2013/02/28

Android上的日文語音合成方案 SVOX 和 N2 TTS

(Rome.Italy)
快忘了自己曾經去過這個城市,這個國度。
在隨處可見古蹟的城市中游盪,照片怎麼拍,都很賞心悅目。

****
之前在這一篇中介紹到svox這個在android上,提供多國語音合成資料庫的軟體,拿來搭配moon+ reader確實很不錯;可惜的是,在台灣竟然沒有辦法購買它的語言資料庫,只能下載free trial version下來試用個一陣子。雖然網路上有提供怎麼可以變成永久試用版,但這始終不是個好的方法。

所以,又再找了一下,看是不是有可以在台灣購買的軟體,或是免費,但是品質又不會太差的替代方案。N2 TTS似乎是個不錯的選擇。這是由日本KDDI的某個lab製作的,除了有免費版可以使用外,還有官網和twitter帳號。就官網來看,它有以下幾點特色:
 - 不佔容量:有男聲,女聲,總共才5.4MB
 - 不需要網路,android單機即可使用
 - 有男聲也有女聲
 - 符合android標準的TTS 介面 (到Settings中可以做切換,並讓其他ap p使用)
 - 可以個別調整語音 (這功能就得付費才有囉)

馬上把它裝到android上測試,試用okay~~ 聽書去!

REF:
官網
網路上可以下載的地方

Big Data and Hadoop

 Spring.Taiwan

Read the book "Understanding Big Data" today. Half way done.
Here's some excerpts from the book:

Understanding Big Data

2013-02-28 15:45:37
the term Big Data applies to information that can’t be processed or analyzed using tradi-tional processes or tools
2013-02-28 15:46:12
An IBM survey found that over half of the busi-ness leaders today realize they don’t have access to the insights they need to do their jobs
2013-02-28 15:48:45
Even if every bit of this data was relational (and it’s not), it is all going to be raw and have very different for-mats, which makes processing it in a traditional relational system impractical or impossible
2013-02-28 15:49:05
variety combining to create the Big Data problem
2013-02-28 15:51:31
Three characteristics define Big Data: vol-ume, variety, and velocity.
2013-02-28 15:54:17
the opportunity exists, with the right technology platform, to ana-lyze almost all of the data (or at least more of it by identifying the data that’s useful to you) to gain a better understanding of your business, your customers, and the marketplace
2013-02-28 15:55:35
a fundamental shift in analysis require-ments from traditional structured data to include raw, semis-tructured, and unstructured data as part of the decision-making and insight process
2013-02-28 15:56:27
To capi-talize on the Big Data opportunity, enterprises must be able to analyze all types of data, both re-lational and nonrelational: text, sensor data,
2013-02-28 15:56:50
Dealing effectively with Big Data requires that you perform analytics against the volume and variety of data while it is still in motion, not just after it is at rest
2013-02-28 15:58:11
Hadoop-based platform is well suited to deal with semistructured and unstructured data, as well as when a data discovery process is needed
2013-02-28 16:07:26
it is about dis-covery and making the once near-impossible possible from a scalability and analysis per-spec-t
2013-02-28 16:09:19
creator Doug Cutting’s son gave to his stuffed toy elephant
2013-02-28 16:09:44
duce)—more on these in a
2013-02-28 16:10:23
this redundancy provides fault toler-ance and a capability for the Hadoop cluster to heal itself
2013-02-28 16:10:55
Some of the more notable Hadoop-related projects include: Apache Avro (for data serializa-tion), Cassandra and HBase (databases), Chukwa (a monitoring sys-tem spe-cifically designed with large distributed systems in mind), Hive (provides ad hoc SQL-like queries for data aggregation and summariza-tion), Mahout (a machine learning library), Pig (a high-level Hadoop programming language that provides a data-flow language and execution framework for parallel computation), ZooKeeper (provides coordination services for distributed ap-plications), and more
2013-02-28 16:12:23
throughout the cluster
2013-02-28 16:13:54
For Hadoop deployments using a SAN or NAS, the extra network commu-nica-tion overhead can cause performance bottle
2013-02-28 16:14:59
an individual file is actually stored as smaller blocks that are repli-cated across multiple servers in the entire cluster
2013-02-28 16:16:09
default size of these blocks for Apache Hadoop is 64 MB
2013-02-28 16:21:41
All of Hadoop’s data placement logic is managed by a special server called NameNode
2013-02-28 16:28:14
All of the NameNode’s infor-mation is stored in memory, which allows it to provide quick response times
2013-02-28 16:30:33
Any data loss in this metadata will result in a permanent loss of corresponding data in the cluster
2013-02-28 16:32:25
devel-oper doesn’t have to deal with the concepts of the NameNode and where data is stored—Hadoop does that for you
2013-02-28 16:38:42
The first is the map job, which takes a set of data and converts it into another set of data, where individual elements are broken down into tuples
2013-02-28 16:38:51
reduce job takes the output from a map as input and combines those data tuples into a smaller set of tuples
2013-02-28 16:41:20
All five of these output streams would be fed into the reduce tasks, which combine the in-put results and output a single value for each city, producing a final result set
2013-02-28 16:43:08
MapReduce program is referred to as a job. A job is executed by subse-quently breaking it down into pieces called tasks
2013-02-28 16:43:45
An application submits a job to a specific node in a Hadoop cluster, which is running a daemon called the JobTracker
2013-02-28 16:44:39
In a Hadoop cluster, a set of continually running daemons, referred to as TaskTracker agents, monitor the status of each task
2013-02-28 16:46:39
This direct-ing of records to reduce tasks is known as a Shuffle, which takes input from the map tasks and directs the output to a specific re-duce task
2013-02-28 16:47:29
under Hadoop are written in Java, and it is the Java Archive file (jar) that’s distributed by the JobTracker to the various Hadoop cluster nodes to execute the map and reduce tasks
2013-02-28 16:48:00
BigDataUniversity.com and download Info-Sphere BigInsights Basic Edi-tion (www.ibm.com/software/data/infosphere/ biginsights/basic.html
2013-02-28 16:49:58
Hadoop Common Components are a set of libraries that support the var-ious Hadoop subprojects
2013-02-28 16:50:50
When you delete an HDFS file, the data is not actually gone (think of your MAC or Windows-based home computers, and you’ll get the point). Deleted HDFS files can be found in the trash, which is automatically cleaned at some later point in time
2013-02-28 16:52:46
we cover three of the more popular ones, which admit-tedly sound like we’re at a zoo: Pig, Hive, and Jaql
2013-02-28 16:53:09
Pig was initially developed at Yahoo
2013-02-28 16:56:02
you can FIL-TER out rows that are not of interest, JOIN two sets of data files, GROUP data to build aggrega-tions, ORDER results, and much more
2013-02-28 16:58:01
There are three ways to run a Pig program: embedded in a script, embedded in a Java program, or from the Pig command line, called Grunt
2013-02-28 16:58:51
Some folks at Face-book developed a runtime Hadoop sup-port structure that allows anyone who is already fluent with SQL (which is commonplace for rela-tional data-base developers) to leverage the Hadoop platform right out of the gate. Their cre-ation, called Hive
2013-02-28 16:59:50
You can use the Hive Thrift Client within applications written in C++, Java, PHP, Python, or Ruby
2013-02-28 17:01:19
Hive is read-based and therefore not appropriate for transaction processing
2013-02-28 17:03:11
JSON is built on top of two types of struc-tures. The first is a collection of name/value pairs
2013-02-28 17:03:35
The sec-ond JSON structure is the ability to create an or-dered list of values much like an array, list, or se-quence you might have in your existing applica-tion
2013-02-28 17:09:13
The operand used to signify flow from one operand to another is an arrow: ->. Unlike SQL, where the output comes first (for example, the SELECT list)
2013-02-28 17:15:16
Jaql is a flexible infras-tructure for managing and analyzing many kinds of semistructured data such as XML, CSV data, flat files, relational data, and so on

REF:
http://www.ibm.com/developerworks/wikis/display/db2oncampus/FREE+ebook+-+Understanding+Big+Data

2013/02/21

在未參加iOS developer program的前提下,在iPhone/iPad實機上測試開發中app

(Taipei.Taiwan)

該做的正事不認真,寫程式倒是花了不少時間。
今天總算有了點進展,先是把addTags的功能搞定,
再來是加上了Favorites的頁面,可以顯示自己加過特定tags的照片。

有了這些自己想要的功能之後,光是在emulator上測試,
就開始覺得不大過癮,想找辦法把程式放到iPhone上玩。
為了要把app deploy到手機上開發測試,
一般來說,得要加入US$99/year的developer program才可以。
加入了這個program之後,除了可以在真機上測試,
還可以發佈app到AppStore上去。
不過,目前我還沒有這樣的需求也沒有這打算,
因此實在是很不想花這筆錢。
而且,也不確定自己之後是不是還有時間研究iOS...

由於我的手機已經JB了,想說總會有辦法藉此到達我的目的吧。
果不其然,網路上從很早以前就有很多教學在講怎麼設定和patch xcode。
但是…全都好複雜呀:
又要加certificate,又是改plist,還要加入scripts。
只要其中哪個步驟沒弄好,應該就不行了吧。

好在,找到有人直接提供了懶人包的程式,
只要輕鬆按幾個按鈕就可以把所有的patch給搞定。
步驟如下:
1. 下載JailCoder,在Mac上執行它
2. 根據畫面最上方的Guide button操作,把certificate先搞定
3. 按左下方的patch xcode
4. 再按右下方的patch project,選擇自己想要deploy的project
5. 替自己的iOS device裝上AppSync (前提是要先JB)
    AppSync的source repo path下面有,
    把它加到Cydia的軟件源中,再尋找AppSync這支app就可以了

REF:
JailCoder
AppSync Installaion Source:
http://cydia.myrepospace.com/Bl00dra1n

2013/02/20

難搞的oAuth Process和文件不夠清楚的flickr Developer Documentation

(JingMei Bridge.Taipei)

試圖為抓下來的圖片加上一個tag,查了flickr api dev guide,對於這種request應該要使用post method才行。可是flickr上的api explorer還是用舊的token方式,而api說明頁又說得不清不楚,只說舊版的token要deprecated,要用新版的oAuth。卻沒說清楚oAuth的signature要怎麼生出來。

後來又回頭看了幾次oAuth signature的生法,並且在我原本找來做oAuth認識的class中找到相關的code;這才把它們改了一下搬到我需要的logic中。

結果發現用post method還是會有問題,可是我把整串request直接貼給browser竟然可以了!這是說如果用oAuth機制的話,不需要用post method也可以嗎?那我又何必大費周章的去改用NSURLConnection呢?再改回原本的[NSString URLWithString]]就好了。

REF:
gitk documentation
oAuth 簡介

2013/02/18

iOS Programming

(Chamonix.France)

看了幾課教學後,總算是有點小進度了。
不看還不曉得,原來xcode的開發環境這麼完整和方便,
是我太久沒有接觸開發環境了嗎?
感覺上eclipse的android開發環境還是那麼地原始呀。

不過,寫程式常會遇到鬼打牆…
需要花時間…

REF:
Stanford CS193p

2013/02/17

如何快速移除iTunes中找不到檔案的mp3 tracks

(GuYiYuan.Shanghai)
冬陽,雖不暖身
但暖人心

****
之前把mp3加到iTunes時,都沒有選擇copy一份到iTunes中,
因為我的MBP是SSD硬碟,容量沒有大到在裡頭放一堆音樂檔。
但是問題來了,一旦外部的mp3檔被我手賤砍掉時,
在iTunes中常常會遇到播放不了的問題。
iTunes又沒有介面可以很方便地移除這些已經找不到檔案的mp3 track。

網路上找了一下,似乎在Windows上有javascript的解法,
只要下載個js檔,從檔案總管中執行一下就可以;
另外,也有很多要錢的軟體可以幫忙解決問題。
(有沒有搞錯啊…這個也需要花錢去買?)
最後,終於找到了Mac上可以使用的解法:
方法是去下載一個applescript的執行檔,
把它塞到iTunes的scripts目錄下,
再想辦法執行它就可以了。
我一直找不到從iTunes執行該script的方式,
最後,還是直接點了該app,它還是很乖地完成了工作。

現在,iTunes乾淨多了,我又可以放新的mp3進去了。

REF:
Windows:
http://www.makeuseof.com/tag/effectively-remove-duplicate-tracks-itunes-windows/
Mac:
http://dougscripts.com/itunes/scripts/ss.php?sp=removedeadsuper


2013/02/15

我可能不會愛你

 (Hualien.Taiwan)

過年期間,花了不少時間把這部老連續劇看完了。
雖然劇情有點老套,大部分的情節也都在掌握之中,
但是,總會不由自主地一集接著一集看下去,
直到看到結局的那一刻。

李大仁,程又青,在沒有看過這部劇之前,
在網路上頂多只會聽到李大仁,和大仁哥的名字。
看來,陳柏霖真的把這個角色給演活了。
雖然在某些劇情上他演起來還是有點生澀,
但由於角包比較討喜,
大仁哥的知名度硬是比更會演戲的林依晨演的程又青更讓人難忘。

2013/02/12

Water Drop

想用iPad上的Paper畫畫,卻發現自己已經忘了怎麼用每隻筆刷。
試了幾次都得不到自己想要的效果,
最後還是只能回到用Brushes來畫。
一開始還有些不習慣,一直想用用兩隻手指來redo畫錯的部分,
不過,沒多久又找回了之前的感覺。
單純地利用color picker來選畫面上的顏色,
然後透過淺透明度來不斷地抹平色塊。

放大來看,老實說,離真實的水滴還有好大的距離。
不過,縮小看起來還不錯就是了。

得再去多找些paper和brush的教學來看看了。

2013/02/11

功夫夢

(Shilin.Taipei.Taiwan)
 人,如果可以跟花一樣潔白就好了。
 ****
現在在iPad上的視頻軟體已經找不到什麼太新的片子了。
前不久不知從哪看到成龍有部新的電影,裡頭有一個黑人小孩學功夫,
恰好在iPad裡看到,就趁過年的時候把它給看完了。
很八股的一部片,劇情幾乎跟很久很久以前的功夫小子一樣:
一個不會功夫的小孩受到欺負,
後來被一位不起眼的人(通常有高強的武功)收為徒弟,
然後用很奇怪的訓練方式教他武功。
一開始小朋友一定都會覺得莫名其妙,
一旦等到師傅稍加提醒後,
馬上變得功力大增,打得有模有樣。
最後,再來參加個比武大會,打敗當初跟小朋友起衝突的對手。

中間加了一小段成龍的內心戲,覺得有點不必要。
很平凡的一部片,如果不是太無聊,可以跳過不看的。

2013/02/09

BBS鄉民的正義

(XianjiYen.Taipei)
回台灣總是遇到好天氣。
好天氣不出門走走有點對不起太陽。
自己吃完早餐就往仙跡岩走。
仙跡岩這幾年來來回回也爬了很多次,
不費力氣一小時多就可以來回一趟。
當散步也好,當運動放空也好,
可以順便拍拍台北南區的鳥瞰照片,
能見度好時,淡水河再過去的觀音山也跟在眼前似的。
台北的地標:101大樓,台北車站旁的新光三越,
頭一轉全都看得到。

****
原以為鄉民的正義是個名詞而已,
沒想到竟然是部電影。
看來我的吸收新知範圍太狹隘了,
很多事常常不知道不懂。

看完該片後,有點似曾相識的感覺。
由於BBS上的對話跟之前電車男中2ch論壇的討論串一樣,
不大適合直接在電影上用平舖直述的列出來,
所以,也採用類似漫畫中的對話方塊,
用特效的方式表現出來;
另外,關於BBS這個虛擬空間,
也利用了類似日本動畫夏日大作戰中的表現方式,
每個人都有不同的形體,在一個人造的世界中行走,交談。

最後的橋段有點芭樂,不過,還蠻催淚的。
陳柏霖演得還好,還是陳意涵的演技比較到位。

雲圖

(Beef.Dapuqiao.Shanghai)
之前同事來上海出差,一起到田子坊吃飯。
隨便找了一間店進去,菜色看起來還okay,可是牛肉好老,不嫩。
不好吃。

田子坊是上海幾個外國人可能比當地人多的地方,
坐在店裡看路上來來往往的人群,
會讓人有種在歐洲街頭的感覺。

****
由很多不同時空的小故事組合起來,
片中的演員都一人分飾多角,有的很起來很明顯,
有的要到片尾的拍片花絮才恍然大悟。

不過,總的來說,我還是看不懂呀,
看來得看小說才行。

一代宗師

(Centry Park.Shanghai)

看不是很懂,不過武打的畫面處理得很不錯。
很多畫面感覺上是故意弄得暗暗的,強調人物的輪廓和表情,
但是多了,就覺得有點拖慢故事的進行。

梁朝偉和章子怡的打戲都很到位,該用力的地方都有使上力。

2013/02/06

iPad更新app時,完全沒反應怎麼辦

(Canne.France)

不知道是不是因為網路不夠快,常常會遇到這問題。
找了一下網路,看來下面這方法有效:
1. 將 /var/mobile/Media/Downloads 裡所有的檔案都殺掉
2. 重開機

Ref:
http://f.pil.tw/thread-895000-1-1.html

iOS 6 Auto Layout versus Springs and Struts

雲,看了總是讓人有種脫離現實的感覺。

在看iOS programming videos時,xcode中的springs and struts我就是一直找不到。影片裡在做操作時,只能看著乾瞪眼。後來上網找了一下才發現,在iOS6中多了Auto Layout的功能,讓開發者能夠用更好(?)的方式調整UI。如果想用回舊的方式,只需要uncheck一下選項就可以了。

ref:
http://msmvps.com/blogs/kevinmcneish/archive/2012/12/10/tutorial-ios-6-auto-layout-versus-springs-and-struts.aspx

2013/02/03

對岸的她 -- 角田光代


是本得獎的作品,對於女性的內心描述很仔細,
不過我有點沒法抓到本書的重點就是了,
畢竟不是個女生(是這樣嗎?

對於書中描述到日本學生時代,在校園內的小團體問題,
以及長大後,帶小孩,和家人之間要注意的種種事情,
感覺上都離我蠻遠的。

Ref:
http://www.books.com.tw/exep/prod/booksfile.php?item=0010394686

西貝 西北菜

和之前同事吃飯。

騎士團長殺人事件 --村上春樹

今年的日文小說看完了。篇幅很長,但規模不大,故事都繞著幾個主要的人物。情節裡有些超現實的場景,但最後卻讓人有點意猶未盡,因為有些想搞懂的疑點,在還沒講清之前就結束了。