this.selectedAnnotation.comment;
     }
 
     deselectAnnotation(keep_position) {
@@ -352,6 +388,7 @@ class Annotator extends EventTarget {
         }
 
         this.wrapperEl.classList.remove('selected-annotation');
+        this.commentEl.value = "";
 
         this.selectedAnnotationI = null;
         this.selectedAnnotation = null;
@@ -409,20 +446,24 @@ class Annotator extends EventTarget {
         fetch(request)
             .then(response => response.json())
             .then(data => {
-                const metadata_req = new Request(`/annotations/${data.file}`, {
-                    method: 'GET',
-                });
-                fetch(metadata_req)
-                    .then(response => response.ok ? response.json() : null)
-                    .then(metadata => {
-                        if (metadata !== null) {
-                            metadata.annotations = metadata.annotations.map((a) => new Annotation(a.tag, a.t_in, a.t_out))
-                        }
-                        this.loadStrokes(data, metadata)
-                    })
-                    .catch(e => console.log(e));
-                // do something with the data sent in the request
-            });
+                if (!this.config.is_player) {
+
+                    const metadata_req = new Request(`/annotations/${data.file}`, {
+                        method: 'GET',
+                    });
+                    fetch(metadata_req)
+                        .then(response => response.ok ? response.json() : null)
+                        .then(metadata => {
+                            if (metadata !== null) {
+                                metadata.annotations = metadata.annotations.map((a) => new Annotation(a.tag, a.t_in, a.t_out, a.hasOwnProperty('comment') ? a.comment : ""))
+                            }
+                            this.loadStrokes(data, metadata)
+                        })
+                        .catch(e => console.log(e));
+                } else {
+                    this.loadStrokes(data, null);
+                }
+            }).catch(e => console.log(e));
     }
 
     updateState() {
@@ -497,7 +538,7 @@ class Annotator extends EventTarget {
                 this.slider.destroy();
             }
 
-            this.annotations.push(new Annotation(tag, t_in, t_out));
+            this.annotations.push(new Annotation(tag, t_in, t_out, ""));
             this.updateAnnotations(true);
 
             this._currentTimeMs = t_out;
@@ -508,15 +549,27 @@ class Annotator extends EventTarget {
 
 
     setUpAnnotator() {
+
         this.playheadEl.min = this.audioOffset < 0 ? this.audioOffset * 1000 : 0;
         this.playheadEl.max = this.getEndTimeMs();
         this._updatePlayhead();
 
+
         this.inPointPosition = this.findPositionForTime(this.currentTime);
         this.inPointTimeMs = this._currentTimeMs;
         this.outPointPosition = this.findPositionForTime(this.lastFrameTime); // TODO: simplify to get the last frame indexes directly
         this.outPointTimeMs = this.getEndTimeMs();
 
+
+        if (!this.config.is_player) {
+            this.buildAnnotator();
+        }
+
+        this.drawStrokePosition(this.inPointPosition, this.outPointPosition);
+    }
+
+    buildAnnotator() {
+
         if (this.scrubberEl.noUiSlider) {
             this.slider.destroy();
         }
@@ -597,21 +650,27 @@ class Annotator extends EventTarget {
                 ttInputEl.addEventListener('blur', submit);
             });
         })
-
-        this.drawStrokePosition(this.inPointPosition, this.outPointPosition);
     }
 
     loadStrokes(drawing, metadata) {
         this.audioOffset = 0;
+
         if (metadata) {
             this.annotations = metadata.annotations;
-            this.audioFile = metadata.hasOwnProperty('audio') ? metadata.audio.file : null;
-            this.audioOffset = metadata.hasOwnProperty('audio') ? Number.parseFloat(metadata.audio.offset) : 0;
+        }
+
+        if ((metadata && metadata.hasOwnProperty('audio')) || drawing.hasOwnProperty('audio')) {
+            if (metadata && metadata.hasOwnProperty('audio')) {
+                this.audioFile = metadata.audio.file
+                this.audioOffset = Number.parseFloat(metadata.audio.offset);
+            } else {
+                this.audioFile = drawing.audio.file
+                this.audioOffset = Number.parseFloat(drawing.audio.offset);
+            }
             this._currentTimeMs = this.audioOffset < 0 ? this.audioOffset * 1000 : 0;
             this._updatePlayhead();
-            //
-            // load any saved metadata
         }
+
         this.filename = drawing.file;
         this.strokes = drawing.shape.map(s => new Stroke(s['color'], s['points']));
         this.viewboxes = drawing.viewboxes;
@@ -619,7 +678,11 @@ class Annotator extends EventTarget {
         this.currentPointI = null;
         this.currentViewboxI = null;
         this.dimensions = drawing.dimensions;
-        this.svgEl.setAttribute('viewBox', `0 0 ${this.dimensions[0]} ${this.dimensions[1]}`)
+        if (!this.config.crop_to_fit) {
+            this.svgEl.setAttribute('viewBox', `0 0 ${this.dimensions[0]} ${this.dimensions[1]}`)
+        } else {
+            this.svgEl.setAttribute('viewBox', `${drawing.bounding_box.x} ${drawing.bounding_box.y} ${drawing.bounding_box.width} ${drawing.bounding_box.height}`)
+        }
 
         // let bgEl = document.createElementNS('http://www.w3.org/2000/svg', 'rect');
         // bgEl.setAttribute("x", 0);
@@ -638,34 +701,19 @@ class Annotator extends EventTarget {
 
         this.setupAudioConfig().then(() => {
             // this.setUpAnnotator()
-            this.updateAnnotations(false);
+            let keyEl;
+            if (this.config.is_player) {
+                keyEl = this.wrapperEl;
+            } else {
+                keyEl = document.body; // always capture
+                this.updateAnnotations(false);
+            }
 
-            document.body.addEventListener('keyup', (ev) => {
+            keyEl.addEventListener('keyup', (ev) => {
                 if (ev.key == ' ') {
                     this.playPause();
                 }
-                console.log('key', ev);
-                if (ev.key == 'i') {
-                    this.setInPoint(this.currentTime * 1000);
-                }
-                if (ev.key == 'o') {
-                    this.setOutPoint(this.currentTime * 1000);
-                }
-                if (ev.key == 'I') {
-                    // shift+i == jump to in point
-                    this.scrubTo(this.inPointTimeMs);
-                }
-                if (ev.key == 'O') {
-                    // shift+o == jump to end point
-                    this.scrubTo(this.outPointTimeMs);
-                }
-                if (ev.key == 'Escape') {
-                    if (this.selectedAnnotation) {
-                        this.deselectAnnotation();
-                    } else {
-                        this.resetInOutPoint();
-                    }
-                }
+
                 // shift+arrow keys, jump playhead (search position)
                 // FIXME doesn't keep playback after initial load. Only after unfocussing the window, and refocussing it, do the keys capture.
                 // Probably a wrong order
@@ -683,6 +731,31 @@ class Annotator extends EventTarget {
                     this.scrubTo(this._currentTimeMs + diff);
                     if (!p) { console.log('play!'); this.play(); } // scrubTo() causes a pause();
                 }
+
+                // additional keys only for annotation mode
+                if (!this.config.is_player) {
+                    if (ev.key == 'i') {
+                        this.setInPoint(this.currentTime * 1000);
+                    }
+                    if (ev.key == 'o') {
+                        this.setOutPoint(this.currentTime * 1000);
+                    }
+                    if (ev.key == 'I') {
+                        // shift+i == jump to in point
+                        this.scrubTo(this.inPointTimeMs);
+                    }
+                    if (ev.key == 'O') {
+                        // shift+o == jump to end point
+                        this.scrubTo(this.outPointTimeMs);
+                    }
+                    if (ev.key == 'Escape') {
+                        if (this.selectedAnnotation) {
+                            this.deselectAnnotation();
+                        } else {
+                            this.resetInOutPoint();
+                        }
+                    }
+                }
             });
         });
 
@@ -714,56 +787,64 @@ class Annotator extends EventTarget {
     setupAudioConfig() {
         // audio config
         return new Promise((resolve, reject) => {
-
-            let audioConfigEl = document.createElement('div');
-            audioConfigEl.classList.add('audioconfig')
-            this.wrapperEl.appendChild(audioConfigEl);
-
-            let audioSelectEl = document.createElement('select');
-            audioSelectEl.classList.add('audioselect');
-            audioConfigEl.appendChild(audioSelectEl);
-
-            fetch('/audio')
-                .then(response => response.json())
-                .then(data => {
-                    data.unshift(''); // add empty, to deselect any file
-                    data.forEach(audioFile => {
-                        let optionEl = document.createElement('option');
-                        optionEl.selected = this.audioFile == audioFile;
-                        optionEl.innerText = audioFile;
-                        audioSelectEl.appendChild(optionEl);
-                    });
-                })
-
-            audioSelectEl.addEventListener('change', (ev) => {
-                this.setAudioFile(ev.target.value);
-            });
-
-
-            let audioOffsetTextEl = document.createElement('label');
-            audioOffsetTextEl.innerText = "Offset (s)";
-            audioConfigEl.appendChild(audioOffsetTextEl);
-
-            let audioOffsetEl = document.createElement('input');
-            audioOffsetEl.setAttribute('type', 'number');
-            audioOffsetEl.setAttribute('step', '.01');
-            audioOffsetEl.value = this.audioOffset ?? 0;
-            audioOffsetEl.addEventListener('change', (ev) => {
-                this.setAudioOffset(ev.target.value);
-            });
-            audioOffsetTextEl.appendChild(audioOffsetEl);
-
-
             this.audioEl = document.createElement('audio');
-            this.audioEl.setAttribute('controls', true);
+            if (!this.config.is_player)
+                this.audioEl.setAttribute('controls', true);
+
             this.audioEl.addEventListener('canplaythrough', (ev) => {
                 console.log('loaded audio', ev);
-                this.audioEl.play();
+                // this.audioEl.play();
             });
-            // this.audioEl.addEventListener('seeked', (ev)=>{
-            //     console.log(ev);
-            // })
-            audioConfigEl.prepend(this.audioEl);
+
+            if (this.config.is_player) {
+                this.wrapperEl.prepend(this.audioEl);
+            }
+            else {
+
+                let audioConfigEl = document.createElement('div');
+                audioConfigEl.classList.add('audioconfig')
+                this.wrapperEl.appendChild(audioConfigEl);
+
+                audioConfigEl.prepend(this.audioEl);
+
+                let audioSelectEl = document.createElement('select');
+                audioSelectEl.classList.add('audioselect');
+                audioConfigEl.appendChild(audioSelectEl);
+
+                fetch('/audio')
+                    .then(response => response.json())
+                    .then(data => {
+                        data.unshift(''); // add empty, to deselect any file
+                        data.forEach(audioFile => {
+                            let optionEl = document.createElement('option');
+                            optionEl.selected = this.audioFile == audioFile;
+                            optionEl.innerText = audioFile;
+                            audioSelectEl.appendChild(optionEl);
+                        });
+                    })
+
+                audioSelectEl.addEventListener('change', (ev) => {
+                    this.setAudioFile(ev.target.value);
+                });
+
+
+                let audioOffsetTextEl = document.createElement('label');
+                audioOffsetTextEl.innerText = "Offset (s)";
+                audioConfigEl.appendChild(audioOffsetTextEl);
+
+                let audioOffsetEl = document.createElement('input');
+                audioOffsetEl.setAttribute('type', 'number');
+                audioOffsetEl.setAttribute('step', '.01');
+                audioOffsetEl.value = this.audioOffset ?? 0;
+                audioOffsetEl.addEventListener('change', (ev) => {
+                    this.setAudioOffset(ev.target.value);
+                });
+                audioOffsetTextEl.appendChild(audioOffsetEl);
+
+            }
+
+            
+
 
             this.audioEl.addEventListener('loadedmetadata', (ev) => {
                 // resolve the 'set up audio' when metadata has loaded
@@ -827,11 +908,15 @@ class Annotator extends EventTarget {
                 console.debug('delay audio playback', t_start, t_diff);
                 // a negative audiooffset delays playback from the start
                 // this.audioStartTimeout = setTimeout((e) => this.audioEl.play(), t*-1000);
-                this.audioStartTimeout = setTimeout((e) => { this.audioEl.currentTime = 0 }, t_start * -1); // triggers play with "seeked" event
+                this.audioStartTimeout = setTimeout((e) => { this.audioEl.currentTime = 0; this.audioEl.play(); }, t_start * -1); // triggers play with "seeked" event
                 // this.audioEl.currentTime = 0;
             }
         } else {
-            this.audioEl.currentTime = t_start / 1000;
+            if (this.audioEl.currentTime !== t_start / 1000) {
+                console.log(this.audioEl.currentTime, t_start / 1000);
+                this.audioEl.currentTime = t_start / 1000;
+            }
+            this.audioEl.play();
             // this.audioEl.play(); // play is done in "seeked" evenlistener
             console.log(this.audioEl.currentTime, t_start, t_in, t_out);
         }
@@ -844,6 +929,10 @@ class Annotator extends EventTarget {
         }, t_diff);
     }
 
+    _scrubAudio(time_ms) {
+        this.audioEl.currentTime = Math.max(0, this.getAudioTime(time_ms)) / 1000;
+    }
+
     getFinalFrameTime() {
         const points = this.strokes[this.strokes.length - 1].points;
         return points[points.length - 1][3];
@@ -924,7 +1013,9 @@ class Annotator extends EventTarget {
         }
         this.currentViewboxI = box_i
         const b = this.viewboxes[box_i];
-        this.svgEl.setAttribute('viewBox', `${b.x} ${b.y} ${this.dimensions[0]} ${this.dimensions[1]}`)
+        if (!this.config.crop_to_fit) {
+            this.svgEl.setAttribute('viewBox', `${b.x} ${b.y} ${this.dimensions[0]} ${this.dimensions[1]}`)
+        }
     }
 
     getNextPosition(path_i, point_i) {
@@ -1053,25 +1144,42 @@ class Annotator extends EventTarget {
                 this._seekByTimeMs(this._currentTimeMs); // prevent playback issue for initial load
             }
 
-            this.startTimeMs = window.performance.now() - this._currentTimeMs;
-            // strokes
-            if (this._currentTimeMs < 0) {
-                this.startVideoTimeout = setTimeout((e) => this.playStrokePosition(this.currentPathI, this.currentPointI), this._currentTimeMs * -1);
-            } else {
-                this.playStrokePosition(this.currentPathI, this.currentPointI);
-            }
-            // viewboxes
-            // const nextViewboxI = Math.max(this.currentViewboxI++, this.viewboxes.length-1);
-            this.playViewboxPosition(this.currentViewboxI);
-
-            // audio
-            this.playAudioSegment(this._currentTimeMs, this.outPointTimeMs);
-            // this.playStrokePosition(this.currentPathI, this.currentPointI);
             this._setPausedFlag(false);
 
-            this.dispatchEvent(new CustomEvent('play', {}));
-            this._animationFrame();
-            resolve();
+            const startPlayback = () => {
+                console.log('start playback');
+                this.startTimeMs = window.performance.now() - this._currentTimeMs;
+                // strokes
+                if (this._currentTimeMs < 0) {
+                    this.startVideoTimeout = setTimeout((e) => this.playStrokePosition(this.currentPathI, this.currentPointI), this._currentTimeMs * -1);
+                } else {
+                    this.playStrokePosition(this.currentPathI, this.currentPointI);
+                }
+                // viewboxes
+                // const nextViewboxI = Math.max(this.currentViewboxI++, this.viewboxes.length-1);
+                this.playViewboxPosition(this.currentViewboxI);
+
+                // audio
+                // TODO: use this.audioEl.readyState == 4 : play immediately, otherwise after event
+                this.playAudioSegment(this._currentTimeMs, this.outPointTimeMs);
+                // this.playStrokePosition(this.currentPathI, this.currentPointI);
+
+                this.dispatchEvent(new CustomEvent('play', {}));
+                this._animationFrame();
+                resolve();
+            }
+
+            if (this.audioEl.readyState !== 4) { // not ready to play after seeking audio.
+                console.log('wait for audio before playback');
+                this.wrapperEl.classList.add('buffering');
+                this.audioEl.addEventListener('canplaythrough', () => {
+                    this.wrapperEl.classList.remove('buffering');
+                    startPlayback()
+                }, { once: true }); // only once
+            } else {
+                startPlayback();
+            }
+
         });
     }
 
@@ -1145,6 +1253,7 @@ class Annotator extends EventTarget {
     _seekByPoint(point) {
         this.dispatchEvent(new CustomEvent('seeking', {}));
         this._currentTimeMs = this.strokes[point[0]].points[point[1]][2];
+        this.audioEl.currentTime = this.getAudioTime(this._currentTimeMs) / 1000;
         [this.currentPathI, this.currentPointI] = point;
         this._updatePlayhead();
         this._updateFrame();
@@ -1158,6 +1267,7 @@ class Annotator extends EventTarget {
     _seekByTime(time) {
         this.dispatchEvent(new CustomEvent('seeking', { detail: time }));
         this._currentTimeMs = Number.parseFloat(time) * 1000;
+        this.audioEl.currentTime = this.getAudioTime(this._currentTimeMs) / 1000;
         [this.currentPathI, this.currentPointI] = this.findPositionForTime(this._currentTimeMs);
 
         this._updatePlayhead();