FAQ: What is SynthesizerV?

This is part of an FAQ I'm working on, a more verbose version compared to David's to accommodate different language versions.This English version is translated and some of the wording may differ from the actual options in the software.

I thought you could ask questions about using SynthV in this thread, and let's build this into a quick (but perhaps somewhat fragmented) knowledge base.

Note: This FAQ is still under constant revision; it does not represent official opinion.

What is Synthesizer V?

Synthesizer V is a "vocal synthesizer" in which the creator writes melodies or input MIDI file, and fills in the lyrics in a piano roll-up and the software sings them out. The rich set of parameters allows the creator to adjust the maximum details of the song, including but not limited to pitch deviation, vibrato envelope, loudness, tension, air, gender, vocal mode, etc.

Synthesizer V consists of the Synthesizer V Studio editor software with a separate database of song voices. After you purchase the editor, you will need to purchase the desired voices before you can sing them.

What is a voice database?

A voice database is the official product name for a synthesized human voice, which represents the product as a collection of data on the voice provider's vocal qualities.

Voice databases are virtual singers independent of editor software, and in the music industry, they are a collection of sampled and compressed singer (instrument) characteristics, just like virtual violins and pianos. Synthesizer V's voice database uses a neural network for learning and training.

Synthesizer V's voice database is divided into two types:

- Std, which represents the traditional phoneme splicing synthesis method, where singers need to record their voices independently according to different phonemes and then the software splices them together to sing.

- AI, which represents the synthesis method combined with artificial intelligence learning, where singers only need to sing real songs as samples, from which the neural network learns the singers' vocal characteristics to simulate real vocals from scratch. The neural network learns the singer's vocal characteristics and simulates the real singing from scratch.

Since early 2022, Synthesizer V has fully switched to AI voice database.

What are the characteristics of Synthesizer V's vocals?

Synthesizer V is characterized by a high degree of emulation and naturalness, and is designed to reproduce the singer's vocal technique to the maximum extent possible, and from it to open up areas that the singer has not tried before, or even languages that the singer has not learned. As a result, Synthesizer V preserves the singer's vocal qualities, and each singer's interpretation of the same song will be different, even from one synthesis to the next.

At the same time, Synthesizer V will faithfully reproduce the accent characteristics of the singers so that they will have a different accent when singing songs in a non-release language (e.g. English songs with Chinese singers), and conversely, the creator can select the singers and carefully edit the phonemes so that the singers have a completely different accent.

Synthesizer V does not use the same vocal "base model" and apply different singers' voices to the song; each singer is sampled and trained independently.

What does the AI in Synthesizer V AI mean?

AI stands for Synthesizer V AI, a neural network-based synthesis engine that learns the singer's timbre, vocal habits, vocal technique, the way different phonemes are articulated, and other details from real vocal material recorded by the vocalist, and applies them to the synthesis of the song.

For all of Synthesizer V AI's vocal databases, the voice providers, are paid with reasonable remuneration, expressly know, and agree that the vocal material they provide will be used for synthetic vocal training and for the production of synthetic vocal commercial products.

What features does Synthesizer V have?

Currently, most of the features of Synthesizer V vocals are based on the Synthesizer V Studio Pro editor and the AI voice database. Some features are limited if you use the Basic version of the editor or the Lite version of the voice database. The features mentioned below are based on the Pro version of the editor and the full version of the AI Voice Database.

Auto Pitch Tuning Mode

By default, the software only sings the lyrics at the appropriate pitch after they are entered, leaving it up to the creator to edit the pitch deviation to make it change.

Auto Pitch Tuning mode automatically generates realistic, emotionally rich vocals with pitch deviations within a very short period of time after the user enters the lyrics. Each vocal has their own unique auto-pitch style, so the same melody will sound different with different artists' auto-pitches.

This is especially useful for creators who want to hear the vocal effects of their lyrics quickly and will greatly reduce the amount of work required for editing.

Cross-Language Synthesis

Based on artificial intelligence technology, all AI voice databases can sing Chinese, English, and Japanese songs, and with a simple switch in the editor software, the voice database can be sung in another language. Since version 1.8.0, Synthesizer V Studio Pro has supported note level cross-language synthesis, which means that you can freely write lyrics in different language mixes within the same track.

In addition to the three officially supported languages, creators can make Synthesizer V sing songs in other languages such as Italian by modifying phonemes or even creating language dictionaries, as is already widely and abundantly practiced in Synthesizer V.

For reference, Synthesizer V uses Modified Arpabet for English, X-SAMPA for Chinese, and Romaji-derived symbols for Japanese.

Vocal mode

Based on the different vocal styles of each voice providers, Synthesizer V can extract different vocal modes from the song samples and numerically assign them to help the creator create a different vocal style. At the same time, different vocal modes can be stacked and combined to create more detailed vocal variations.

With the parameter panel, you can draw vocal mode curves to guide the different vocal mode combinations over time, so that the texture of the song changes naturally and smoothly throughout the song.

AI Retake

In a real singing environment, every rendition of the same song by the singer in the studio, or in the concert venue, will have subtle differences. Producers can assemble the best performance from the vast number of overdubs and re-recordings to create the perfect song.

In Synthesizer V, you can ask the synthesized vocals to "retake" the pitch or timbre of any length of notes, just as you would ask the singer to re-sing them in the studio, and by comparing them over and over again, select the most suitable re-recorded version.

VST Support

You can introduce Synthesizer V into the DAW to integrate your workflow by loading a VST3 plug-in.


What are the limitations of Synthesizer V?

- It is not a substitute for basic composition.

Synthesizer V can only synthesize songs in an "input-output" form and does not include any auto-generated content such as auto-lyrics, auto-melodies, etc. It is not a substitute for producing a song from scratch.

- It is not a substitute for post-production.

Synthesizer V is focused on the vocal synthesis workflow and is not optimized for song post-production, mixing, etc. Creators need to provide their own DAW to complete the song post-production.

- No elimination of vocal accents.

Due to the high level of authenticity that Synthesizer V seeks and the in-depth learning of each voice provider's vocal techniques and habits, the voice provider's accent is reflected in the synthesized vocals. In cross-lingual singing, different vocals singing different languages may have accents and pronunciation habits that do not match the standard pronunciation of the target language.

This is a limitation for the creators who want the synthesized vocals to be "standard" for cross-lingual singing, but at the same time, because different pronunciation habits are recorded and there is no so-called "standard" pronunciation, it is possible to edit phonemes to sing various languages and dialects. It is also possible to edit phonemes to sing different languages and dialects.

- It is not possible to synthesize conversational vocals.

Synthesizer V cannot be used to synthesize any kind of conversational or narrative vocals, and it cannot synthesize pure narration in a song, or it would take a lot of effort.

RAP, as a special musical subject, is currently not implemented in Synthesizer V. It is also unclear whether it is in the future singing roadmap.

- Vocal performance depends on the sound provider

Since the synthesized vocals are learned from the provider of the vocals, it means that the performance of the synthesized vocals is somewhat limited by the strength of the provider of the vocals. Especially in the extreme high and low registers, if the strength of the provider does not show these pitches consistently in the song samples, it is difficult for the synthesized voice to reproduce them.

Note: This FAQ is still under constant revision; it does not represent official opinion.



What products does Synthesizer V have?

The Synthesizer V product line consists of two parts: the editor and the voice database.

Synthesizer V Studio Basic and Synthesizer V Lite Voice Database are free products for creators to try out before purchase but are not available for commercial use.

Synthesizer V Studio Pro is sold directly by Dreamtonics. The vocal database is sold by the respective producers.

You can find Synthesizer V's current voice database lineup at https://dreamtonics.com/synthesizerv/ in the Meet Our Voice section.

Specifically, Mai is the Japanese female voice database available to Synthesizer V Studio Pro users for free. No additional fees are required.

How much do Synthesizer V products cost?

Synthesizer V Studio Pro costs $89.99, and the voice database from Dreamtonics costs $79.

Prices for third-party voice databases vary and are determined by the publisher. Prices range from $73 to $119. The popular price for the English voice database is $89.99.

You can also choose to purchase the editor and voice database as a package, starting at $151.20.

What is the Synthesizer V product licensing and renewal policy?

The Synthesizer V Studio Pro editor software and song database are licensed for use on one device only. The software has a built-in deactivation feature that allows you to safely migrate your license to a new device.

Both the Synthesizer V Editor and the Song Database grant the user lifetime rights, as well as the right to receive upgrades within major releases free of charge, without any subscription options.

When the editor features are updated, the song database will also receive feature updates. However, the right to update the third-party singing database version is always ultimately owned by the corresponding producer, and the producer has the right to decide whether to participate in any function update of Synthesizer V.

At this time, Synthesizer V Studio has not yet undergone a major release, so the strategy for major upgrades from 1.x to 2.x is not yet disclosed.

Where can I buy Synthesizer V products?

You can find the specific lineup of song databases from each company in the Meet Our Voice section of https://dreamtonics.com/synthesizerv/.

For the Synthesizer V editor software and the Dreamtonics vocal database: https://store.dreamtonics.com/

For the Eclipsed Sounds voice database: https://www.eclipsedsounds.com/

For Audiologie voice database: https://audiologie.us/

For Animen voice database: https://www.anicute.com/

For AH-Soft voice database: https://www.ah-soft.com/product/download.html#synth-v

For the VOICEMITH voice database: http://www.voicemith.com/6084-2

For Quadimension voice databese: https://quadimension.taobao.com/

What is the commercial use policy for Synthesizer V products?

The vast majority of Synthesizer V AI song databases grant the user full commercial use rights, which means that after purchase the creator is free to use the Synthesizer V songs profitably without having to apply for a commercial use license and pay any further fees. The exceptions are a few voice databases (Stardust Infinity, Eleanor Forte, Xia Yu Yao), which have their own commercial use restrictions or procedures.

Please note that for voice database products that have both song and character designs, the song license and the character license are independent of each other. You usually have the commercial use rights for the voice, but not the commercial use rights for the character image.

For example, with the purchase of SOLARIA, you are free to use SOLARIA's vocals in a commercial setting or for commercial purposes or commercial release without paying additional fees or obtaining additional licenses, but you are not free to use SOLARIA's character images unless specifically licensed to do so.

What is the Synthesizer V product use attribution policy?

It depends on the situation. However, in general, there are two options, using the product name for the song byline or no byline. You may not use a name other than the product name for song attribution.

Please note that this restriction is for vocals and not songs. In other words, you are free to use your real name, pseudonym, stage name, or creative group name to attribute the artist of the song, but you may not use a name other than the product name to describe the singer of the song, such as Sing by, Feat., Featuring by, etc.

You are free to fill in any name you want for entries with only one artist field, but in the case of lyrics or song details, the above policy still applies to the attribution of voices/vocals.

Not attributing a vocalist or voice is always an option, so you are free to describe the song work as your own creation without leaving a place for the Synthesizer V artist.

Also, you do not have to state or disclose that the work was written using Synthesizer V. Again, this is always an optional option.

Who is Dreamtonics?

Dreamtonics is the development company for Synthesizer V, providing the Synthesizer V Studio Pro editor software and private label voice database to creators worldwide. Dreamtonics also acts as a technology provider for third-party companies that produce Synthesizer V song databases.

Dreamtonics is based in Tokyo, Japan. You can contact them for commercial and business purposes at [email protected].

Quite a FAQ! Can you enter lyrics with the free version? I'm using Eleanor lite, and if I use anything except laa laa laa, there's no sound. One more question..is there a shortcut key for starting playback from the begining..Using the mouse to rewind is tedious..thanks.

YaniDee said:

Can you enter lyrics with the free version?

Click to expand...

It should be can. You might need to check these videos:

YaniDee said:

is there a shortcut key for starting playback from the begining..

Click to expand...

You can set any shortcut you like in the setting.

FIXING PRONUNCIATION IN SYNTHV

These suggestions are aimed at getting SynthV to pronounce a word the way you want when the default pronunciation doesn't work. This can be necessary when working with cross-lingual synthesis.

SPELL THE WORD AS IT SOUNDS

Artfully respelling the word may give the pronunciation you want.

A common example is "the", which has multiple pronunciations - /dh ah/ and /dh iy/. By default, SynthV selects schwa /ax/, which is a softer version of /ah/.

One simple solution is to respell it as "thee", so SynthV selects /dh iy/.


SPELL THE WORD PHONETICALLY

If you have the "display phonetics" option turned on, the phonetic spelling of a word is displayed above the word. If you select a note and then double-click on the phonetic spelling, you can type into the phonetic box.

The selected note displays the phonemes /f ow/ above the note:



Double-clicking on /f ow/ above the note allows direct editing of the phonemes:



Phonemes must be separated by spaces. Here's a helpful reference of the Arpabet phonemes, which is what SynthV uses for English: CMU Pronouncing Dictionary

If you mis-type any phonemes, SynthV will simply ignore that phoneme.

You can also type phonemes directly into the note text by putting a period into the first character of the text, such as .dh iy




SELECT AN ALTERNATE TAKE OF THE PHONEME

You can make the Note Properties panel visible, either via the button:



or from the menu:



or the hotkey Ctrl+B

When you select a note, the properties can then be edited in the panel. If you scroll to the bottom of the panel, you can see the phonemes associated with the note:



There is a button labeled "DFLT" (short for "default") next to each phoneme. Each phoneme has up to three "takes" in the database to select from. Clicking the phoneme will toggle alternate takes of that phoneme.

Selecting a different version of a phoneme that precedes or follows it may also alter the phoneme.


ADJUST THE PHONEME DURATION AND STRENGTH

There is a Strength and Duration slider for each phoneme in the Note Properties panel:



See the prior section for how to open the Note Properties panel.


USE ALTERNATE PHOENEMES

Quite often in speaking English, a "related" phoneme may be replace the "correct" phoneme.

Many English phoeneme have "voiced" and "unvoiced" versions. This is reflected in the Arapabet spelling of the phoneme:

For example, here are some common replacements where the unvoiced phoneme is replaced with the voiced phoneme:

/s p/ → /s b/
/s t/ → /s d/
/s k/ → /s g/

In some cases, selecting a phoneme with a similar articulation he helpful. For example, /ch/ has a similar articulation to /y/, which means that "you" is sometimes pronounced as /ch ew/ or /jh ew/.

Similarly, vowels are often swapped out for more neutral vowel sounds such as /aa/ and /eh/.

This can be combined with altering the phoneme Duration and Strength for good results.

You can also combine vowels phonetically. For example, imagine you wanted to make the /eh/ portion of the diphthong /ey/ in the word "say" longer. One option would be to use the phonemes /s eh ey/, and adjusting the duration and strength of /eh/ and /ey/.

You can also use the closure phoneme (see next section) to replace a very soft stop consonant, such as /t/ or /p/.


FORCE A CLOSURE

Sometimes, SynthV simply has the wrong sound in the database. Because the prior phoneme affects the pronunication of a phoneme, you can try inserting a closure into the syllable before the problematic phoneme.

A closure can be written in English by using an apostrophe, such as ta'co. You can write them phonetically with the /cl/ phoneme, as in /t aa cl k ow/.

In most voices, a closure is treated as a stop similar to what precedes a /t/, but for some voices (such as SOLARIA) it may trigger a glottal fry. So results may vary.

Using a closure to force an alternate version of a phoneme - that is, one preceded by a stop - is typically most useful when trying to resolve pronunciation issues that come about with cross-lingual synthesis.

This is gold!

Just one little remark: surely the default phonemes for ‘the’ should be /DH AH/?

Licinus said:

This is gold!

Just one little remark: surely the default phonemes for ‘the’ should be /DH AH/?

Click to expand...

Yep! I'll correct that. Although SynthV chooses /dh ax/, which is what I intended to write. The /ax/ is more relaxed than /ah/.

The irritating thing is, I've been working with these phonemes for freaking ever, and I still get them wrong.

Thanks for writing all this down. Thanks for all the help.

Billy

First post in the forum, hello everybody.
I'm a fan of vocal emulators, I use them very often in my productions: Emvoice, Synth V, Vocaloid, Plogue, Loquendo, commodore 64, Acapela voices, etc.. I've been told that in this forum there are people who know a lot about emulation and, from what I've read so far, it's absolutely true. Thank you so much for this interesting insight

I purchased the full version of Synthesizer V back in May. It sounds amazing but I found its interface inscrutable. Be prepared to spend a LOT of time learning how to wrangle it. I gave up on it.

Hey guys, do I understand correctly, that SynthV can't be used as a vst-plugin in Ableton for now?

Gedren said:

Hey guys, do I understand correctly, that SynthV can't be used as a vst-plugin in Ableton for now?

Click to expand...

As I researched, it's possible to use SynthV in Ableton, but it requires some extra actions...
My Ableton doesn't see SynthV, but works as a standalone. Some users suggest to enable 'Use VST3 Plug-In System (or Custom) Folders', but it doesn't show SynthV still, hmm...
I guess I need to show the correct folder path, but I can't figure out yet which one in my Mac...

Gedren said:

Hey guys, do I understand correctly, that SynthV can't be used as a vst-plugin in Ableton for now?

Click to expand...

I don't use Ableton, but it should be nicely possible ... check this video, which inspired me to check, if this multi-track routing is also possible in Studio One:



PS: For Studio One users ... when Synthesizer V's audio channel layout is set to Track instead of Master (check above video or the unofficial manual), the channel routing is similar as with Studio One's Impact.

Gedren said:

Some users suggest to enable 'Use VST3 Plug-In System (or Custom) Folders', but it doesn't show SynthV still, hmm...

Click to expand...

Synthesizer V should by default install in the default locations, but definitely no VST2, only VST3 and AU too, if you use a Mac.

PS: Maybe reinstall SynthV to the default location and then rescan VSTs in Ableton.

@Aymara, thank you so much for your comment and suggestions! I'll keep exploring and trying!
I was thinking that's maybe because I have Lite version... As in the video above you shared, I saw the YT author comment: "You do need the full Pro version of SynthV to get the plugin (vst)."

Gedren said:

"You do need the full Pro version of SynthV to get the plugin (vst)."

Click to expand...

Yes, that's confirmed in the unofficial manual.

I hear the metronome click sound while recording, but not during playback? (Metronome option is active)

I have read a lot of articles and comments about the license issue and there is one thing I am not convinced about. I have Synth V Pro and I bought Solaria.

I'm going to release a song with my lyrics and music on Spotify, Apple Music, etc. I won't attribute the vocalist's name on these platforms and I'll only share it under my name since it's my song. My question is;

Can I share the promotional video of this song on Instagram, YouTube, etc. with a name that is not real?

For example; Vocal: Roberta Travis (not a real person)

sweeperxx said:

I hear the metronome click sound while recording, but not during playback? (Metronome option is active)

Click to expand...

The metronome is disabled for playback by default. You can deactivate this in Settings: