Hello User,
Recording a German voicebank can be a bit challenging if you have never recorded one before. If you're new to voicebank recording, I recommend trying a Japanese CV voicebank first. With patience and some time, however, it's definitely possible to create a German voicebank. Voicebank creation can be a fun and exciting project, but it requires some preparation and knowledge to ensure that your voicebank turns out well. Here's a step-by-step guide on how to record a German UTAU voicebank: |
|
1) Requirements
Firstly, you'll need a set of headphones and a microphone to record your voice. It's important to use headphones to avoid any feedback or echoes in your recordings. A decent microphone is recommended to ensure good audio quality, but if you don't have access to one, you can still achieve decent audio quality by using your smartphone. Simply record your voice using the voice recorder app on your phone, making sure to hold the phone close to your mouth and speaking clearly. If you only have access to a low-cost microphone, you can still improve the overall recording quality by using audio software such as Voicemeeter Banana.
Secondly - unless you use you smartphone for recording - you'll need a computer to install and run Oremo, which is the software used for recording and managing UTAU voicebanks. It's important to note that Oremo and other UTAU-related programs will only work properly if your computer's local settings are set to Japanese. This may take some time to set up, but there are many online resources available to help you with this process.
Lastly, you'll need to download and install Oremo as well as the German CVCV Recording Set-Up.
Once you have Oremo installed and your computer's local settings are configured properly, you can begin recording your German voicebank.
Secondly - unless you use you smartphone for recording - you'll need a computer to install and run Oremo, which is the software used for recording and managing UTAU voicebanks. It's important to note that Oremo and other UTAU-related programs will only work properly if your computer's local settings are set to Japanese. This may take some time to set up, but there are many online resources available to help you with this process.
Lastly, you'll need to download and install Oremo as well as the German CVCV Recording Set-Up.
Once you have Oremo installed and your computer's local settings are configured properly, you can begin recording your German voicebank.
2) Choosing a reclist
When you are preparing to record your German voicebank, it's important to choose one of the three standalone Core reclists available. It's important to note that this decision will only impact the recording process and will not affect any further voicebank creation steps. Choosing the right Core reclist depends on your recording experience and preferences.
Core (1) is great for people who prefer short recordings in one or two mora style. It can also be used as a word-based reclist, which makes it ideal for native speakers. To use it as a word-based reclist, you must put the OREMO-comment.txt if the Core (1) folder in your recording folder. This reclist is perfect for beginners who want to get started with recording a voicebank. Core 1 has a simple structure and is easy to record. The simplicity of Core 1 makes it ideal for people who are new to voicebank recording and may not be familiar with more complex reclists.
- Example voicebank -
Core (2) is a great option for those who have experience with CVCV v.1 and prefer a recording style that allows for longer recordings. It's also ideal for those who prefer reclists that are sorted by consonants. If you don't have an existing CVCV voicebank to expand, you can simply pick Core (2) (100).txt . However, if you do have an old voicebank that you want to expand, you can create a new folder inside the voicebank and record with CORE (2) (56) as add on for CVCV v. 1.txt .
- Example voicebank -
Core (3) is my personal favorite due to its simplicity and efficiency. It consists of only five-mora recordings, and the reclist is sorted by vowels, making it easy to pronounce since it's easy to get the vowel right. Furthermore, an OREMO-comment.txt is provided to assist with getting the vowels right, including German, English, and French pronunciation descriptions. If you're unsure which reclist to choose, I highly recommend Core 3.
The Extra reclist is an add-on reclist for those who have already recorded a Core reclist and want to add more complexity to their voicebank.
I highly recommend recording Extra for several reasons. Just like Core, there are two options to choose from: Extra (1) and Extra (2). Generally, Extra (1) features easier-to-pronounce and shorter samples, but with a greater number of them. On the other hand, Extra (2) contains longer samples, which might pose a bit of a challenge initially. Choose the option that best suits your needs and preferences, and remember that both offer valuable contributions to your recordings.
Core (1) is great for people who prefer short recordings in one or two mora style. It can also be used as a word-based reclist, which makes it ideal for native speakers. To use it as a word-based reclist, you must put the OREMO-comment.txt if the Core (1) folder in your recording folder. This reclist is perfect for beginners who want to get started with recording a voicebank. Core 1 has a simple structure and is easy to record. The simplicity of Core 1 makes it ideal for people who are new to voicebank recording and may not be familiar with more complex reclists.
- Example voicebank -
Core (2) is a great option for those who have experience with CVCV v.1 and prefer a recording style that allows for longer recordings. It's also ideal for those who prefer reclists that are sorted by consonants. If you don't have an existing CVCV voicebank to expand, you can simply pick Core (2) (100).txt . However, if you do have an old voicebank that you want to expand, you can create a new folder inside the voicebank and record with CORE (2) (56) as add on for CVCV v. 1.txt .
- Example voicebank -
Core (3) is my personal favorite due to its simplicity and efficiency. It consists of only five-mora recordings, and the reclist is sorted by vowels, making it easy to pronounce since it's easy to get the vowel right. Furthermore, an OREMO-comment.txt is provided to assist with getting the vowels right, including German, English, and French pronunciation descriptions. If you're unsure which reclist to choose, I highly recommend Core 3.
The Extra reclist is an add-on reclist for those who have already recorded a Core reclist and want to add more complexity to their voicebank.
I highly recommend recording Extra for several reasons. Just like Core, there are two options to choose from: Extra (1) and Extra (2). Generally, Extra (1) features easier-to-pronounce and shorter samples, but with a greater number of them. On the other hand, Extra (2) contains longer samples, which might pose a bit of a challenge initially. Choose the option that best suits your needs and preferences, and remember that both offer valuable contributions to your recordings.
3) How to use Oremo
To set up Oremo for recording your German voicebank, you will need to make sure again your PC's local is set to Japan and all format settings are set to Japanese standard.
Once you have downloaded Oremo, open it and go to "File >> Load Voice List" to select the reclist you want to record. You will now see a list of syllables that need to be recorded in the upper left corner. Next, choose an empty folder on your computer where you want to save your recordingsand make sure the OREMO-comment.txt is included in case you use one. The example on the right demonstrates the setup for Core (2), but if you use a reclist with an OREMO-comment.txt, it should appear directly under the syllables you must record. |
Next, go to "Option >> Recording Style Setting" and select "Automatic recording 2". In the right border, click on the folder symbol and navigate to Oremo >> guideBGM >> your BGM to select you background music. With Oremo comes a standard BGM on F4 pitch. Use your DAW (something like Audacity) to change the pitch until it feels comfortable for you. Copy paste the F4-100bpm.txt file and name it like your pitched audio.
For recording the Extra reclist, I personally prefered using a 120bpm background music. Once you have set up Oremo and everything is ready to go, take a deep breath and relax for a moment. Remember that recording a voicebank can be a time-consuming process, so make sure you have a comfortable and quiet environment to work in. Get yourself a big cup of tea, put on some comfortable headphones, and prepare to bring your best vocal performance. And remember, even if things don't go perfectly the first time around, don't be discouraged - practice makes perfect! So, take your time and have fun with the recording process. Best of luck! |
4) Prounounciation guide
Navigating the phonetic system of German CVCV can be challenging at first, but don't worry! To assist you, I've created a handy table that will help you understand the different phonemes. Additionally, I've included a link to Wikipedia that provides equivalents of the phonemes in different languages, which can be particularly helpful for non-native speakers. With these tools at your disposal, you'll be able to tackle the German CVCV phonetic system with confidence!
If you're having trouble with pronunciation, don't worry, you're not alone! To help you out, I've created a YouTube tutorial where I provide guidance and audio examples. Watching the tutorial might make it easier for you to get the hang of the German CVCV phonetic system. Give it a try!
If you're having trouble with pronunciation, don't worry, you're not alone! To help you out, I've created a YouTube tutorial where I provide guidance and audio examples. Watching the tutorial might make it easier for you to get the hang of the German CVCV phonetic system. Give it a try!
Vowels
* Wikipedia
The unrounded open central vowel in German (Standard) pronunciation is represented by the letter <a>. There is no simple IPA symbol for this sound. Usually, one of the following three phonetic symbols is used for the unrounded open central vowel:
[a] - the IPA symbol for the unrounded open front vowel: This is the most commonly used written representation. The use of the IPA symbol [a] is one of the reasons why front and central vowels are not distinguished.
[a̱] - composed of the above IPA symbol [a] and a diacritic for retraction.
[ɑ] - symbol for the unrounded open back vowel.
The unrounded open central vowel in German (Standard) pronunciation is represented by the letter <a>. There is no simple IPA symbol for this sound. Usually, one of the following three phonetic symbols is used for the unrounded open central vowel:
[a] - the IPA symbol for the unrounded open front vowel: This is the most commonly used written representation. The use of the IPA symbol [a] is one of the reasons why front and central vowels are not distinguished.
[a̱] - composed of the above IPA symbol [a] and a diacritic for retraction.
[ɑ] - symbol for the unrounded open back vowel.
Consonants
Reclist |
oto.ini |
Example |
Note |
b |
b |
Biene |
|
ch |
ch |
ich |
Say "k" and keep your tongue like that - pressed against the roof of your mouth. Force air through the narrow space in the back of your mouth. It shounds like a softer "sh". |
d |
d |
dann |
|
dz |
dZ |
Dschungel |
|
f |
f |
Hafen |
|
g |
g |
Gang |
|
h |
h |
hoffen |
|
j |
j |
jung |
"y" like in "you" |
k |
k |
Kind |
|
l |
l |
Lamm |
|
m |
m |
Maus |
|
n |
n |
Nord |
|
ng |
N |
lang |
|
p |
p |
Papier |
|
pf |
pf |
Apfel |
Place your bottom lip against your top front teeth and then blowing out a puff of air while releasing the lips. Think of it like saying the "p" sound with a little extra "f" at the end. |
r |
r |
rot |
I say: Gargle so that it makes a noise. Reduce the vibration. ChatGPT says: No, gurgling is not a proper way to produce the German "r" sound. While gurgling may involve a similar sound to the German "r," it is not the correct method for producing the sound in speech. The proper way to produce the German "r" sound involves positioning the back of the tongue close to the soft palate and then exhaling air while vibrating the vocal cords. With practice and proper technique, one can learn to produce the German "r" sound correctly.
I don't know how that is supposed to work. You will get away with the French "r" as well. |
R |
R |
Bär |
I say: Mumble an "a" sound at the end of the vowel. ChatGPT says: No, mumbling an "a" sound at the end of the vowel is not the correct way to pronounce German end r like in Bär. The correct way is to pronounce the r sound by vibrating the back of your tongue against your uvula, while keeping the lips and jaw relaxed. |
s |
s |
See |
"z" like in "zoo" |
ss |
ss |
Wasser |
"s" like in "say" |
sh |
S |
Schale |
|
t |
t |
Topf |
|
w |
w |
Wand |
Round the lips and make a sound that is similar to the English "v" but with more emphasis on the lips. |
x |
x |
Dach |
Clean your throat but remove the coughing part. Make the remaining sound less agressive and strech it. |
z |
z |
Zebra |
"ts" like in "what's" |
zh |
Z |
Genie |