拍樂雲 CEO 趙加雨：音視頻領域的發展與挑戰 | QCon

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"疫情過後，人們對在線課堂、視頻會議等音視頻服務的需求顯著上升，對音視頻的使用體驗、產品質量也有了更高的要求。拍樂雲作爲實時通信雲服務提供商，需要用專業的技術滿足用戶在不同場景、設備上的需求，採用不同的策略應對多樣化的網絡環境，才能不斷提升用戶在觀看音視頻過程中的體驗。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"InfoQ 記者有幸在"},{"type":"link","attrs":{"href":"https:\/\/qcon.infoq.cn\/2021\/beijing\/","title":null,"type":null},"content":[{"type":"text","text":"QCon 2021 全球軟件開發大會"}],"marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}]},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"上，採訪到了拍樂雲的創始人&CEO趙加雨，由他親自爲我們講解音視頻領域的發展與挑戰。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"以下是視頻採訪內容，爲方便讀者查看，視頻下方也附上了文字內容。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"embedcomp","attrs":{"type":"video","data":{"id":"385276","name":"拍樂雲 CEO 趙加雨：音視頻領域的發展與挑戰 .mp4","poster":"https:\/\/static001.infoq.cn\/resource\/image\/fa\/cf\/fa39646d6cdc31e637069bf74e0fd8cf.png","url":"https:\/\/media001.geekbang.org\/0ec2886027c44b1aa5685d9c006307ad\/d9042efc66d0445aa2b760af9bf9d3af-579b02c47b66a55a30ac5b497003950f-sd.m3u8"}}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"非常感謝您參加 QCon 2021 的視頻採訪。首先，請您做一個簡單的自我介紹。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"趙加雨："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"我曾經在 WebEx 工作了14年，17年從美國回來後就加入了網易，擔任網易雲信的 CTO。2年後出來創業，創立了"},{"type":"link","attrs":{"href":"https:\/\/xie.infoq.cn\/article\/f369fcffcfca5fdef2088539e","title":"xxx","type":null},"content":[{"type":"text","text":"拍樂雲"}]},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"您一直在做專業視頻音頻方面，所以我對您的工作經驗非常感興趣。之前在網上看到您有聊到音視頻領域的昨天、今天和明天，那在這裏能和我們再進一步談談音視頻領域裏技術的演進過程嗎？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"趙加雨："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"音視頻領域經過這十幾年的發展也是在快速演進的，這裏可以從多個緯度來討論。在我2003年最開始做視頻會議的時候，那時的音視頻還是以語音爲主，視頻還是低清的，但這之後視頻就逐漸變得越來越高清。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"從設備角度來說，剛開始很多都是在 PC 上使用，隨着移動互聯網的發展，更多的移動設備也接入進來。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"另外，起初視頻會議的使用場景一般就是在電腦上開會，但現在已經發展到了有各種場景都會使用音視頻，包括娛樂、在線教育等等。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"從技術的角度來說，近幾年也有 "},{"type":"link","attrs":{"href":"https:\/\/xie.infoq.cn\/article\/816d873ed8c0ad24c249b14d9","title":"xxx","type":null},"content":[{"type":"text","text":"AI 與音視頻"}]},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"的結合，這些都是非常好的發展。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"音視頻技術以及處理流程都比較複雜，處理過程中包含編碼、數據存儲、弱網處理等等，所以能請您簡單介紹一個實時音視頻系統的處理流程是怎樣的嗎？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"趙加雨："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"比如說我們兩個人是線上通話的，我們都需要聽到對方的聲音，看到對方的畫面，這裏就涉及到一個鏈路。一般情況下，系統會採集音頻和視頻，接下來會有一個前處理，前處理可能會完成美聲、美顏和其他的一些視頻優化。接下來，視頻要經過編碼再進入到網絡傳輸，在互聯網傳輸的過程中又需要實現網絡加速。服務端分發後，需要有一個接收端，接收端收到數據包後會應用一些抗弱網的技術來保障質量，最後完成視頻解碼並在用戶面前播放。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"音視頻處理流程有這麼多環節，各個環節的關鍵技術是怎樣的？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"趙加雨："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"因爲音視頻都非常強調兩高一低，尤其是在低時延的情況下，因此我們要求所有通話都最好能控制在400毫秒以內。同時，低時延的限制導致技術人員在各個環節都要把很多性能方面的問題調整到最優點。就像之前提到的，從採集到前處理，從編碼到服務端，這裏的每個鏈條都要考慮兩高一低，最終實現一個真正的高清流暢的低時延視頻通話效果。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在各個環節中有哪些難點，它們分別是什麼？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"趙加雨："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"這裏一般會面臨兩大難點。第一個是網絡環境是多種多樣的，比如不同的網絡環境裏可能有很多弱網，所以在這些場景裏就要考慮到抗弱網。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"第二個是視頻運行的設備也是多種多樣的，比如各種手機設備、低端或高端設備等等，但我們還是要在各個設備上都做到表現最優。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"所以，多樣化的網絡環境和多元化的設備，給音視頻的處理帶來了各種挑戰。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"我瞭解到擁塞控制是保證音視頻質量的一個關鍵環節，能談談你們在這方面的理解和解決方案嗎？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"趙加雨："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"其實可以從兩個維度來探討擁塞控制。第一是要知道有哪些擁塞控制的方式，這裏可能涉及到幾十種策略算法，而且在各個環節都會有不同策略的算法來幫助技術人員處理對應的情況，所以首先是要知道這個工具箱裏有這麼多工具。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"第二點是在面對真實的弱網環境時，不同的弱網環境適用的擁塞控制策略不同，所以應該具體情況具體分析，採取合適的抗弱網算法，而使用這些算法最終都是爲了保證兩高一低的音視頻體驗。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"所以，不能爲了抗弱網而抗弱網，丟棄了兩高一低，一定要避免在錯誤的地方使用錯誤的工具，從而帶來負向效果。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"從勢能上看，整個音視頻這個行業進入了快速增長期。那麼從技術上來看，目前行業面臨的挑戰在哪裏？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"趙加雨："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在我看來，因爲這次疫情，音視頻變得更下沉，又變得更上浮。這裏可以這樣理解，下沉的意思是說有更多的場景和人來使用音視頻，對應的設備種類也增加了。因此低端設備就變多了，差的網絡變多了，這就是一種挑戰。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"而上浮則代表着有更專業、高端的場景也開始需要音視頻，比如怎樣讓音視頻更高清、更有沉浸感和互動感，這些方面也出現了更多的需求。所以，比起過去，音視頻的應用場景是在不斷拓寬的，需求也變得更多樣了。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"您覺得音視頻行業未來的趨勢會是怎樣的？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"趙加雨："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"未來，音視頻一定會隨着新的技術以及大環境，包括 IT、基建等等來同步迭代。在音視頻領域裏，有很多東西是幾十年都沒有變過，也有很多東西在持續迭代。隨着 AR、VR、AI 這些新技術的出現，以及 5G 的商用，音視頻也一定會和它們有更多的結合，在更多場景裏出現。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"最後想問問您的創業，音視頻的創業賽道非常激烈，您在幾年創業過程裏有什麼感觸？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"趙加雨："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"我原來做了將近20年的技術，這兩年創業的過程也看到了這個賽道里激烈的競爭。從自己的工作體會來說，現在會看到各種客戶不同的場景、不同的問題，這個很有意思。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"以前只是專注把技術做深，離市場很遠。現在帶着一羣專業的技術人員去解決客戶面臨的問題，一方面是積累專業的技術，一方面也縮短客戶反饋的鏈路。在解決問題過程裏，會有非常多的成就感。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"雖然這個賽道競爭很激烈，但我們也堅信要把音視頻產品的質量做到最優，這樣才能在市場獲得發展的機會。所以我們希望在技術上做到極致，在產品上追求極簡，把產品做好，給客戶帶來最好的產品，最終在市場上佔領一席之地。"}]}]}