So you want to learn Synthesizer V

Hi there!

If you’re reading this post, then most probably you are like me, who enjoyed listening to Hatsune Miku and other Vocaloid (or Piapro Characters) songs. And you thought to yourself, “This is so cool! I want to produce something like this”. With your new-found determination, you are looking for some way to create them sweet, sweet music using a vocal synthesizer.

Background of me

First and foremost, I’m not a music producer. Though I have some knowledge of music since I have been playing the guitar for several years already, and my music theory knowledge is just so-so. I have never touched any vocal synth software similar to Vocaloid, nor have I used any DAW (Digital Audio Workstation) ever. I’m picking up music production as a hobby. Maybe in the future, I would like to try to produce some original work.

From this perspective, the only objective for this post is as follows:

To learn Synthesizer V as quickly as possible.

Methodology

Several approaches to learning can be enumerated as follows:

  • Creating random vocal notes, pressing all the buttons and turning all switches, and trying to break the software. (Not the recommended way, but it’s really fun to do it anyway)
  • Creating a music cover from a well-known work (recommended way)
  • Creating original music using DAW in tandem with Synthesizer V (advance to pro level)

Based on the list above, the 2nd methodology is the best of the 3 since you are treating your learning experience as a project (using the SMART principle, Specific, Measurable, Achievable, Resource-bounded, Time-bounded).

Gathering Resources

Since I’m a big fan of Wowaka’s work (R.I.P.), I have picked “Unknown Mother Goose” to be covered, specifically, the chorus part. Now, I’m not Japanese nor speak in Nihonggo, so I referred to one of the covers of Will Stetson’s English Cover

I also searched for the “karaoke”, or the minus-one file. Fortunately, the karaoke version was provided by the producer and is found in this link (https://www.obvibase.com/p/8aeNzuNgdcTt).

Using Synthesizer V: My Experience

Before I did anything else, I have read the manual (yes, people like me exist) and consulted with the forum for the video manual (credits to Run_djp_Run for his amazing work). It took me around 1.5 to 2 hours of learning all the essentials, including the bells and whistles of the software.

Starting the software, you are greeted with the piano reel, in which all the vocal magic happens. At first, I only used the mouse for the music note input. When it became cumbersome, I started using keyboard shortcuts to hasten my work. The keyboard shortcuts that I used frequently are as follows:

  • Ctrl-B – Edit node properties
  • A, D – Timeline navigation
  • Spacebar – Play/stop
  • Alt-A – Go to beginning of the timeline
  • Ctrl-Alt-M: Open Track Manager
  • Alt-V – Open Phoneme Viewer
  • Ctrl-P – Convert to Phonemes
  • Alt-Q – Hide/Show Panels

Tips on creating vocal notes:

  • Set the BPM to match the song. After some trial and error of me pounding to the quarter notes of the instrumental on the table (I don’t have a metronome), I have found that the BPM of the song is 111.
  • Mute the instrumental, and lower the BPM when you have difficulty with the timing of the vocal notes. This tip helped me a lot especially when there are fast lyrics.
  • Try to sing the lyrics of the song. This helps with the timing of each phoneme of the lyrics. Try to take notes when you sing.
  • Convert weird, and off-timing pronouncing words into phonemes. Break them apart, syllable-by-syllable. Use the Phoneme Viewer to manually encode the phonemes into the lyrics (you can tell when you have converted the lyrics into phonemes by the forward slashes “/ /” before and after the phonemes).
  • Lastly, have lots and lots of patience to yourself. Adjusting the tone, breathiness, and other parameters of the notes are painfully cumbersome. Always remind yourself that this work is worth it.

Conclusion:

It was really fun and it took me around 12 hours to finish the excerpt. Much improvements can be made to the vocals, but I’m really happy with the result. I really enjoyed Eleanor Forte’s voice and vocal range for this song. Here is the link to the finished product.

Hope this post can help people who are interested in trying and producing using Synthesizer V!

Have fun creating!

11 Likes

Hi,

Can you tell me where I can buy a license to use SynthV commercially - the download version.

Cheers,

George.

Hi George,

From what I know with the new Synthesizer V Studio, you can purchase the vocal bank licenses in their new website:

https://dreamtonics.com/en/synthesizerv/

The official release is on July 31, 2020. Eleanor Forte is not available as of release date but the Japanese and Chinese voice banks are already available for purchase as far as I know.

Cheers!

1 Like

Thanks for the tip! hmm… After importing an instrumental music to Synth V, how do you begin plotting the voice in the Piano Roll? I mean do you just listen frequently to the music and play it by ear or do you use a software / guide to know where to place the first word in the lyrics of a song?

Most people will do this by ear. It is common to chart the melody in a DAW instead of SynthV because then you can choose which instrument to use. Some people prefer a clean sine wave, others a square wave, some people use a piano, etc.

It’s mostly based on what the user feels is easiest to listen to and identify that the note entered is the same pitch as the reference material, and often that’s easier with a consistent instrumental tone than with a vocal.


There are some tools which can help provide a guideline, notably Newtone and Melodyne. Both have the option to analyze an acapella and produce MIDI estimating the note pitch and timings. The MIDI produced by these can then be imported into SynthV, however it will usually require significant cleaning-up since the note start/end points are difficult to detect and humans don’t sing perfectly on-pitch.

If you use one of these tools, of course you first need an acapella of the original track, but I would recommend importing the MIDI onto one track and charting the notes on a separate one, using the MIDI as a guide. It ends up meaning you still do things by ear, but it can be much faster this way.

Notice how the notes I’ve charted in this example don’t match up precisely with what the software has detected from the acapella. This is usually due to vibrato, notes where the original singer is slightly off-pitch, early/late note transitions, etc.


Some producers like CircusP will provide untuned SVP files so that people can skip this whole process when covering their songs.

Also keep in mind this thread is from over two years ago, before SynthV Studio was announced. Some of the information above might be referring to the legacy version.

3 Likes

Thanks for the help Claire! :slightly_smiling_face: At the beginning, I plan to just cover existing songs as practice. But it’s difficult to find an acapella of the songs I like. So I’m considering if I can just find or convert the instrumental track of a song to MIDI, maybe I can use that as a rough guide to plot the vocals in Synth V.

In Melodyne Assistant’s “Audio-To-MIDI” feature (or similar software), would you know if it automatically separates the instruments upon conversion and/or import to DAW? I was thinking maybe by recognizing where particular instrument notes fall in a section of a song in the Piano Roll, I might be able to more easily locate the nearby placement for the voice that would hopefully sound in tune. Have you tried that approach and would you know if that is do-able?

I’m new to using Mixcraft 9 Pro Studio DAW (had older & lower version in the past) and it came with the basic Melodyne Essentials. Sadly that does not yet include the Audio-To-MIDI feature and I’m still holding back from upgrading to the higher edition as I’m unsure if it would be suitable for me.

I’ve also only started checking out some free SVP’s online and will see how that works out. :grin:

Ah I just discovered here how to convert an Audio track to MIDI using both Mixcraft 9 Pro Studio and Melodyne Essentials. No need to upgrade Melodyne to a higher edition as Mixcraft 9 Pro Studio has a built-in “Convert Audio-to-MIDI” feature that works well with even just Melodyne Essentials. :grin: I will now test if I can do something with this together with Synth V.

You can use many web service to separate music and vocal, for me the easier way is to record your own MIDI and fill in the lyrics manually, or you can write or open any score with lyrics to Dorico, Sibelius, Musescore etc and export the melody +lyrics and convert it to svp easily

2 Likes

The best way to get better at charting notes by ear is to start doing it. I expect the approach you’re considering will make things more difficult rather than easier, since instrumental tracks have many layers and the instrumental melody won’t always match the vocals.

Most original vocal synth songs on YouTube will provide a download for this. If the song was made with SynthV it will already be in SVP format, but if it was made with Vocaloid you can convert the VSQ to SVP using UtaFormatix. Keep in mind in the latter case you may need to correct some phonemes since SynthV and Vocaloid use different phoneme syntax.

https://sdercolin.github.io/utaformatix3/

2 Likes

Hi! Can these software isolate, separate, and extract the vocal from the music from an imported audio file like WAV or MP3? Sort of generating an acapella sound file?

Synth V cannot extract audio stems, but the online web service can. Or you can buy RIPX to extract all the tracks

1 Like

You’re right. It helped a little but it’s still difficult placing the lyrics in the proper position on the Piano Roll to sound right. I noticed when I reopen my Synth V project file, it sounds different than when I saved the song last time. I have to stop using the program for a few hours and have to listen to the song again to re-check the tune. :laughing: It’s just my second try doing this with Synth V. I need more practice with it.

How I wish there was a software that can extract only the vocals from a song file and convert it to MIDI. It does not need to be really accurate because it can just serve as a guide in Synth V. :grin:

By the way, does all Synth V voice bank use ARPABET phoneme format? If I edit a phoneme and that particular lyric suddenly falls silent, what does that mean? Maybe I’ve made a mistake I guess?

Will delve into using SVPs soon. Thanks for the link to UtaFormatix! :slight_smile:

What online web service?

Whoa, that’s the first time I learned about RIPX. It’s like a dream as it being able to separate the voice and convert to MIDI! The price is kinda expensive though for a hobbyist. But will try it out. Thanks for the suggestion! :slight_smile:

It looks like vocal separation can also be done in Melodyne. I’m just not sure in what edition of the software. Will check mine out.

But using these websites with free services might be easier:

Will see what I can do with these. Thanks for all the help! :smiley:

For English, yes. SynthV also uses romaji for Japanese and x-sampa for Chinese, since each language requires a different set of sounds.

A note will be silent in two main situations:

  1. It contains an invalid phoneme
  2. It is overlapping with another note on the same track or in the same group
1 Like

Another good online voice / instrument separation I’ve used is this:

There are free and subscription options to this one [if you pay, there are extra features – I’ve only used the free]. This has worked pretty well for me, but I’ve never heard a really clean separation with any app. This may not matter so much with some genres.

1 Like

Thanks again! :slight_smile: