昨日不忿氣，又得閒無事做，又見有材料，就重造一條 2.5 mm 轉 3.5 mm 線。於是…它出世了。

第一條實用的轉接線。為減省體積，連接頭的膠殼都棄用了，只用熱縮膠包住焊接部份。

上星期我一尐聲氣都無，我去了哪？

Continue reading →

人問問題時，最常犯的錯誤，除問了等於無問外，就係問得太深入。

不時收到這類問題：「(Microsoft) Excel 中怎樣改圖表類型？」如果我不在場，好多時我會老實答「唔識」，有時我會撞下；如果我在場，我會親自操刀。不能理解我的回答？試想像一下我問你以下問題：「Foobar2000 (一播放軟件) 如何顯示專輯封面？」、「Cowon S9 (PMP) 如何較大細聲？」答案，當然係要試過先知，又或者要睼說明書。當你不是用家，亦沒有軟件 / 機在手，甚至連圖都無幅的時候，你怎知道實際操作？

一般人應該都越來越面對得多這種事，因為軟件巨頭 Microsoft 近年產品的界面變化越來越大。由 Microsoft Office2007 的「Ribbon」到 Windows 7 的 toolbar，鑊鑊新鮮鑊鑊甘。

至於為何我不能講出做法而實際做又做到？因為我從不記做法。我只記原理。每一次我做都只係由原理開始，重新摸索如何做。無法，經常遊走不同平台，每個平台都有自己的做法，唯背後原理相似。總不能死記做法，否則走到新平台時，不懂找方法，就死定了。

This is true, in the end. I don’t believe this previously, but after a study of the notorious Loudness War, I changed my mind.

Some may already get the idea by the name of “loudness war”.

Laymen usually think a louder track sounds better. Since the way the human ear perceive voice, a louder (technically, higher sound pressure) sound is heard more by the low and high frequency, hence the feel of sounding better. And of course a louder voice always takes more attention than a wimpier one. So, Artists and publishers both want their record to be louder than their competitors in the same volume, and they requested the mastering engineer to do it.

But there is a caveat, it is the way CDs store audio information. Ever heard of 16 bits / 44.1 kHz PCM? That is the standard format of audio CD. Every second there are 44100 (44.1k) sample blocks that record the current sound level, each 16 bit wide, that means a combination of 16 zeros and ones. This 16 zeros and ones translate to a spectrum from totally silence to the saturate point 96 dB. So 96 dB is the full scale of 16 bit audio, which should never be exceeded, otherwise the information will be lost. This full scale may be referred to as the 0 dBFS, all sample should be below 0 dBFS.

In the ’70s and early ’80s, CDs were recorded in a conservative -20 dBFS. Some day in some year, one (mindless) guy said, hey, his music is louder than mine! I want my music be as loud as his! No, even louder! And suddenly, it comes. All records are being louder and louder, until… until they hit the barrier of the medium, 0 dBFS.

Since they can’t get anywhere beyond 0 dBFS, otherwise the waveform will be distorted, they can only look back, not bring down the volume back, but pump up the quite part without peaking the loud part. This is the concept of dynamic compression. With this evil practice, the mastering engineer managed to continue making records even louder. But things gets ugly rapidly, as the originally loud parts were all gone. Why? Course when there is no quiet, there can be no loud. When there is no dynamics, there is no impulse, no punch, no emotion. No record is getting any louder hence, since all listeners start turning their volume knob. Well, something seems very wrong…

Then here we go to the 21th century, and we don’t see any stop sign. Metallica’s 2004 album is at average -4 dBFS, always peaks 0 dBFS, one volume from start to end, and it sounds like hell. Some songs are re-released on the game Guitar Hero, which didn’t get dynamic compressed, and sound much better, though still can be better.

So we’re here, in the age of crap music.

It’s not the fault of CD. In fact, CD is superior to vinyl. But, well… I am looking at you, mastering engineers.

Next time when you listen to your music collection, and found some drum shot are soft, muffled, then go damn the publisher, the master engineer AND the artist.

Turn Me Up! (which means “use your volume knob”) was raised to encounter the problem, but sadly it’s still not doing quite well after 1 year of launching. Last time I go to HMV, which is on today, not even one album has this sticker on it.

Further reading:

• Music Machinery: The Loudness War Analyzed (Very detailed)
• Dynamic Range Database
• European Broadcasting Union: Level and distortion in digital broadcasting

上個星期搞了個所謂的問答比賽 (這個電腦學會主席總於有事做了)。收回答案，批改之時，發現我學校的人，係唔識英文的。

第二十條問題：互聯網英文 www 全寫是什麼？

1. world wide web
2. world wild web
3. world war web
4. what was web

這問題，其實與電腦無大關係，主要係睼你識唔識英文。四項都有人揀，這一早預徂，但誓估不到，好多人圈了B。

Holy cow.

英文老師一進來，只是分發着單行紙，在黑板貼上液晶計時器，一聲「七十五分鐘內完成。」後，就坐在教師座，一聲不發，像從沒進入過這課室一樣。

看了看題目，總是沒什麼動力下筆。

成績表上的紅字，老師與父母的責備，輔導老師的提問…猶如穿越時空的記憶，清晰地浮現在腦海。

無意識地望向你的座位，看見高才生的筆寫個不停，已經遊走於頁末了。這影像把思緒拉回了現在，眼中方能看到課室的現狀。一些文字在腦中閃過。一反常態地，我把腦中未經整理的文字寫在紙上。

作文非我強項，七十五分鐘過去了，計時器「嗶嗶嗶」地響，紙上只零零碎碎、斷斷續續的寫了二百多字，距離目標四百五十字，還有七十五分鐘。結果還是沒變吧，我想着。然而一切已顯得不重要了，我把所有東西拋諸腦後，眼前所剩的，就只有高才生，那個倦意滿臉，伏在桌上休息的高才生。

寫作耗時：二十八分鐘
字數：三百零七字，加四十二個標點符號。

後記：細節位不夠細節，重點不突出。有寓意的描寫，沒有加以擴展，反而沒有用的寫得太多。原本的重點在哪？寫完後已經忘了。不，應該動筆時就已忘了，否則寫出來的不會如此空洞。

這令我多麼的安心。看見你在面前，在面前與眾人談笑着。你是如何的歡樂，連少少的笑點也不放過，連那冷得不可再冷的笑話也不放過，臉上應該掛着笑容吧？我看不到，只能靠估。你以前也是這樣的嗎？我答不上，不是忘記了，只是影像已變得模糊，好像被些什麼扭曲了。

看見你在面前，你應該是在笑吧，在跟朋友說着些有的沒的，什麼也能談上，什麼也能開懷大笑。

好像沒有發生過什麼似的。

也許真的什麼也沒發生過。

寫作耗時：十一分鐘
字數：一百六十八字，加二十一個標點符號。

後記：呃…原本還想寫「你好像在隱瞞什麼」的，結果完全忘記了。

一時積下不少作業…先記低下星期到深水埗時的購物清單

日前到友人家檢查電腦。事主無聊，借了 S9 去聽。

聽了一陣，問到為何我聽的歌那麼怪。「是嗎，我覺得幾好聽。」「全部都聽不懂的。」正巧連續播了好幾首日文歌。「我比較喜觀聽一些聽懂歌詞的歌。」

和我正好相反。我聽歌多數會聽一些我聽不明的語言，或乾脆聽音樂算。其實我可否算作聽歌呢…就算聽一些聽得懂的語言，我都會忽略歌詞，除非那是帶強烈訊息 (非強烈感情) 的歌。正因為人聲在心中排位較低，所以我才會選擇 ER-4P 這種人聲較「後」的耳機。還是因果倒轉了？是因為選擇了 ER-4P 才無可奈何地忽略人聲？無論如何，結果就是，我近來聽人講話時，經常知道有人在說話，但不知他在說什麼。

畫公仔畫出腸，是好無癮的。但我話大家知，聽 (看) 的那個更無癮。當字字句句都聽得明，你到底是在聽歌的感情，還是詞的感情？是歌的意思，還是詞的意思？而由於全世界大多數的歌都在講那幾件事，詞的意思即係無意思。

至於為何是日語嘛，第一當然係因為我不懂，第二係因為簡單。日語只有五個元音，五個十分簡單的元音，而輔音亦不怎複雜，如此，說日語的人好容易給人一種「在學說話」的感覺，稱之為萌化亦未嘗不可。所以我頗討厭男生講日語…

今日再次體現到萌化的威力，一個樣貌、聽線皆不太出眾的同學，在早會時的日文學會宣佈中說起日語，頓時迷倒萬千男同學，引用副校長：「真是十分精彩的宣佈。」連副校長都話一個宣佈精彩，日語魔法可見一班。

還有一個同日語無關的原因，就是日本人喜好豐富音軌。雖然韓國歌亦多如此，但…樣的問題，韓國歌手多數造佢，而且，我對韓國無好感。係，係好主觀，但恁客觀恁理智做乜。用 Cowon S9 + Etymotic ER-4P 這種高分析力組合，聽者會有意無意地追細節，所以對音軌不多的音樂容易覺悶；但對音軌豐富的則會聽上癮，每次聽都有可能聽到曾經忽略的聲音。講起這組合，聽了大半年，開始覺得這組合不能滿足我了。雖然偶爾戴得好的話還是有驚喜，但 S9 的電子味還是太重了，尤其高音，簡直像在聽方波，經朋友的「DIY PtoS 線+」(實為 100 Ω 抗阻線，一般 PtoS 線為 75 Ω) 後更加表露無遺。加上 S9 無 Line-out 輸出，又不能換OPAMP，限制日後升級，所以…兩年不升級可能做不到了。

Finally.

Getting rid of Windows had been my dream for sometime. For monetary reason, Mac is a no go, so I go for the Linux way. Road was rough, but I got through it… except there is a (not so small) problem: audio. ALSA recognised my on-board Realtek AIC880 sound card, but somehow misconfigured it, leading to low volume and hiss and cracked sound. This is a problem started from day one, and I can’t find a solution even now. I doubt if there is a solution without a firm-released driver.

Year after year, I have no choice but dodging it. And now… technically I’m still dodging it (XD). But things changed, by friend’s help. He lent me his unused Creative SB X-Fi Surround 5.1 USB sound card, sine die (:D). Better sound quality than any Realtek chip, fully supported by ALSA (though no button can be used, but that’s enough), I am fully satisfied and listening music through my beloved ER-4P in Ubuntu now.

Thank you very much, Phenix01.

Because my ALC880 supports digital output, a fast, low technical knowledge required, but high cost solution would be an external DAC and / or amplifier.