In this brief article I will attempt to explain what the vocaloid resonance controls are,
and give some pointers as to how they can be used. Please note that this is a small article,
and will only scratch the surface of what is a massive subject. I hope this tutorial will point
users in the right direction, and inspire them to experiment.
在这篇简短的文章中，我将解释什么是VOCALOID的“RESONANCE”共鸣腔，
并给出一些如何使用它们的例子。但是请认识到这仅仅是一篇短文，我只能简单的描
绘出共鸣腔这个复杂而庞大的控制参数的表象。我希望能借此文为广大使用者指出正
确的研究和试验的方向。
What is a resonance?
什么是共鸣腔？
What they aren't: The Vocaloid resonance controls are NOT a parametric EQ or a filter!!
.... Well not in the conventional sense anyway, but you can get some crazy synth like filtering
effects when using extreme settings.
首先指出它不是什么：VOCALOID的共鸣腔不是一个传统意义上的参量EQ控制或者一个过滤器，
但是你可以用极端的参数设置去得到类似过滤器一样的效果。
这文章我太喜欢了……处处都说到了我说不出的感受……我向别人传教的时候无法解释什么的是
共鸣腔，解释了对方也很难以理解，于是我总是先从EQ的角度出发，然后对他们说，这是一个
和EQ类似的但是算法和效果完全不一样的参数，可以用EQ去触类旁通，但是实际用法是不一样
的，需要自己揣摩。从最开始接触的时候我也以为它是EQ来着，但是用完了发现效果完全不是
一回事。于是请记住这货绝对不是后期里面用的那种参量EQ！！！！！！
What they are: Vocaloid singing happens by attempting to synthesise the resonances of the
human voice. There are many resonances covering the whole vocal frequency range, these
resonances are known as formants. Formants are crucial as they not only govern the timbre
(gender, richnes etc) of the voice but also the pronunciation of the vowels. During synthesis,
Vocaloid is shifting the resonances around, effectively morphing the sounds into each other.
共鸣腔是：VOCALOID的歌声是模拟人类共鸣腔的结果。人声波段由许多共鸣腔构成，这些共
鸣腔表现为共振峰。共振峰至关重要，因为它们不仅仅影响声音的音色（例如性别和饱满度等）
还影响元音的发音。VOCALOID是通过调整声音的共鸣腔来有效的合成歌声。
For example the phoneme 'aI' (sounds like 'eye') starts with the sound 'uh' and then changes
to 'ee', with a slight 'y' sound. During synthesis each of the resonances will move producing the
required sound.
举个例子：音素“ai”（发音类似英文eye 中文“爱”）由“uh”开始，然后变换到“ee”最后用一个较
轻的“y”结束。在发声的过程中，每个共鸣腔都将会提供需要的声音部分。如下图：
The 4 resonance controls in the vocaloid editor modify the values of the first 4 formants (0-3 in the graph).
They are the lowest in frequency and they are the most important as they are not only the loudest,
but they also dictate the percieved shape of the mouth and how a vowel is pronounced.
VOCALOID里面的4个共鸣腔参数对应控制如图的0到3区间里的共鸣腔声音。这四个频段的
共鸣腔是最重要的，因为它们不仅仅是声音中最重要和最响的部分，还决定了口型和元音的
发音。
Each of the 4 resonances has 3 controls... amplitude (volume), frequency (pitch) and band width (Q),
each of these ranges from 0-127 with 64 being the default.
每一个共鸣腔参数由三个部分构成：
amplitude 对应部分的共鸣腔音量大小
frequency 对应部分的共鸣腔的震动频率。
（原文用了pitch来表述frequency的作用 Pitch是一种音高定义：用一秒钟的振动数来
表示。频率次数多者音高，频率次数少者音低。将每秒振动440次的声音定为“a”，是目
前国际通用的标准音。）
band width 对应部分共鸣腔的影响波段宽度。
每个参数范围从0到127，64是默认值。
It is important to note that the resonance controls are relative..... So, the changes
you make to theresonance tracks are affecting the synthesis that is already taking place.
还有重要的一点是共鸣腔的参数控制是相对的，也就是说你对参数的改变是相对已经存
在的合成声音的改变。
For example if a resonance frequency control is set to 64 you will get no change in sound.
And if you increase the frequency of the resonance, you are increasing the frequency of that
formant above where vocaloid would normally be synthesising it. The same principle applies to
both Amplitude and bandwidth as well. This is why the controls are measured in a relative manner,
and NOT in specific values like dB and hz.
举个例子：如果一个frequency的参数设置为64，你会发现合成的声音没有变化。如果你
升高这个共鸣腔的frequency参数，VOCALOID在合成的时候就会增加了这个频段的震动
频率，然后就能听到合成声音变化了。Amplitude和bandwidth也一样是相对参数。这就是
为什么这些参数控制给出的数值都是0到127的相对值，而不是一个精确到HZ或者DB的
绝对值。
怎么用？
One of the main uses for the resonance controls is to change the pronunciation of the vowels.
You canchange one vowel into another if you want, or you can make hybrid vowels that are
in between 2 normal vowels. Its a tricky technique, but you can even use this to change the
dialect of vocaloid.
经常用到的一种共鸣腔控制是改变元音发音。你可以把一个元音发音改变成另一个，或者做
出介于两个元音发音之间的元音。这是个复杂的技巧，但是通过改变共鸣腔你甚至可以改变一
个VOCALOID的口音。
You can also use resonances to thin or thicken up both vowel or voiced consonant. But as the
resonances do not affect the noise component of the synthesis, they will not affect unvoiced
consonants in any way.
你也可用通过修改共鸣腔削薄或者加厚辅音和元音。但是共鸣腔不影响无声的气息部分，
它只对有发声部分产生作用。
图片附件: 004.jpg (2010/12/30 00:47, 31.13 KB) / 下载次数 1
In this brief article I will show a couple of examples of how vocaloid resonances
can be used to change the pronunciation of vowels. Please read part 1 in order
to get the best understanding of what is being discussed.
在这篇短文中我将用几个例子来演示如果用共鸣腔参数修改元音的发音。请阅读part1来获得
对我将要讨论的参数的认识。
Example 1: 'enfold'
This example is from the Zero-G Miriam demo "You are the light".
There is a line that goes 'whispers enfold me'. I am an English person with a very neutral accent
and to me the way vocaloid sings the word 'fold' sounds very incorrect, it sounds over-precise
and posh. The vowel phoneme for 'fold' is '@U', this is the same phoneme that is used in 'phone'
and 'coat', but the way I speak these words doesn't rhyme with the word 'fold', but vocaloid
makes them rhyme.
这个例子是 Zero-G Miriam 的demo曲"You are the light"里面的一句话。那句话是“whispers enfold
me”，这句话里面的enfold中的fold的元音从我的角度听起来不够正确，fold的元音音素是“@U”
虽然和“phone”“coat”里面用到的音素一样，但是从我的角度出发，我念这个单词的时候元音发音
和VOCALOID合成出来的不一样。
Now I am aware that the issue I am describing is a dialect issue, some people may use the same
vowel sound for the above 3 words, but I don't.
现在我发现这个不同是口音的问题，一些人会用同样的元音发音念例子中这三个单词。
But none of this matters at all, because this is simply an example of how the resonance controls can
be used to change the pronunciation. My point is this... If I can change the resonances so the
dialect is good to my ears, then theoretically you can make the pronunciation right for you.
但是没关系，这刚好是个简单的例子，可以用共鸣腔参数来改变发音。我的观点
是如果我可以用共鸣腔参数把发音改成我喜欢的，那么你同样可以！
想要midi的请去原地址下载，但是例子内的歌手是Miriam。
注意：在这篇文章中描述的共鸣腔调整参数不是所有歌手都适用的，但是他们的理论上是相同的，
用在其他歌手身上也应有相似的效果。
It asumes you have Miriam as singer ID 0, if this is not the case on your system then simply open
the singer control track (bottom left of the screen) and double click the block underneath the word
enfold, and select Miriam.
这个例子使用的歌手是Miriam，如果她不是默认歌手，请打开左下角的singer参数把她设置为默认，
前提是你的电脑里有Miriam。
Now to change the sound of the vowel.
现在让我们开始修改发音。
Select the 'Res2 Freq' control track and draw a line at the start of the word fold at 40 with a slope
going back up to 64. Make it look something like this...
选择 res2 f 如图画一条斜线如下图
Select 'Resonance2-Bandwidth' and set it to 51, also with the slope going back upto 64, it should
look like this...
在res2 BW 画一条斜线如下
听前后对比效果请去原地址！
这个作者在这里说的是enfold这个单词的fold的元音发音像“嗷哦”，他认为“哦”他更喜欢，
于是用res2 f 把元音发音口型变圆了，然后用res2 BW 控制了稍微缩小了res2的影响范围。
PS：BW对鼻音的影响真的很不明显……那本身就是个范围……我忘记了谁跟我说BW和
鼻音有关了……然后我也一直以为这东西能调鼻音，不过当然调整res的f去修鼻音的话，用
BW自然可以控制鼻音修的范围了……可以解释为什么我尝试只用BW修鼻音效果会比较微……
------------2011年2月18日更新------------------------
最近调歌的时候顺手测试了下4个res的BW,其中RES2 BW对鼻音的
影响较为大一些，其他3个微弱。我觉得是2频率范围的原因。
----------------------------------------------------
Compare the before and after examples to hear the difference more clearly. To my ears the
pronunciation is greatly improved, especially when in the context of the song. But like I said, it's
just my opinion.
对比一下修改前后的发音，就我听来发音变好了很多，在歌里面也协调了。当然这只是我的个人
观点。
于是这个短文在继续举例子讨论如何使用共鸣腔修改元音发音。如果你不知道我们在讨论什么，
请好好阅读PART1 和PART2。
Example 2: 'die'
In this example I want to use the phoneme 'aI', sounds like 'eye'. There has been various words that I haveused where I have not liked the pronunciation of 'aI', and have been able to fix it very easilly by changing
resonance2. One such word was 'die', so I will use that as an example. Please open this vocaloid midi file.
This example is also made using the Miriam font so remember to switch the singer to Miriam.
需
要midi例子文件的去原地址下载。note 'die' starts.
于是我打开文件，找到die这个单词作了如图的修改：在res2 f里面降低参数到45
Again, to my ears this sounds more correct now. But it is entirely my opinion and the difference is
subtle. It's completely irrelevant anyway, as this is purely an example of how to change the dialect
of Vocaloid. Like I said in part 2, if I can make the accent right for my ears, then you can make it
sound right for you.
但是这是修改VOCALOID口音的一种方法。就像我说过的那样，如果我可以把他们的发音修改的
更合我的口味，你也一样可以做到。
So what now?
Unfortunately messing around with the resonances in Vocaloid is largely trial and error. In
theory you could make any vowel sound like any other vowel, but in practise that is impossible.
My advice therefore is to start with the vowel that sounds the closest to the sound you want,
and then use that as a starting point. It might not be the 'correct' vowel phoneme, especially if
you are trying to simulate a specific accent.
不幸的是，在VOCALOID里面做大量的共鸣腔试验是令人蛋疼的事情。理论上，你可以把任何一
个元音发音改成别的发音，但是实践证明，修改发音的程度是有局限性的。我的建议是，尽量找
和你想要的发音接近的音去作为修改的基础。即便那个音素在你看来可能是不对的，尤其是当
你想要模拟出一种特别的口音。比如说不是单单修正发音，而是完全做出另一种语言什么的！
I find that if I change the frequency of a resonance it's sometimes hard to clearly hear the bit thats
changed. Sometimes it can be helpful to boost the amplitude of each resonance in turn (not too
much or you could destroy your speakers..... and ears!!), and then move the frequency around
untill you find the point that sounds most correct, then set the amplitude back to 64 and move on
to the next resonance.
我发现有时候我修改了res f但是却听不出变化，这个时候可以试试调高那个res amp，这样就曾
加了你修改的那部分res的音量，可以帮助你听出修改前后的区别，但是千万不要把amp加的太多，
爆囧音会让你聋掉！然后你慢慢调整res f的位置，直到找到合适的点，再把amp调回去默认。
It is possible to change the resonances so they swap positions, for example if you increase the frequency
of resonance 2 and reduce resonance 3, they may cross over. This usually produces a synthetic and
sometimes unpleasant sound, you will know it when you hear it.
那4个共鸣腔之间是有交叉影响的，例如你增加了res2 又减少了 res3，他们会产生互相影响。通常
这样会爆囧，当你听到了你就知道了……
不过我个人不赞同这一点……
正因为四个共鸣腔之间会互相影响，所以才能产生更多的可能性。
需要变换语种的话，只用一只会很快上手，但是用几只联动会更美。
一般产生交叉作用的时候：
作用相同的时候可以减低参数修改值，你修改的越少越能保持原来的音质。
作用相反或者可以做一些微妙的变化，比如你降低了某个共鸣腔的频率，那么其实频率变低了，声音会略沉，你完全可以升高另一个共鸣腔的频率保持和原来的发音效果统一……当然这个情
况你直接修GEN什么的也可以啦~我只是举个例子。
剧终！
图片附件: 003.jpg (2010/12/31 08:32, 28.63 KB) / 下载次数 1
图片附件: 001.jpg (2010/12/31 08:32, 30.49 KB) / 下载次数 1
图片附件: 002.jpg (2010/12/31 08:32, 8.45 KB) / 下载次数 0
）时间紧迫心情不好人士还是表看了．．．
）啊啊 吾辈是说大家对Ｖ１里无论哪款有爱＋无比根气都可以调教成中文？！
没用过Ｖ１的初心者的妄想罢了，承认对文章概念不严格谢谢！
| 欢迎光临 iVocaloid论坛 (http://bbs.ivocaloid.com/) | Powered by Discuz! X2 |