In the past few years, the product of smart speaker went viral in families, especially for some mid-level group.  Known products such as 小米音箱 , Chromcast, Apple HomePod became a vital application in a occasion, even some hotel provide one in a room as a digital house keeper.

Based on the 資訊處理途徑 and the using scenario, I would  predict this product is redundant and will be obsolete without audacity.

There are two main reasons I stand the base on:

• The inefficient  information dealing process

• The inconvenient user scenario

Furthermore, a better replacement to this product is your mobile phone with a bluetooth speaker.

我們會使用到智慧音箱的情境有哪些？

Play music, Manage IOT devices, set alarm

First, the information processing is as below:

Input 智慧音箱的麥克風

—>  使用embedded / or cloud 聲音辨識

—> 語言處理

—> 形成指令

—> 音箱中的應用程式，音樂、鬧鐘、IOT 等等，response/執行指令

Output 喇叭

Several awful feedbacks:

• 語音收發的回饋感不明顯

• 等候執行的同時無法做其他事

• 網路連接內部 wifi 須確保穩定

若我們將該使用情境改成用「手機連接藍牙喇叭」各環節比較：

• 手機可以隨手拿起來講話，收音距離大勝，不需要走到智慧音箱旁邊。

• 語音輸入的回饋感更好，有螢幕可以確認，指令是否輸入正確、裝置處理進度

• 手機本身記憶體更強，網路也穩定

手機輸的地方 

喇叭音質 音量

Solution 穩定連接的音箱

比較 wifi 藍芽 各protocal 穩定度

因為喇叭不小心開到最大聲，使用者對使用者說「把音量調小」的指令被淹沒在喇叭的聲音裡，像這樣粗糙的將喇叭跟麥克風試做單一音源的產品，欠缺考慮了兩個元件本質上衝突的問題（聲音輸入ys輸出）可以想想，iphone在處理這個問題的細緻度。（將手機貼近耳朵，就會將聲音改由聽筒輸出）