deadbyte

Hi, I'm deadbyte! Your local English vocal-synth enthusiast and Synthesizer V stanHe/Him
28
Gay🌈
ENG - Native
日本語 - BasicAccounts are SFW, though not averse to spicier content or references to adult content (13+)Manager & Voicer of UTAUs "ATLAS", "LongWei", and "db-svs".Co-Manager of VIRTUAL VOCALIST「eseph」

• Ko-fi
• Email
• Resources
• Commissions

• UTAU/OpenUtau

• Synthesizer V Studio

• Alternative Mugshots & Portraits

Includes:
- Kotonoha Akane & Aoi
----- Mug: Removed background/Transparent- Eleanor Forte(Lite)
----- Mug & Portrait: Matches AI DB----- Mug & Portrait: Separate Mug & Portrait based on anniversary artwork is also available----- Mug: Added "Std" icon- Eleanor Forte AI
----- Mug: Removed background/Transparent----- Mug: Added "AI" icon- GENBU
----- Mug & Portrait: Uses his updated artwork ( From AHS Store )- Saki
----- Mug: Added "Std" icon- Saki AI
----- Mug: Added "AI" icon- Yamine Renri
----- Mug: Removed Background/Transparent

• Eleanor Forte Logo

Fanmade recreation of Eleanor Forte's logo from her boxart.

• Reclists

• OpenUtau Basic Dict

A basic arpasing.yaml file which allows for [ax], [zh], and [dx] to be used in phonetic hints.This does not contain additional word entries.

• OpenUtau Arpasing 0.2.0 Dict

• Terms of Service

This dictionary is now considered redundant. It is now recommended to use Cadlaxa's [EN ARPA+] Phonemizer instead.

An arpasing.yaml file made to significantly improve to usage of 0.2.0 Arpasing voicebanks in OpenUtau.More details regarding the dictionary can be found in it's "readme.txt" file.Click here for a demonstration

• OpenUtau Arpasing AgO Dict

• Terms of Service

This dictionary is now considered redundant. It is now recommended to use Cadlaxa's [EN ARPA+] Phonemizer instead.

An arpasing.yaml file made to improve to usage of AgO Arpasing voicebanks in OpenUtau.More details regarding the dictionary can be found in it's "readme.txt" file.Hiragana input is now supported as well.

Japanese

• Eve-Style Reclist (Modified)

A Japanese CVVC reclist. It is a modified version of the "Eve" reclist used for 波音リツ(Namine Ritsu) Eve. It corrects several issues and inconsistencies with the original list, as well as adding English translated instructions, and an extra optional list for Vowel-Vowel samples.

English

• db 3-Mora ARPAsing Basic list

A "Basic" Phoneme-based 3-mora ARPAsing reclist made of 526 samples. It covers all necessary diphones for English.For multi-pitch libraries, an "additional pitches" reclist is available that cuts out samples that do not change significantly on other pitches. This list only contains 345 samples.In addition, contrast to other English recording lists, it is meant to be friendly for non-rhotic accents, as it is meant to have enough duplicate "r" diphones to compensate for most contexts.Other traits include:- "Index" and "No Index" variants- Various oremo-comment files in different formats ( SAMPA, VCCV, VOCALOID, etc. )- An add-on "dx" reclist- IPA transcripts to help experienced singers unfamiliar with SVS.

• db 3-Mora ARPAsing Liverpool list

A phoneme-based ARPAsing reclist designed for the Liverpool accent. This list features a number of additional phonemes made to accommodate Liverpool English.The full list is made of 697 samples. Without CC's, it is 411 samples.Originally made for Toge Uroko -ARPAsing-Other traits include:- Base oto.ini files included- Phoneme chart for new phonemes- Dictionary files to use voicebanks with OpenUtau- An "OtoSplitter" Python script to help quickly and easily make voicebanks compatible with ARPAsing Assistant.

Mandarin Chinese

• VCCV Chinese

**COMING SOON**

This reclist is currently a work-in-progress. If you would like to help record a voicebank for it, please get in touch with me.A custom condensed format for Mandarin Chinese. It is made to be an in-between to popular formats such as CV Chinese and CVVChinese. It is designed to be compact, while still allowing for smooth and high-quality vocals. This format is made particularly so that bigger projects involving Chinese singers ( such as multi-expression or voicebanks with many pitch-layers ) are possible without sacrificing synthesis quality comparable to CVVChinese, nor having the large file sizes known to that format, or causing an overflow error in UTAU.There are currently two finished reclists. A 3-syllable list with 216 samples, and a 6-mora^[1.] list with 104 samples.^{[1.] This list currently still experimental, and requires testing}

Terms of Service

This is a Terms of Service for the Arpasing 0.2.0 and Arpasing AgO Dictionaries made by deadbyte, intended for use in the open-source software "OpenUtau" created and managed by stAkira.The Arpasing 0.2.0 and Arpasing AgO Dictionaries are based on Arpasing syntax as made by Kanru Hua, which is based on Arpabet, developed by the Advanced Research Projects Agency.The Arpasing AgO recording script is developed and managed by 水音 (みなと).These dictionaries are heavily edited versions of the Open-Source "CMU Pronouncing Dictionary" developed by Carnegie Mellon University.You may use this dictionary at your own discretion. Given that the rules of this ToS are followed, permission is not required.Attribution is not required when using the dictionary. You are free to publish works using it without attribution to use of the dictionary, or to the creator(deadbyte).Editing the arpasing.yaml file is allowed.Redistributing the unedited file is prohibited. However, if it is edited with relevant changes or improvements, redistribution is allowed. Otherwise, please link to the original file.If the link it was being distributed at is no longer available, redistribution of the unedited file is allowed.You may distribute this dictionary with your own UTAU/OpenUtau VB's, plugins, or other UTAU-related software. This includes commercial software as well. However, selling the dictionary itself is strictly prohibited. Please ensure that in the case of being bundled with commercial software, the dictionary does not add to the total price of the product. In addition, if you do distribute software with the dictionary, including voicebanks, attribution to deadbyte and Carnegie Mellon University are required.Usage of the dictionary in ways that are illegal, discriminatory, or aimed at harming others is strictly prohibitedUsage of the dictionary for pedophiliac expression and acts that encourage it are strictly prohibited.Claiming the unedited file as your own is strictly prohibited.deadbyte is not responsible for any damages caused by this file. Use at your own discretion.This Terms of Service may be subject to change without prior notice.For questions or inquiries, please contact <[email protected]>

Prices

UTAU & DeepVocal

If the voicebank format is not listed, please contact me at [email protected] for a customized quoteIf you would like to begin a commission, I can be reached by email([email protected]), via DM's on Twitter(@dead_byt), or Discord(deadbyte#3530)***Every price listed is in USD- Please note that being asked to work on a base configuration/oto.ini file that was automatically generated ( such as with moresampler or setparam ) with little to no further adjustment, will still constitute an entire pitch, and is not applicable to any sort of discount or price reduction -Terms of Service can be found at the bottom of the page. Please read before commissioning.

VocalSharp

For the time being, commissions for VocalSharp VB's will only available for formats listed on the price chart above. The list will be expanded over time.If you are looking to convert an already existing UTAU VB to VocalSharp, please ensure that the VB in question has a COMPLETED oto.ini file. Otherwise, it will be treated as an original VB.If you would like to begin a commission, I can be reached by email([email protected]), via DM's on Twitter(@dead_byt), or Discord(deadbyte#3530)***Every price listed is in USD

Terms of Service

Within this transaction, I, the service provider, will create a configuration file to make a set of specified voice samples compatible with the UTAU, OpenUTAU, DeepVocal, or VocalSharp software.Once I have completed the configuration file, I will create a short demo for proof of completion. From that point, you will receive the file upon payment.Upon completion of the transaction, you are free to use it freely for both commercial and non-commercial purposes. Permission for commercial usage is not required upon completion of the transaction. You may use it for other voicebanks.Unless otherwise agreed upon, I will require that the client have me credited as the creator of the configuration, either within the contents of the voicebank itself, the voicebank's distribution source, or other publicly available venue associated with the voicebank.You may not:- Use it for illegal activity or hate speech.- Sell it individually for profit without the accompanying voice library.- Take credit for the creation of the completed configuration file. Any unedited variant of the configuration file is not to be claimed as another persons work.- Use it for the purposes of being used in, with, or as a peripheral of NFTs, cryptocurrency, or other blockchain related/adjacent endeavors.If you use the configuration file as a base configuration for another voicebank, and adjust it to allow that voicebank to function, you may claim it as your own. If you edit the configuration for it's intended voicebank, you may credit yourself as an additional creator or co-contributor.In the event there are incorrect or broken recordings, I will compile a list of samples needed to finish the bank, and ask that they be re-recorded.If you need instruction or advice for a revised set of recordings, I am willing to assist or advise for no extra charge.You generally will not be charged extra for extraneous recordings ( such as trilled "r"'s or "l"-related samples in voicebanks which do not require them ), unless there is an excessive amount of them. If you are unsure, please inquire beforehand.Generation of .frq files is free, however, if you wish to have frq files refined, there will be an additional charge for significant adjustment of .frq files, including repair or refinement of .frq files. This includes frequency files of all resamplers. The amount charged will vary depending on the scope of the project.Clipping and/or shortening of UTAU samples will be provided for free upon request if specified beforehand.If in order for the voicebank to function properly, I have to modify a large amount of the provided audio samples ( particularly due to incorrect pronunciation, or technical issues such as distorted, corrupted, or otherwise problematic audio ), there may be an additional charge proportional to the amount of extra work which needed to be done.In the cases of ARPAsing, or similar formats which utilize contextual duplicates, I will personally decide on the maximum number of contextual duplicates ( usually 3 additional duplicates or less ) if the client does not specify beforehand.I reserve the right to cancel or decline the commission for any reason.By accepting the completed product, you agree to the terms and conditions listed here.----------For any questions or concerns, please contact via DM on BlueSky, Discord, or email at [email protected]

Commissions

• oto.ini ( UTAU, DeepVocal, VocalSharp )

• Labeling ( DiffSinger, ENUNU, etc. )

• Training ( DiffSinger )

• vocal midi ( .ust, .ustx, .svp, .vsqx, etc. )

Prices

Labeling

Label commissions are currently operating on a trial basis. Mistakes and hiccups may be possible, prices may fluctuate, and I will only be taking one commission at a time.If you would like to begin a commission, I can be reached by email([email protected]), via DM's on BlueSky(@deadbyte.bsky.social), or Discord(deadbyte#3530)***Every price listed is in USD- Please note that being asked to work with base label files that were automatically generated ( such as with labelmaker ) with little to no further adjustment, is not applicable to any sort of discount or price reduction -Keep in mind that these commissions are primarily for Singing Voice Synthesis (SVS), and not speech synthesis.Due to the amount of time it takes to complete labels for the average dataset, UTAU/oto.ini commissions will take priority. I can give a projection for how long I believe it will take, but it may end up taking longer depending on how busy my other commissions are, and how complex the dataset is.Terms of Service can be found at the bottom of the page. Please read before commissioning.

Rate - $5 per minute of data labeled.Willing to do any language. However, assistance may be required with languages I am not familiar with.Best with: English
Experienced with: Japanese, Korean
Confident with: Spanish, Mandarin Chinese, FrenchAdvising for DiffSinger/NNSVS recording and model creation is available free of charge.

Terms of Service

Within this transaction, I, the service provider, will create a set of label file(s) compliant with the client's choice of language and phoneme format, for the purposes of being trained for a neural-network based synth such as DiffSinger or NNSVS.Payment will only be requested upon completion, at which point, you will receive the file upon payment.Upon completion of the transaction, you are free to use it freely for both commercial and non-commercial purposes. Permission for commercial usage is not required upon completion of the transaction. You may use it for other ethical datasets.You may absolutely not request labels be made for a dataset(s) sourced from a vocalist(s) who has not explicitly given consent for their voice to be used for this purpose. Ethical Commissions Only. No datasets made of celebrities, popular fictional characters, public figures, etc., without the explicit consent of the appropriate parties.Unless otherwise agreed upon, I will require that the client have me credited as the creator of the labels, either within the contents of the dataset/model itself, the dataset/model's distribution source, or other publicly available venue associated with the finalized model or dataset.You may not:- Use it for illegal activity or hate speech.- Sell it individually for profit without the accompanying dataset or model.- Take credit for the creation of the completed configuration file. Any unedited variant of the configuration file is not to be claimed as another persons work.- Use it for the purposes of being used in, with, or as a peripheral of NFTs, cryptocurrency, or other blockchain related/adjacent endeavors.If you use the configuration file as a base configuration for another dataset, and adjust it to allow that dataset to be utilized, you may claim it as your own. If you edit the configuration for it's intended dataset, you may credit yourself as an additional creator or co-contributor.In the event there are incorrect or broken recordings, I will compile a list of samples needed to finish the bank, and ask that they be re-recorded.If you need instruction or advice for a revised set of recordings, I am willing to assist or advise for no extra charge.If in order for the dataset to function properly, I have to modify a large amount of the provided audio samples ( particularly due to incorrect pronunciation, technical issues such as distorted, corrupted, or otherwise problematic audio ), there may be an additional charge proportional to the amount of extra work which needed to be done.I reserve the right to cancel or decline the commission for any reason.By accepting the completed product, you agree to the terms and conditions listed here.----------For any questions or concerns, please contact via DM on BlueSky, Discord, or email at [email protected]

Prices

Training

Training commissions are currently operating on a trial basis. Mistakes and hiccups may be possible, prices may fluctuate, and I will only be taking one commission at a time.If you would like to begin a commission, I can be reached by email([email protected]), via DM's on BlueSky(@deadbyte.bsky.social), or Discord(deadbyte#3530)***Every price listed is in USDKeep in mind that training is mostly offered here as a courtesy to those who have had labeling commissions. Training is cloud-based, and not local, so if you require local training, it's best to look elsewhere.Due to training being cloud-based, and reliant on Google Colab, the fidelity of required Python libraries, and the status of available tools, issues are liable to occur and may delay the process of training significantly. Please keep this in mind before commissioning.Before commissioning, please ensure that your dataset has, at minimum, 15 minutes of valid singing data, to ensure the stability of the final model. The recommended minimum amount of data for an adequate model is 30 minutes or more. A high-quality model typically requires 2+ hours of data.If you are primarily using UTAU sample data as your dataset, it is strongly advised to utilize additional speakers to fill in gaps in the dataset ( See "If you want additional speakers for parallel training" section below ).Terms of Service can be found at the bottom of the page. Please read before commissioning.

Rate - $20 per dataset trained, acoustic + variance both included ( Max of 160000 steps each ).Before commissioning, please specify the following:

Terms of Service

Within this transaction, I, the service provider, will create a set of label file(s) compliant with the client's choice of language and phoneme format, for the purposes of being trained for a neural-network based synth such as DiffSinger or NNSVS.Payment will only be requested upon completion, at which point, you will receive the file upon payment.Upon completion of the transaction, you are free to use it freely for both commercial and non-commercial purposes. Permission for commercial usage is not required upon completion of the transaction. You may use it for other ethical datasets.You may absolutely not training for a dataset(s) sourced from a vocalist(s) who has not explicitly given consent for their voice to be used for this purpose. Ethical Commissions Only. No datasets made of celebrities, popular fictional characters, public figures, etc., without the explicit consent of the appropriate parties.Unless otherwise agreed upon, I will require that the client have me credited for the training of the model, either within the contents of the dataset/model itself, the dataset/model's distribution source, or other publicly available venue associated with the finalized model or dataset.You may not:- Use it for illegal activity or hate speech.- Sell it individually for profit without the accompanying dataset or model.- Take credit for the creation of the completed configuration file. Any unedited variant of the configuration file is not to be claimed as another persons work.- Use it for the purposes of being used in, with, or as a peripheral of NFTs, cryptocurrency, or other blockchain related/adjacent endeavors.In the event there are incorrect or broken recordings, I will compile a list of samples needed to finish the bank, and ask that they be re-recorded.If you need instruction or advice for a revised set of recordings, I am willing to assist or advise for no extra charge.If in order for the dataset to function properly, I have to modify a large amount of the provided audio samples ( particularly due to incorrect pronunciation, technical issues such as distorted, corrupted, or otherwise problematic audio ), there may be an additional charge proportional to the amount of extra work which needed to be done.I reserve the right to cancel or decline the commission for any reason.By accepting the completed product, you agree to the terms and conditions listed here.----------For any questions or concerns, please contact via DM on BlueSky, Discord, or email at [email protected]

Prices

Vocal Midi Making

Midi commissions are currently operating on a trial basis. Mistakes and hiccups may be possible, prices may fluctuate, and I will only be taking one commission at a time.If you would like to begin a commission, I can be reached by email([email protected]), via DM's on BlueSky(@deadbyte.bsky.social), or Discord(deadbyte#3530)***Every price listed is in USDPrice will vary depending on the complexity of the song, and what specific parts of the song you want transcribed. The prices displayed below are a loose metric for what to expect.Coms will likely have some mild tuning unless otherwise specified.Terms of Service can be found at the bottom of the page. Please read before commissioning.

Willing to create for most vocal synth programs.Experienced with: VOCALOID1-6, UTAU, Synthesizer V Studio, OpenUtau, DeepVocal, VocalSharp, VoiSona

Will not create for: Ace Studio

Willing to do any language. However, assistance may be required with languages I am not familiar with.Experienced with: English, Japanese, Korean, Spanish, Mandarin Chinese, French, German

Examples



Japanese:

Black Out ft. UTAU 騒音ロコ

Memeriser ft. UTAU LongWei & Pruno Faustino

FIRST ft. UTAU Peiton

flag ft. UTAU LongWei



English:

Into You ft. Synthesizer V Asterian

Hitomania (English ver.) ft. UTAU eseph_COLORS

Kaleidoscopic Core (Vocal Cover) ft. Synthesizer V Saki & Kevin

Feeling of Falling ft. UTAU eseph_COLORS



French:

Intergalactia (French ver.) ft. Kenzo Hoshine

Terms of Service

Within this transaction, I, the service provider, will transcribe a midi-style file ( .ust, .vsq, .svp, or similar equivalents ) for a song of the client's choice, for the purposes of making a vocal-synth sing it.Payment will only be requested upon completion, at which point, you will receive the file upon payment.Upon completion of the transaction, you are free to use it freely for non-commercial purposes. Permission for commercial usage is not required upon completion of the transaction if you own the song; however, you will have to ask for consent from the appropriate parties if you do not own the song.Unless otherwise agreed upon, I will require that the client have me credited for the creation of the file, either within the contents of the file itself, the file's distribution source, or other publicly available venue associated with the file in question.You may not:- Use it for illegal activity or hate speech.- Sell it individually for profit without explicit permission from the song's owner.- Take credit for the creation of the completed file. Any unedited variant of the configuration file is not to be claimed as another persons work.- Use it for the purposes of being used in, with, or as a peripheral of NFTs, cryptocurrency, or other blockchain related/adjacent endeavors.I reserve the right to cancel or decline the commission for any reason.By accepting the completed product, you agree to the terms and conditions listed here.----------For any questions or concerns, please contact via DM on BlueSky, Discord, or email at [email protected]