Update README

README.md

@@ -96,8 +96,8 @@ model-index:
 
 ### Vietnamese Zero-Shot Text-to-Speech (TTS) with real-time streaming and voice cloning from seconds of audio. Fast, natural, and optimised for CPU inference.
 
-**The most accurate open Vietnamese TTS we know of — 13× fewer word errors than
-the next open model**, and it runs faster than real time on a laptop CPU.
+**The most accurate open Vietnamese TTS we know of — 4× fewer word errors than
+the next best model**, and it runs faster than real time on a laptop CPU.
 
 * 🎯 **Ultra-natural** — 2.91 UTMOS, ~0.5 MOS above every other open Vietnamese
   system, with near-zero dead air (0.029 s vs 0.23–0.53 s).
@@ -107,7 +107,7 @@ the next open model**, and it runs faster than real time on a laptop CPU.
 * ⚡ **Real-time on CPU, streaming** — ~2× faster than real time (RTF 0.5×),
   first audio chunk in ~70 ms. No GPU required.
 * 🇻🇳 **Built for Vietnamese** — tones, code-switched English, and a built-in
-  normalizer that reads `31/12/2025` and `1.250 tỷ` the way a person would.
+  normalizer that reads `31/12/2025` and `ZeroTTS` the way a person would.
 
 * Code, examples, browser demo: **https://github.com/zeroweight-ai/ZeroTTS**
 * Benchmark dataset: **https://huggingface.co/datasets/zeroweight-ai/ZeroBench-TTS**
@@ -127,8 +127,33 @@ tts.save_audio(audio, "out.wav")
 Streaming, with first audio in roughly 70 ms:
 
 ```python
-for chunk in tts.synthesize_stream("Một đoạn văn bản dài hơn…", voice="maichi"):
-    play(chunk)   # (1, n) float32 at 48 kHz
+import queue
+
+import numpy as np
+import sounddevice as sd   # pip install sounddevice
+
+TEXT = ("Đây là chế độ phát trực tuyến. Âm thanh được tạo ra và phát ngay lập tức, "
+        "không cần chờ toàn bộ đoạn văn hoàn thành. Nhờ vậy, người nghe chỉ mất "
+        "khoảng 70 mili giây là đã nghe thấy câu đầu tiên, ngay cả khi mô "
+        "hình đang chạy trên CPU của một chiếc laptop bình thường.")
+
+pending, tail = queue.Queue(), np.zeros(0, dtype="float32")
+
+def feed(outdata, frames, _time, _status):
+    global tail
+    while len(tail) < frames and not pending.empty():
+        tail = np.concatenate([tail, pending.get_nowait()])
+    n = min(frames, len(tail))
+    outdata[:n, 0] = tail[:n]
+    outdata[n:] = 0
+    tail = tail[n:]
+
+with sd.OutputStream(samplerate=tts.sample_rate, channels=1,
+                     dtype="float32", callback=feed):
+    for chunk in tts.synthesize_stream(TEXT, voice="maichi"):
+        pending.put(chunk.reshape(-1))   # chunk is (1, n) float32 at 48 kHz
+    while not pending.empty() or len(tail):
+        sd.sleep(50)                     # let the buffer drain before closing
 ```
 
 ## Benchmarks
@@ -146,11 +171,13 @@ spoken out, from the benchmark's own curated reading.
 | **Dead air** (excess silence) ↓ | **0.029 s** | 0.386 s | 0.568 s | 0.215 s |
 | **RTF, CPU** ↓ | **0.50×** | 6.12× | 0.71× | 0.73× |
 | **Time to first audio, CPU** ↓ | **~70 ms** | ~34 s | ~6.1 s | ~5.1 s |
-| Size | **81 M**, CPU | 3.1 GB, GPU | 1.9 GB, GPU | 1.9 GB, GPU |
+| Size | **202 M** params, 0.86 GB fp32, CPU | 3.1 GB, GPU | 1.9 GB, GPU | 1.9 GB, GPU |
 
-**4× fewer word errors than the next-best system**, and the only one of the four
-that's actually fast on CPU — the others are sized and tuned for a GPU, and it
-shows.
+**4× fewer word errors than the next-best system**, and the fastest of the four
+on CPU. The gap is much wider in latency than in throughput: the two XTTS
+fine-tunes also beat real time (0.71×) but need seconds to emit their first
+sample, while OmniVoice is 6× *slower* than real time. All three are sized and
+tuned for a GPU, and it shows.
 
 Full comparison tables, per-subset breakdowns, and CPU speed methodology:
 **[docs/BENCHMARKS.md](https://github.com/zeroweight-ai/ZeroTTS/blob/main/docs/BENCHMARKS.md)**