SoftPointer · Sep 28, 2026 · Sep 28, 2026
@@ -5,7 +5,7 @@
 <section id="examples"><h2>Qualitative editing examples</h2>
 <h3>Action substitution</h3><p>The requested change replaces walking toward the door with drinking water. In the shown Gemini Omni example, drinking is added near the end while much of the walking remains. Check both the presence of the target event and removal of the source event. <a href="presentation.html#action-case">View the example</a>.</p>
 <h3>Clothing edits across turns</h3><p>The sequence edits clothing colors, preserves the edits in the next turn, then restores both colors. It does not test undoing only one change while preserving another. <a href="presentation.html#multi-turn-case">View the sequence</a>.</p>
-<h3>Dog–bone interaction</h3><p>The source contains a yellow-brown bone on the ground. The archived target has a differently shaped white bone in the mouth from the first frame, so it does not demonstrate the original bone being picked up. The Gemini Omni Flash output leaves the original bone on the ground and changes the dog's coat. Model attribution follows the presenter's identification of the result. These three clips belong to one example; they are not three independent trials. <a href="presentation.html#dog-object-case">View the comparison</a>.</p>
+<h3>Dog–bone interaction</h3><p>The source contains a yellow-brown bone on the ground. The archived target has a differently shaped white bone in the mouth from the first frame, so it does not demonstrate the original bone being picked up. The third clip is <a href="https://d2drjpuinn46lb.cloudfront.net/Lucy_Edit__High_Fidelity_Text_Guided_Video_Editing.pdf" target="_blank" rel="noopener noreferrer">LucyEdit</a>: the original bone stays on the ground and the dog's coat changes. The fourth clip, newly supplied by the presenter, is <a href="https://ai.google.dev/gemini-api/docs/models/gemini-omni-flash" target="_blank" rel="noopener noreferrer">Gemini Omni Flash</a>: the dog approaches the original bone near the end, but the clip does not show it lifting and carrying the bone. Attribution follows the presenter's correction. These four clips belong to one example. Use Play all to start the comparison together and Pause all to pause it. <a href="presentation.html#dog-object-case">View the comparison</a>.</p>
 <p>These examples diagnose event timing, object continuity, interaction execution, and preservation. They do not establish overall model rankings or failure rates.</p></section>
 <section id="benchmark"><h2>Evaluation snapshot — 2026-09-26</h2><p>The three results shown in the deck use different metrics and protocols. Each is a diagnostic of the judge, and their scales must not be combined.</p>
 <table><thead><tr><th>Subset</th><th>Judge and protocol</th><th>No-edit score</th><th>Interpretation</th></tr></thead><tbody>

@@ -3,7 +3,7 @@
 <body><header class="site-header"><div class="nav-wrap"><a class="brand" href="../../">CausalVideoGroup</a><nav class="site-nav" aria-label="Discussion navigation"><a href="./">Overview</a><a href="presentation.html">Presentation</a><a href="summary.html">Summary</a><a href="references.html">References</a><a href="../">Discussions</a></nav></div></header>
 <main><header class="page-header"><p class="eyebrow">2026-09-28 · Peiyuan Zhu</p><h1>Video Editing Frontiers</h1></header><article class="prose">
 <p class="lede">Editing mechanisms, controllable changes, and persistent video editing.</p>
-<p>Led by Peiyuan Zhu. A bilingual, 46-slide survey with dedicated method explanations, paper figures, playable examples, and technical hover notes.</p>
+<p>Led by Peiyuan Zhu. A bilingual, 42-slide survey with dedicated method explanations, paper figures, playable examples, and technical hover notes.</p>
 <div class="actions"><a class="button" href="presentation.html">Open presentation</a><a class="button secondary" href="summary.html">Read topic summary</a><a class="button secondary" href="references.html">References</a><a class="button secondary" href="evidence.html">Examples &amp; evaluation notes</a></div>
 <h2>Discussion focus</h2><p>How can a video editor change an event while keeping the intended source objects and unrelated content? The discussion connects sampling and learning mechanisms to object interaction, edit history, world models, agents, and reliable evaluation.</p>
 <p><a href="presentation.html#open-directions">Open problems</a> · <a href="presentation.html#dog-object-case">Dog–object interaction</a> · <a href="presentation.html#open-world-models">World models</a> · <a href="presentation.html#open-editing-agents">Editing agents</a></p>

@@ -64,7 +64,18 @@
   document.querySelectorAll('[data-notes]').forEach(el=>el.onclick=()=>toggle('notes'));
   document.querySelectorAll('[data-image]').forEach(el=>el.onclick=()=>{document.querySelector('#zoom img').src=el.dataset.image;toggle('zoom')});
   let index = Math.max(0, meta.findIndex(s => s.id === location.hash.slice(1)));
+  let mediaRequest = 0;
   const videos = () => [...slides[index].querySelectorAll('video')];
+  function updatePlaybackControls() {
+    const vs=videos(),playing=vs.some(v=>!v.paused&&!v.ended);
+    const buttons=[document.getElementById('mediaButton'),...slides[index].querySelectorAll('[data-media-toggle]')];
+    buttons.forEach(button=>{
+      button.textContent=(playing?'Pause':'Play')+(vs.length>1||button.hasAttribute('data-media-toggle')?' all':'');
+      button.setAttribute('aria-pressed',String(playing));
+      button.title=playing?'Pause all videos (P)':'Play all videos from the start (P)';
+    });
+    document.getElementById('mediaButton').hidden=vs.length===0;
+  }
   function resize() {
     document.documentElement.style.setProperty('--scale', Math.min(innerWidth / 1280, innerHeight / 720));
     placeTerm();
@@ -75,27 +86,44 @@
   }
   function go(n) {
     closeTerm();
+    mediaRequest++;
     index = Math.max(0, Math.min(slides.length - 1, n));
     slides.forEach((s, i) => { s.classList.toggle('active', i === index); s.inert = i !== index; if (i !== index) s.querySelectorAll('video').forEach(v => v.pause()); });
     history.replaceState(null, '', `#${meta[index].id}`);
     document.getElementById('counter').textContent = `${index + 1} / ${slides.length}`;
-    document.getElementById('mediaButton').hidden = videos().length === 0;
-    document.getElementById('mediaButton').textContent = 'Play';
+    videos().forEach(v=>{v.preload='auto';});
+    updatePlaybackControls();
     document.getElementById('prev').disabled = index === 0;
     document.getElementById('next').disabled = index === slides.length - 1;
     updateNotes();
   }
   function toggle(id) { closeTerm(); const d = document.getElementById(id); if (d.open) d.close(); else { dialogs.forEach(x => x.close()); d.showModal(); } }
   async function play() {
-    const vs = videos();
-    if (vs.some(v => !v.paused)) { vs.forEach(v => v.pause()); document.getElementById('mediaButton').textContent = 'Play'; }
-    else { vs.forEach(v => { v.currentTime = 0; }); await Promise.all(vs.map(v => v.play().catch(() => {}))); document.getElementById('mediaButton').textContent = 'Pause'; }
+    const vs=videos(),request=++mediaRequest;
+    if(!vs.length)return;
+    if(vs.some(v=>!v.paused))vs.forEach(v=>v.pause());
+    else {
+      vs.forEach(v=>{v.currentTime=0;});
+      const results=await Promise.allSettled(vs.map(v=>v.play()));
+      if(request!==mediaRequest)return;
+      if(results.some(result=>result.status==='rejected')){
+        vs.forEach(v=>v.pause());
+        updatePlaybackControls();
+        [document.getElementById('mediaButton'),...slides[index].querySelectorAll('[data-media-toggle]')].forEach(button=>{button.title='A video could not play. Check the video controls, then retry.';});
+        return;
+      }
+    }
+    updatePlaybackControls();
   }
   document.getElementById('prev').onclick = () => go(index - 1);
   document.getElementById('next').onclick = () => go(index + 1);
   document.getElementById('tocButton').onclick = () => toggle('toc');
   document.getElementById('notesButton').onclick = () => toggle('notes');
   document.getElementById('mediaButton').onclick = play;
+  document.querySelectorAll('[data-media-toggle]').forEach(button=>button.onclick=play);
+  document.querySelectorAll('video').forEach(video=>['play','pause','ended'].forEach(event=>video.addEventListener(event,()=>{
+    if(video.closest('.slide')===slides[index])updatePlaybackControls();
+  })));
   document.getElementById('fullButton').onclick = () => document.fullscreenElement ? document.exitFullscreen() : document.documentElement.requestFullscreen();
   document.querySelectorAll('[data-go]').forEach(b => b.onclick = () => { go(Number(b.dataset.go)); document.getElementById('toc').close(); });
   dialogs.forEach(d => { d.querySelector('.close').onclick = () => d.close(); d.addEventListener('click', e => { if (e.target === d) { const r = d.getBoundingClientRect(); if (e.clientX < r.left || e.clientX > r.right || e.clientY < r.top || e.clientY > r.bottom) d.close(); } }); });

@@ -5,6 +5,6 @@
 <h2>Paper figures</h2><ul><li><a href="assets/dynaedit.png">dynaedit.png</a> — <a href="https://arxiv.org/pdf/2603.17989" target="_blank" rel="noopener noreferrer">DynaEdit</a></li><li><a href="assets/bernini.png">bernini.png</a> — <a href="https://arxiv.org/pdf/2605.22344" target="_blank" rel="noopener noreferrer">Bernini</a> (v1 figure)</li><li><a href="assets/propfly.png">propfly.png</a> — <a href="https://arxiv.org/pdf/2602.20583" target="_blank" rel="noopener noreferrer">PropFly</a></li><li><a href="assets/motionv2v.png">motionv2v.png</a> — <a href="https://arxiv.org/pdf/2511.20640" target="_blank" rel="noopener noreferrer">MotionV2V</a></li><li><a href="assets/void.png">void.png</a> — <a href="https://arxiv.org/pdf/2604.02296" target="_blank" rel="noopener noreferrer">VOID</a></li><li><a href="assets/joyai-video-edit.png">joyai-video-edit.png</a> — <a href="https://arxiv.org/pdf/2608.03974" target="_blank" rel="noopener noreferrer">JoyAI-Video-Edit</a></li><li><a href="assets/goku.png">goku.png</a> — <a href="https://arxiv.org/pdf/2606.30599" target="_blank" rel="noopener noreferrer">Goku</a></li><li><a href="assets/coinve-framework.png">coinve-framework.png</a> — <a href="https://arxiv.org/pdf/2608.17566" target="_blank" rel="noopener noreferrer">CoinVE</a></li><li><a href="assets/world-in-world.png">world-in-world.png</a> — <a href="https://arxiv.org/pdf/2609.11548" target="_blank" rel="noopener noreferrer">World in World</a></li><li><a href="assets/mse-framework.png">mse-framework.png</a> — <a href="https://arxiv.org/pdf/2608.17559" target="_blank" rel="noopener noreferrer">MSEditor</a></li></ul>
 <p>Figures retain their original proportions. The presentation can enlarge the complete paper figures; the Bernini and DynaEdit slide views use crops of those originals.</p>
 <h2>Formula components</h2><p>Eleven SVG components typeset the FlowEdit ODE, DynaEdit guidance, and Bernini training objectives with MathJax 3.2.2. The public presentation uses the pre-rendered assets and does not load a MathJax CDN at runtime.</p>
-<h2>Example videos</h2><p>Nine videos accompany the action-substitution, clothing-edit, and dog–bone comparisons. They are the presentation examples supplied by Peiyuan Zhu. The dog clips are 832 × 480, 25 fps, 81 frames; their static poster images show 2.40 s. Original video files are retained for playback.</p>
+<h2>Example videos</h2><p>Ten videos accompany the action-substitution, clothing-edit, and dog–bone comparisons. They are the presentation examples supplied by Peiyuan Zhu. In the four-video dog comparison, the source, archived target, and <a href="https://d2drjpuinn46lb.cloudfront.net/Lucy_Edit__High_Fidelity_Text_Guided_Video_Editing.pdf" target="_blank" rel="noopener noreferrer">LucyEdit</a> clips are 832 × 480, 25 fps, 81 frames, with posters at 2.40 s. The <a href="https://ai.google.dev/gemini-api/docs/models/gemini-omni-flash" target="_blank" rel="noopener noreferrer">Gemini Omni Flash</a> clip is 640 × 368, 24 fps, 80 frames, with its poster at 3.25 s. Original video files are retained for playback. Use Play all to start all four clips together; shorter clips hold their final frame.</p>
 <p><a href="evidence.html#examples">Example observations and limits</a> · <a href="presentation.html">Open presentation</a></p></article></main>
 <footer class="site-footer"><div class="site-footer-inner"><span>© CausalVideoGroup</span><a href="../">All discussions</a></div></footer></body></html>