atlas-iac/dockerfiles/hermes-webui-atlas-voice.css
jenkins 09ba5f8ac0 hermes(voice): fix first-word clipping, always-stitch, panel glow, conversation mode
Round 2 from live testing:
- Clipping root cause: the endpointer committed on any 1.1s pause, so a
  thinking pause after a sentence opener sent one word; a speculative
  Whisper pass over that fragment then stalled the real commit ~3.5s on
  the Jetson. Young utterances now hold a 1.8s endpoint until 1.2s of
  speech accrues, speculative decode waits for 700ms of speech, resume
  is unconditional after any 650ms gap (server-frozen snapshots can
  never reach commit), and the noise floor is capped so playback echo
  cannot deafen onset. Deterministic capture-continuity probe added.
- Stitching now fires for any barge-cancelled send within 20s,
  regardless of partial assistant output.
- The workspace-drawer dark rectangle was our own HUX chrome resolving
  undefined theme tokens (--bg-primary) to a flat box; bootstrap.css
  bridges the real theme tokens and drops a compositor-hazard
  backdrop-filter.
- Conversation mode: full-viewport hands-free overlay with an energy-
  driven orb (mic RMS in, TTS activity out), state caption, live
  transcript/reply captions, mute and exit controls, focus trap,
  Escape, reduced-motion support. Voice lane 253 tests green.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BvMSXH8VH2tMWXanb8SJdf
2026-08-24 11:21:35 -03:00

571 lines
15 KiB
CSS

/* Private hands-free conversation instrument for the Atlas voice bridge. */
.voice-mode-bar {
--voice-accent: 72, 207, 204;
--voice-accent-secondary: 76, 164, 205;
position: relative;
isolation: isolate;
box-sizing: border-box;
min-height: 106px;
padding: 9px 12px 10px;
display: flex;
flex-direction: column;
align-items: center;
justify-content: center;
gap: 6px;
overflow: hidden;
border-bottom: 1px solid rgba(174, 192, 218, 0.1);
background:
radial-gradient(circle at 50% 38%, rgba(var(--voice-accent), 0.075), transparent 46%),
linear-gradient(180deg, rgba(255, 255, 255, 0.025), rgba(4, 8, 18, 0.035));
}
.voice-mode-bar::before {
content: "";
position: absolute;
z-index: -1;
top: 0;
left: 18%;
right: 18%;
height: 1px;
background: linear-gradient(90deg, transparent, rgba(233, 244, 255, 0.16), transparent);
}
:root:not(.dark) .voice-mode-bar {
border-bottom-color: rgba(38, 55, 78, 0.1);
background:
radial-gradient(circle at 50% 38%, rgba(var(--voice-accent), 0.09), transparent 48%),
linear-gradient(180deg, rgba(255, 255, 255, 0.72), rgba(226, 232, 240, 0.12));
}
.voice-mode-indicator {
--voice-ripple-scale: 1.035;
--voice-ripple-opacity: 0.3;
position: relative;
width: 70px;
height: 70px;
flex: 0 0 70px;
border-radius: 50%;
transform: translateZ(0);
}
.voice-mode-indicator.listening {
--voice-accent: 69, 218, 207;
--voice-accent-secondary: 73, 169, 209;
}
.voice-mode-indicator.transcribing {
--voice-accent: 73, 193, 218;
--voice-accent-secondary: 93, 129, 224;
}
.voice-mode-indicator.thinking {
--voice-accent: 135, 111, 244;
--voice-accent-secondary: 81, 62, 190;
}
.voice-mode-indicator.speaking {
--voice-accent: 249, 169, 105;
--voice-accent-secondary: 236, 112, 137;
}
.voice-mode-indicator.error {
--voice-accent: 235, 135, 88;
--voice-accent-secondary: 198, 74, 78;
}
.voice-instrument-halo,
.voice-instrument-ripple,
.voice-instrument-orbit,
.voice-instrument-core {
position: absolute;
display: block;
box-sizing: border-box;
border-radius: 50%;
pointer-events: none;
}
.voice-instrument-halo {
inset: 2px;
opacity: 0;
background: radial-gradient(circle, rgba(var(--voice-accent), 0.2), rgba(var(--voice-accent), 0.05) 43%, transparent 70%);
filter: blur(4px);
}
.voice-instrument-ripple {
inset: 3px;
opacity: 0;
border: 1px solid rgba(var(--voice-accent), 0.56);
box-shadow: 0 0 12px rgba(var(--voice-accent), 0.13);
}
.voice-instrument-orbit {
inset: 2px;
opacity: 0;
border: 1px solid rgba(var(--voice-accent), 0.18);
box-shadow:
inset 0 0 8px rgba(var(--voice-accent), 0.06),
0 0 11px rgba(var(--voice-accent), 0.07);
}
.voice-instrument-orbit::before {
content: "";
position: absolute;
inset: -1px;
border-radius: inherit;
border-top: 1.5px solid rgba(var(--voice-accent), 0.88);
border-right: 1px solid rgba(var(--voice-accent-secondary), 0.35);
border-bottom: 1px solid transparent;
border-left: 1px solid transparent;
}
.voice-instrument-orbit::after {
content: "";
position: absolute;
top: -2px;
left: 50%;
width: 5px;
height: 5px;
margin-left: -2.5px;
border-radius: 50%;
background: rgb(var(--voice-accent));
box-shadow:
0 0 5px rgba(var(--voice-accent), 0.92),
0 0 12px rgba(var(--voice-accent), 0.44);
}
.voice-instrument-core {
inset: 12px;
display: grid;
place-items: center;
color: rgba(242, 250, 255, 0.94);
border: 1px solid rgba(238, 247, 255, 0.2);
background:
radial-gradient(circle at 38% 29%, rgba(255, 255, 255, 0.26), transparent 24%),
radial-gradient(circle at 50% 66%, rgba(var(--voice-accent), 0.36), transparent 64%),
linear-gradient(145deg, rgba(36, 45, 61, 0.96), rgba(10, 15, 27, 0.98));
box-shadow:
inset 0 1px 1px rgba(255, 255, 255, 0.22),
inset 0 -9px 18px rgba(0, 0, 0, 0.24),
0 4px 13px rgba(0, 0, 0, 0.25),
0 0 17px rgba(var(--voice-accent), 0.17);
}
:root:not(.dark) .voice-instrument-core {
color: rgba(31, 48, 66, 0.9);
border-color: rgba(255, 255, 255, 0.82);
background:
radial-gradient(circle at 38% 29%, rgba(255, 255, 255, 0.96), transparent 27%),
radial-gradient(circle at 50% 66%, rgba(var(--voice-accent), 0.2), transparent 65%),
linear-gradient(145deg, rgba(248, 250, 251, 0.98), rgba(216, 225, 231, 0.96));
box-shadow:
inset 0 1px 1px rgba(255, 255, 255, 0.94),
inset 0 -8px 17px rgba(64, 84, 100, 0.1),
0 4px 12px rgba(35, 53, 72, 0.15),
0 0 15px rgba(var(--voice-accent), 0.13);
}
.voice-instrument-symbol {
display: block;
width: 22px;
height: 22px;
filter: drop-shadow(0 1px 2px rgba(0, 0, 0, 0.32));
}
.voice-instrument-symbol svg {
display: block;
width: 100%;
height: 100%;
}
.voice-symbol {
display: none;
}
.voice-mode-indicator.listening .voice-symbol-listening,
.voice-mode-indicator.transcribing .voice-symbol-transcribing,
.voice-mode-indicator.thinking .voice-symbol-thinking,
.voice-mode-indicator.speaking .voice-symbol-speaking,
.voice-mode-indicator.error .voice-symbol-error {
display: inline;
}
.voice-mode-indicator.listening .voice-instrument-halo {
opacity: 0.62;
animation: voice-instrument-breathe 3.8s ease-in-out infinite;
}
.voice-mode-indicator.listening .voice-instrument-ripple {
opacity: var(--voice-ripple-opacity);
transform: scale(var(--voice-ripple-scale));
transition: transform 140ms ease-out, opacity 160ms ease-out;
}
.voice-mode-indicator.transcribing .voice-instrument-orbit {
opacity: 0.92;
animation: voice-instrument-orbit 3.4s cubic-bezier(0.58, 0.12, 0.42, 0.88) infinite;
}
.voice-mode-indicator.thinking .voice-instrument-orbit {
opacity: 0.96;
animation: voice-instrument-orbit 4.6s cubic-bezier(0.58, 0.12, 0.42, 0.88) infinite;
}
.voice-mode-indicator.thinking .voice-instrument-core {
animation: voice-instrument-core-breathe 5.7s ease-in-out infinite;
}
.voice-mode-indicator.speaking .voice-instrument-orbit {
opacity: 0.86;
transform: rotate(38deg);
}
.voice-mode-indicator.speaking.is-playing .voice-instrument-halo {
opacity: 0.66;
animation: voice-instrument-speaking-pulse 2.15s ease-out infinite;
}
.voice-mode-indicator.error .voice-instrument-orbit {
opacity: 0.9;
transform: rotate(-32deg);
border-color: rgba(var(--voice-accent), 0.44);
}
.voice-mode-indicator.error .voice-instrument-core {
color: rgba(255, 232, 219, 0.95);
box-shadow:
inset 0 1px 1px rgba(255, 255, 255, 0.2),
inset 0 -9px 18px rgba(0, 0, 0, 0.24),
0 4px 13px rgba(0, 0, 0, 0.25),
0 0 12px rgba(var(--voice-accent), 0.17);
}
.voice-mode-label {
position: relative;
z-index: 1;
min-height: 16px;
max-width: min(100%, 34rem);
overflow: hidden;
color: var(--text, #f4f6fb);
font-size: 12px;
font-weight: 650;
line-height: 1.35;
letter-spacing: 0.018em;
text-align: center;
text-overflow: ellipsis;
white-space: nowrap;
text-shadow: 0 1px 3px rgba(0, 0, 0, 0.28);
}
:root:not(.dark) .voice-mode-label {
color: #263648;
text-shadow: 0 1px 1px rgba(255, 255, 255, 0.8);
}
@keyframes voice-instrument-breathe {
0%, 100% { opacity: 0.38; transform: scale(0.94); }
50% { opacity: 0.72; transform: scale(1.06); }
}
@keyframes voice-instrument-orbit {
0% { transform: rotate(0deg); }
50% { transform: rotate(177deg); }
100% { transform: rotate(360deg); }
}
@keyframes voice-instrument-core-breathe {
0%, 100% { filter: saturate(0.96) brightness(0.98); }
50% { filter: saturate(1.08) brightness(1.06); }
}
@keyframes voice-instrument-speaking-pulse {
0% { opacity: 0.58; transform: scale(0.88); }
58% { opacity: 0.22; transform: scale(1.12); }
100% { opacity: 0; transform: scale(1.2); }
}
@media (max-width: 640px) {
.voice-mode-bar {
min-height: 94px;
padding: 7px 10px 8px;
gap: 4px;
}
.voice-mode-indicator {
width: 62px;
height: 62px;
flex-basis: 62px;
}
.voice-instrument-core {
inset: 11px;
}
.voice-instrument-symbol {
width: 20px;
height: 20px;
}
}
@media (prefers-reduced-motion: reduce) {
.voice-mode-bar *,
.voice-mode-bar *::before,
.voice-mode-bar *::after {
animation: none !important;
transition: none !important;
}
.voice-mode-indicator.listening .voice-instrument-halo {
opacity: 0.45;
transform: none;
}
.voice-mode-indicator.listening .voice-instrument-ripple {
opacity: 0.34;
transform: scale(1.035);
}
.voice-mode-indicator.transcribing .voice-instrument-orbit,
.voice-mode-indicator.thinking .voice-instrument-orbit {
transform: rotate(24deg);
}
.voice-mode-indicator.speaking.is-playing .voice-instrument-halo {
opacity: 0.42;
transform: scale(1.03);
}
}
/* ── Conversation mode overlay ─────────────────────────────────────────── */
/* Full-viewport hands-free conversation surface. Created lazily by
atlas-voice.js on activation and removed on deactivate. The orb breathes in
idle listening, follows live microphone energy through
--conversation-level, and pulses while Hermes speaks (.is-playing, driven
from TTS playback activity). State accents reuse the indicator palette. */
.voice-conversation {
--voice-accent: 69, 218, 207;
--voice-accent-secondary: 73, 169, 209;
--conversation-level: 0;
position: fixed;
inset: 0;
z-index: 420;
display: flex;
flex-direction: column;
align-items: center;
justify-content: center;
gap: 24px;
padding: 34px 22px calc(28px + env(safe-area-inset-bottom, 0px));
background:
radial-gradient(circle at 50% 30%, rgba(var(--voice-accent), 0.11), transparent 52%),
radial-gradient(circle at 22% 92%, rgba(157, 142, 224, 0.07), transparent 44%),
linear-gradient(160deg, #0b1120, #070a12 55%, #081018);
color: rgba(240, 247, 255, 0.94);
}
.voice-conversation[data-voice-state="transcribing"] {
--voice-accent: 73, 193, 218;
--voice-accent-secondary: 93, 129, 224;
}
.voice-conversation[data-voice-state="thinking"] {
--voice-accent: 135, 111, 244;
--voice-accent-secondary: 81, 62, 190;
}
.voice-conversation[data-voice-state="speaking"] {
--voice-accent: 249, 169, 105;
--voice-accent-secondary: 236, 112, 137;
}
.voice-conversation[data-voice-state="error"] {
--voice-accent: 235, 135, 88;
--voice-accent-secondary: 198, 74, 78;
}
.voice-conversation-orb {
position: relative;
width: min(46vmin, 220px);
height: min(46vmin, 220px);
flex: none;
}
.voice-conversation-orb-halo,
.voice-conversation-orb-core,
.voice-conversation-orb-ring {
position: absolute;
inset: 0;
display: block;
border-radius: 50%;
pointer-events: none;
}
.voice-conversation-orb-halo {
background: radial-gradient(circle, rgba(var(--voice-accent), 0.34), rgba(var(--voice-accent-secondary), 0.1) 52%, transparent 74%);
filter: blur(10px);
opacity: 0.75;
transform: scale(calc(1 + (var(--conversation-level) * 0.3)));
transition: transform 130ms ease-out;
}
.voice-conversation-orb-core {
inset: 12%;
background:
radial-gradient(circle at 36% 28%, rgba(255, 255, 255, 0.34), transparent 34%),
radial-gradient(circle at 52% 68%, rgba(var(--voice-accent), 0.62), rgba(var(--voice-accent-secondary), 0.4) 62%, transparent 86%),
linear-gradient(150deg, rgba(34, 44, 63, 0.95), rgba(9, 13, 24, 0.98));
box-shadow:
inset 0 1px 2px rgba(255, 255, 255, 0.28),
0 10px 42px rgba(0, 0, 0, 0.45),
0 0 44px rgba(var(--voice-accent), 0.3);
transform: scale(calc(1 + (var(--conversation-level) * 0.14)));
transition: transform 110ms ease-out;
}
.voice-conversation-orb-ring {
inset: 3%;
border: 1.5px solid rgba(var(--voice-accent), 0.4);
box-shadow: 0 0 22px rgba(var(--voice-accent), 0.16);
opacity: 0.85;
}
/* Idle listening: slow breathing. */
.voice-conversation[data-voice-state="listening"] .voice-conversation-orb-halo {
animation: voice-conversation-breathe 4.4s ease-in-out infinite;
}
/* Transcribing/thinking: the ring sweeps. */
.voice-conversation[data-voice-state="transcribing"] .voice-conversation-orb-ring,
.voice-conversation[data-voice-state="thinking"] .voice-conversation-orb-ring {
border-top-color: rgba(var(--voice-accent), 0.95);
border-right-color: rgba(var(--voice-accent-secondary), 0.4);
animation: voice-conversation-sweep 2.6s linear infinite;
}
/* Hermes speaking: the halo waves outward while playback is live. */
.voice-conversation.is-playing .voice-conversation-orb-halo {
animation: voice-conversation-wave 1.9s ease-out infinite;
}
.voice-conversation.is-playing .voice-conversation-orb-core {
animation: voice-conversation-swell 1.9s ease-in-out infinite;
}
.voice-conversation-state {
min-height: 20px;
font-size: 15px;
font-weight: 650;
letter-spacing: 0.02em;
text-align: center;
color: rgba(238, 246, 255, 0.9);
}
.voice-conversation-captions {
display: flex;
flex-direction: column;
gap: 10px;
width: min(100%, 40rem);
max-height: 30vh;
overflow: hidden;
text-align: center;
}
.voice-conversation-caption-user {
margin: 0;
font-size: 14px;
line-height: 1.5;
color: rgba(178, 196, 214, 0.88);
font-style: italic;
}
.voice-conversation-caption-assistant {
margin: 0;
font-size: 16px;
line-height: 1.6;
color: rgba(240, 247, 255, 0.95);
}
.voice-conversation-controls {
display: flex;
gap: 14px;
margin-top: 6px;
}
.voice-conversation-controls button {
font: inherit;
font-size: 13px;
font-weight: 600;
color: rgba(233, 243, 252, 0.92);
background: rgba(255, 255, 255, 0.07);
border: 1px solid rgba(var(--voice-accent), 0.36);
border-radius: 999px;
padding: 10px 18px;
cursor: pointer;
}
.voice-conversation-controls button:hover {
background: rgba(255, 255, 255, 0.12);
}
.voice-conversation-controls button:focus-visible {
outline: 2px solid rgb(var(--voice-accent));
outline-offset: 2px;
}
.voice-conversation.is-muted .voice-conversation-mute {
color: #ffd9c7;
border-color: rgba(235, 135, 88, 0.6);
background: rgba(235, 135, 88, 0.14);
}
.voice-conversation.is-muted .voice-conversation-orb-halo,
.voice-conversation.is-muted .voice-conversation-orb-core {
animation: none;
transform: none;
filter: blur(10px) saturate(0.4);
}
.voice-conversation.is-muted .voice-conversation-orb-core {
filter: saturate(0.35) brightness(0.8);
}
@keyframes voice-conversation-breathe {
0%, 100% { opacity: 0.55; transform: scale(calc(0.96 + (var(--conversation-level) * 0.3))); }
50% { opacity: 0.85; transform: scale(calc(1.05 + (var(--conversation-level) * 0.3))); }
}
@keyframes voice-conversation-sweep {
0% { transform: rotate(0deg); }
100% { transform: rotate(360deg); }
}
@keyframes voice-conversation-wave {
0% { opacity: 0.8; transform: scale(0.96); }
55% { opacity: 0.5; transform: scale(1.12); }
100% { opacity: 0.75; transform: scale(0.98); }
}
@keyframes voice-conversation-swell {
0%, 100% { transform: scale(0.99); }
50% { transform: scale(1.05); }
}
@media (max-width: 640px) {
.voice-conversation { gap: 18px; }
.voice-conversation-captions { max-height: 34vh; }
}
/* Reduced motion: static orb states — the accent color and the state caption
still communicate every phase without animation or level-driven motion. */
@media (prefers-reduced-motion: reduce) {
.voice-conversation *,
.voice-conversation *::before,
.voice-conversation *::after {
animation: none !important;
transition: none !important;
}
.voice-conversation-orb-halo,
.voice-conversation-orb-core {
transform: none !important;
}
}