|⚠||This is a controversial topic.
Any changes should be notified to ensure that edits meet VOCALOID Wiki's policies and guidelines. Discussions for controversial topics are here and more subjects categorized here.
POCALOID and illegal modifications
- Main article: POCALOID
One of the most common concerns for any software is pirated versions. It has been proven over the course of time that there is little that can be done to prevent software from being cracked by software crackers and the ensuing impact on legitimate copies of the software.
The main concerns about illegal software versions are normally as followed:
- Harm to user computers - illegal download links can at times contain malware or viruses. When a user downloads pirated versions of software from unknown sources, they are at risk of being infected.
- Scams - There are cases where voicebanks sold over the Internet are advertised as legitimate, but are actually cracked copies in disguise. Buyers of VOCALOID need to check their software sources when purchasing the software to ensure they are purchasing from a legal vendor.
- Lack of technical support - there is no technical support offered for illegal versions of software such as VOCALOID. Users of POCALOID and other such software would have to seek technical support themselves, which can be difficult to acquire.
- Legality - the act of piracy is considered a form of theft; therefore, people can be held accountable for their actions. As such, illegal versions of VOCALOID™ do not comply with the end user agreements of VOCALOID™.
- Publishing - work published that utilizes illegal versions of software can lead to shame within the music industry and destroy producer reputation. In the case of the Japanese, shunning is a common reaction; however, this doesn't always work.
- Support for further development - it is only financially feasible to update software if it sells. Piracy often takes away potential sales from the software, having a negative effect on its future development. The studios are concerned about illegal software because of the loss of potential customers; there will always be users who are unwilling to ever purchase the software as long as illegal "free" versions exists. The impact of the loss of sales can potentially hurt Vocaloid development; the sales expectancy for vocal synthesizers is often as low as 1,000 units. If the software fails to sell, then there is less of a chance that the studio or company behind the software will produce more. For some studios this is an important issue, as the financial support they give to producing software is based on the success of previous releases, and less sales result in less money for the next vocal.
- Capabilities - sometimes the capabilities of pirated versions are different from those of legitimate versions of the software. Sometimes they contain unsupported functions, altered interfaces or "bug fixes" that can interfere with software updates or functionality. One of the most common function removals of illegal versions of VOCALOID is the security measures put in place within the software.
- Software conflict - illegal versions of a software are known to at times cause conflict with legal versions of the software.
These concerns aside, there is often little overall difference between the illegal versions of software and legal versions. However, due to the moral implications behind support of illegal Vocaloid, it is often shunned within the VOCALOID fandom.
Within the Vocaloid fandom, there is support for alternative software such as UTAU as a legitimate method for producing works, rather than use of illegal VOCALOID™ software.
During the development of Ruby, Syo spoke out about his payment and involvement of her production. He noted he had worked on Ruby for 17 months mostly alone, taking 10 hours out of his school time a week and was only offered 1/4 of the minimum wage amount, plus a bonus if she sold well. He noted the fact that others will steal his work bummed him out.
Cross-synthesis ("XSY") was first introduced for this engine, but was limited to a set few VOCALOID packages. XSY between two VOCALOIDs or two languages were not programmed into the feature's capabilities.
The software can be, however, modified to allow XSY to occur between any VOCALOID of any language, creating a "Frankenstein-esque" vocal with a mixture of both voicebanks' traits. The resulting vocal can often sound very different in comparison to the products used to make it, making it difficult to tell which two were used.
While a number of fans do not see XSY modding as harmful, several VOCALOIDs do not share the same licensing as each other. An example of two VOCALOIDs with different licenses are Hatsune Miku and IA. As a result, problems are caused when a producer publishes their work, as any work produced using modded XSY will not comply with the End User Agreement license of VOCALOID.
The issue with XSY modding is often overlooked and producers have been known to publish works using modded XSY on sites such as YouTube or Nico Nico Video. Not all producers see it as destructive, compared to issues related to general pirating, because XSY modding does not affect the sales of VOCALOIDs.
Most of the faults with XSY modding are caused by the mathematical equations used to render the results. When two voicebanks not designed to be used together for XSY are used, the softwares calculations cannot handle them so well. These are due to the technical differences between the vocals, such as extra phonetic data, differences in phoneme usages and set up. Whereas in the case of vocals set up to compliment and work well together (example being Megpoid V4 Native and NativeFat) the results are higher in quality. This is also why in the modified version of XSY vocals such as V3 Megpoid and Megpoid V4 are not set up to XSY with each other. No matter what, the results of mixing the V3 and V4 versions is impossible to get quality results out of due to Megpoid V4 having increased triphones amount and vocal differences.
XSY of vocals between languages (such as English x Japanese) can produce unintelligible or low quality results, as the sounds were not designed to be combined. However, it is noted that the method can produce entirely new sounds for use, for example, mixing V3 Megpoid - Native and Megpoid English can be used to make the Megpoid package recreate other languages more easily. Though it can be inconsistent, as in theory only samples with share coding will react in XSY and is unpredictable in samples that are not the same in both languages such as English "S" and Japanese "S". It will have far less issues with English "k" and Japanese "k" since they are similar enough.
- See also: English - Japanese
Due to the differences between any Vocaloid, quality is almost always lost due to this and this means the mixing of two voicebanks not intended to be used together in XSY almost never produces high enough quality results to warrant the modifying of the vocal. As confirmed, mixing two vocals can at times produce a result that is the equivalent of an entirely new voicebank and the greater the differences between the voicebanks the more different the result is than vocals set up to be used with it. Effectively the potential tones the User has access to is double by just mixing two vocals, producing two possible variations in addition to the original two voices. The main attraction to XSY modding therefore falls upon the results that XSY itself offers, making a few voices produce many results.
Regardless of the advances that XSY modding can offer, producers should be aware that works published using the altered software violates the End User Agreement.
As with any part of Vocaloid, job plug-in support gave way to unsupported plug-ins being made. For example, one of the more common plug-ins are language conversion based and many perform basic executions like transforming English phonetics to Japanese and vice versa, a practice that already occurs within the Vocaloid fanbase without such plug-ins. Many of these Job plug-ins are harmless and often are just made to make life easier for users by speeding up the process. Though in the case of phonetic conversions even with supported plug-ins such as Maika's Spanish to English plug-in are not perfect and can even be bested by vocals of that language.
As with any unsupported adaptions of Vocaloid or its functions, users should be aware of devious practices such as hidden Trojan files within the download, though this does not mean all are designed with malicious intent.
As of VOCALOID5, job plug-ins are not compatible and this issue is less notable.
Another form of illegal VOCALOID usage is the act of creating an entire new voicebank from existing ones or claiming to create a new product without creating a licensing agreement with Yamaha in the process. This latter issue is much rarer than illegal downloads as it requires the re-recording of entire vocal banks. To date this has yet to be a serious problem. All groups or parties claiming to have done so have proven to be hoaxes, either created using UTAU or XSY modding.
Vocaloids are created using a developers construction Kit or "Dev Kit". It was confirmed without the access code, the Dev Kit can only pack/unpack the script - a full working Vocaloid voicebank is not possible.
VOCALOID into UTAU
UTAU does not permit VOCALOID vocals within the software and is even stated within the agreements of both software, if not by name then by user agreement. VOCALOID is a proprietary software and VOCALOIDs are made to be used only by the VOCALOID software itself. This covers not only the samples within the voicebank itself, by everything such as the character itself and all vocal results from creating lyrics within VOCALOID.
This has not prevented voicebank creators from attempting to create voicebanks from VOCALOIDs, for example in 2010 a user by the name of "RICARDO" was known to have created a Lola one. Several other users at the time were discovered to have made voicebanks on Kagamine Len and Hatsune Miku. Due to a number of issues, RICARDO was forced to cease distribution of the voicebank, as this was not his only illegal activity.
In 2019, a user "sought" permission from Miriam Stockley to be able to make a UTAU based on her VOCALOID MIRIAM. However, Zero-G Limited stated their authority on the matter and noted how illegal the voicebank was. The justification put forward was the lack of an update to MIRIAM and despite Zero-G asking several times to only speak via their Customer Support service, much of the user's attempts to explain their actions was made public on Twitter. The user has since kept the voicebank private and the issue has been handled privately. However, the distribution of the illegal voicebank has reappeared according to a tweet made by Zero-G on July 29th, 2019. In the tweet, Zero-G stated they will be taking action to protect their IPs, and thanked everyone for notifying, and rallying behind the company. This may possibly lead to lawsuits being filed by both Zero-G and YAMAHA.
In late 2019, a user uploaded a video to YouTube of KAITO V1 imported into UTAU singing a short sample of Yura Hatsuki's song "Hameln". Despite attempts by others, they have refused to take down the video and cease their illegal actions, showing blatant disregard of the end user agreements. They and their follows went on to mock those who voice their stance against their actions. As of January 2020, the video remains up on YouTube. They also uploaded a video of MEIKO V3 imported into UTAU in which they again mocked the situation in the video's description, aside from stating that all MEIKO's V3 vocals were imported, along with Megurine Luka V4X and VY1v4. In all videos, their followers have requested some VOCALOID into UTAU videos of characters such as Kagamine Len and Utatane Piko (which encouraged another user to import him into the engine), encouraging their illegal behavior, as well as saying if companies aren't going to update VOCALOIDs, they might as well do it themselves. Their justifications are said to having ownership of the purchased VOCALOID voicebanks. It's unknown if YAMAHA and Crypton are aware of their actions.
In the case of English vocals, UTAU only is created to recreate the Japanese language. Though voicebanks are available in other language, they still work within the basic set up of the Japanese layout of samples, unable to create missing phonetics such as ending or final consonants. The non-Japanese language creation of UTAU is considered poor, even when it has clarity. It has been witnessed that many who create a voicebank of a English VOCALOID is recreate the vocal in Japanese, as was the case in both previous mentioned user cases. The English voicebanks are created for the English language and therefore have English Phonetics. Using their voicebanks for English - Japanese comes with all the issues of using a language for another, and while it is generally much easier to use a English vocal for Japanese then the other way around, there are still issues. Clarity may be achieved, but it is important to remember that clarity is not the only sign of quality.
The other issue is due to the way in which often users obtain the sample for UTAU, quality can be lost. The UTAU engine also does not sound like the VOCALOID one, thus even if the two would share the same sample, the resulting singing voice is not identical. This is usually the case for multi-medium methods as different engines can produce very different results. This has been witnessed several times, with the most known example of this being Macne Nana who has been on Reason Garageband, UTAU and Vocaloid, each producing different vocal results.
Either way, no matter what the UTAU vocal will not match how a VOCALOID sounds entirely, and looses much of the quality making a VOCALOID -> UTAU voicebank creation pointless. It is often far better to create a UTAU vocal for UTAU from scratch, simply because the vocal can be catered for the UTAU software restrictions and quirks and fix bugs themselves. A lot less control is given for a VOCALOID -> UTAU conversion and often a UTAU built vocal will end up being overall superior to a VOCALOID conversion, simply because VOCALOID is calibrated for an entirely different engine, and making a UTAU from the Vocaloid engine will make the voice less realist because it will have UTAU's and Vocaloid's engine noise at the same time. The known examples of VOCALOID providers having given their voices to both (DEX , DAINA, Macne Nana and RUBY) have overall demonstrated how voice don't sound the same depending of the software. Thus ultimately, VOCALOID -> UTAU is a pointless effort as for rendering purposes.
Though Editors may be aware of how to create UTAU voicebanks using VOCALOID voicebanks, please note we cannot supply this information. For legal reasons, users should not attempt this either.
Controversy has been caused by users breaking the VOCALOID license agreement. As explained by the End-User license agreement, a User is responsible for their own actions, and any misconduct will fall solely upon the actions of the user. For example, despite the existence of pirated versions, it is the user who chooses whether to buy the VOCALOID software legally, pirate the software, or not use it at all. If the user allows another to use their software or hardware for any of these acts, the user risks being held responsible, as they allowed permission of their sources to be used in the first place.
One of the most notorious cases of user misconduct in the overseas fandom came from a former producer, who had a questionable reputation among overseas VOCALOID fans owed to past controversies. One of these was distributing links to the cracked trial voicebanks of anon & kanon, despite the package not being released yet nor the trial having been declared free for distribution by the developers.
Despite his reputation, he wanted to aid promotion of several projects within the overseas community and had a small following of fans as a result. This allowed him to gain a position of respect and trust despite the more questionable actions he had been known for.
In January 2015, Stella was announced as an upcoming Japanese VOCALOID being produced by the team PL Tech, for the nonprofit organization Americans for the Arts. It was stated that Stella was private because the expenses needed to release her as a commercial VOCALOID were too great. She had three 'voicebanks' listed, known as "core", "sun" and "moon".
Over time suspicion grew among fans that Stella might be a hoax, with a turning point in public view occurring when a discrepancy was spotted in one of her demos, revealing the videos to be doctored. On January 19, the user finally confessed that Stella was fake, leading to large amounts of vitriol within the community. Within hours of the confession it was revealed Stella was created using XSY modding. Consequently, any work that had been produced using her was illegal. Though two of the vocals were initially revealed to be created using XSY, the third was stated to be made using a VOCALOID Dev kit. He declared that he acquired them from a group called "VIBE" and Bplats, Inc., as he had additionally claimed to have aided in the development of the XSY function. VIBE was forced to speak out, stating that user's connection to them was fabricated. The team expressed frustration at statements and worried that his lies may have put their own project in jeopardy. They confirmed they had not passed over a Dev kit to him and the third vocal was yet another XSY modded voice result.
The user's aim to sell albums containing Stella's "voice" for charity would have been a breach of the license agreement, regardless of transparency. The monetization added to the seriousness of the situation and the charitable intentions were brought into question.
The scale of trust involved between the project and its members was initially high, though three other members involved with PL Tech stated they were in the dark regarding Stella's origins. These members worked on Stella on the basis of trust and friendship with the user, causing large emotional fallout among the group when the illegitimacy of the project was revealed. The fact that he used the name of a well known charity to promote her development was considered a direct attempt to mislead individuals, as well as the forged emails provided to 'prove' her legitimacy. At the time the user was also involved with the promotion of an UTAU known as "Jewel", a tribute to Ring Suzune, which also became caught up in the controversies. The sudden news of what happened with Stella led the team behind Jewel to overhaul large parts of their project to distance themselves from this user, including renaming the character.
Furthermore, investigations by fans during and before the event revealed that the user frequented websites involved in the distribution of illegally cracked VOCALOID software, leading fans to question whether Stella's voice was formed using such voicebanks or not. Additionally, due to comments made at the time by himself on the VOCALOID wiki, his copy of Rana was also suspected of being illegal. This makes the situation even further problematic as any published work would not have complied with the End-User agreement. Even if XSY modding was allowed, he would have not been in a position to commercially release his works if they were created using pirated software.
Some of his practices were known before the event, but were not touched on until afterwards, and had no connection with the events at the time. Even though the user's actions did not lead to criminal charges against him by any legal authority, his reputation as a VOCALOID producer was tarnished. The user's actions are not an isolated case, though it highlights the potential and consequences of fraudulent practices within the fanbase.
It is also important to note that in his confession he mentioned that he was aware of breaching the license agreement and was aware of the immoral nature of his actions.