• We're currently having issues with our e-mail system. Anything requiring e-mail validation (2FA, forgotten passwords, etc.) requires to be changed manually at the moment. Please reach out via the Contact Us form if you require any assistance.
IO+

IO+

Resonance47
Messages
244
Media
58
Resources
11
Reaction score
2,044
Medals
0
Joined

Profile posts Latest activity Postings Media Albums Resources About

  • IO+
    Did youtube analytics is broken or something? My videos only last for 4 minutes but YT show me average view duration over 7 minutes...
    • Haha
    Reactions: AddictiveCUL (Add)
    View previous comments…
    AddictiveCUL (Add)
    AddictiveCUL (Add)
    @IO+ I guess it doesn't, I was just joking XD But your vids surely are amazing.
    • Haha
    Reactions: IO+
    mobius017
    mobius017
    Probably someone's doing DJ-style "scratching"...a lot...while watching them. :LOL:

    Seriously, I guess they could maybe be watching, pausing, and resuming? Depends how they calculate average view time.
    IO+
    IO+
    I'm not sure about looping/replay function on youtube count as views duration since my videos have very low views count so i guess it might be a bot or some random people forget to turn off replay function.
    IO+
    I see people talk about YouTube audio compression and this topic is pretty much everywhere start from reddit, twit, audio engineer commu, musician groups to audiofoolphile groups so i made a blind test here. Only "one" of these choices is extracted from Original CD and it got a RAW lossless WAV files and others choice are MP3 that i personally ripped it off from youtube myself.

    However. . . can you tell a different?

    In order to make it fair, if you want to play and pick one of these 3 choice, please write a comments about the reason why you pick it. (Example: I pick A because....) or you can just say i'm just guessing, it's perfectly fine too.

    This is not about making fool of someone , you can call it a scientific experiment.

    A

    B

    C

    Music: 40mP
    Song title: evidence (bonus track)

    Then again, it not about shaming, there is absolutely nothing wrong with the choice you choose.
    • Like
    Reactions: WyndReed, MillyAqualine and MagicalMiku
    View previous comments…
    IO+
    IO+
    It's time to reveal the answer, but first I have to say I'm sorry, it was an accident. The host I uploaded to (vocaroo) convert it to mp3 so it might not accurate, but doesn't mean this blind test are useless. I have to upload on vocaroo for the sake of convenience.

    Every track has been turn down -6.0dB (Track C turn down by -5.5dB)
    LUFS are very close to each other, loudness error only 0.12 LUFS (barely any difference) except Track C. . . i was purposely make it louder.
    MP3 files is Varible BitRate

    I do blind test myself so i wanted to share my choice too.

    I have such a hard time to choose between A and C and in the end choose C because i feel this track is sound more "in the face" more energy and because of this i can hear miku voice better (even reverb tails) but it's quite difficult to focus what i should listen to because i do like what i hear so my bias start to consume me.

    The true answer is:
    Track A: RAW lossless WAV files with 100% sample to sample accuracy compare to CD i own.
    Track B: Ripped off from this channel (uploaded around 12 yrs ago).
    Track C: Is the same from track B but it's a trap with +0.5dB louder than other tracks.

    I choose this song to do a blind test because it's easier to detect any difference and you got all usual piano, kick, snare, bass, lead guitar and miku. If i pick something like rock, metal, EDM it's practically impossible. . .
    IO+
    IO+
    @MagicalMiku
    especially now youtube "likes" to slow down on purpose or going down on playback quality because of the whole "ad block thing"
    Click to expand...
    They're so busy for this one :ROFL:
    but from my experience, usually warmer sound can be a result of sound compression (for some songs it can give a warm effect, it really depends and is very subjective)
    Click to expand...
    You are correct, MP3 with lower bit rate (anything below 256kbps) will apply a High Pass Filter to filter anything above 17kHz because that area consume a lot of data space so it got cut out. However, it's depend on type of filtering, some old school conversion method use shallow filter and already start cutting from 15kHz but most standard filter usually apply brickwall like filter to completely remove anything above 16kHz.

    @pico Track C is a trap that i created but... i'm the one who got caught in it. Original song both from youtube and CD is get push to absolute limit (like 5.7 LUFS) so 0.5db that i boost is enough to make it sound crush, flat.

    @WyndReed I don't know how you detect any detail with such a tiny speaker... it's impossible to me.

    In my opinion, Audio compression is necessary (May god have mercy if we don't compress it. . . imagine you have to download +400GB just for one game) the method of audio conversion is so much advance compare to the algorithm from 20yrs ago. A tiny detail will be loss that for sure but it's good enough not to be a bottle neck, or hampering the actual content.
    MagicalMiku
    MagicalMiku
    ohhh loved everything, also the trap! hehe :ryuuto_lili:
    it's interesting that track B has been uploaded 12 years ago and quality is good, I guess playback 12 years ago was very limited (480p probably was the standard, I don't remember..) and the source file must've been CD.
    also, I totally agree that compression for audio, video and images is needed. I like to rip my CDs both in FLAC (lossless) and AAC (compressed), and my blu-ray discs rip all are compressed in h264 and convert the DTS HD audio track into AAC 2.0. for most important blu-ray concerts, I buy some 50 or 100GB blu-ray XL and make a backup copy:lapis_smile_lili:
    IO+
    I do a rough mix covers for a comparison between UTAU and Synth V basic (AI lite), i think it's fair because both voicebank are free.
    I make it fair and equal as much as i can. leveling the play-field by make it equal in terms of perceived loudness and amounts of attention i spend for both.


    Teto is quite troblesome to mix due to how very dynamic she is (and her vibrato), this remind me when i used luka. Gain automation is mandatory.
    Much less input while tuning, i like it but manual input is recommended.


    Yamine Renri is easier to mix, this kinda unfair because i work with Renri and UTAU more than Teto and Synth V, still... i don't even need a compressor for this one, it's take me like 20 minutes from start to export. Although she need manual tuning just like any traditional voicebanks.
    IO+
    This year was a very sad new year.
    IO+
    I might as well do a little survey here, just to have a rough idea.

    Do peoples still use an old version of vocaloid editor(v3-v4) on daily basis?
    IO+

    Miku Dark can sound sweet but more mellow. You can find a vsq file in the video description as usual.
    This time I put more effort into using miku, I hope it will be worth it.
    IO+
    I never expected to be able to use Luka for this long. Only 2 years left before hit a decade old...

    I've been struggling a lot with luka but it's feel so rewarding when i can get it right.. it's like a dopamine.
    IO+
    Free plugins suite from Venn Audio.
    A barebone general-purpose kind of tools.

    Free Suite | Venn Audio

    Free Suite is a completely free, performant, light weight, resizeable, zero fuss set of cross-platform audio plugins. All the tools needed to fully mix a track to professional standards. No registration needed, simply download and enjoy! Free Suite is available for Windows, Mac, Mac M1 native...
    www.vennaudio.com www.vennaudio.com
    IO+
    I start to think that music is a mathematic at heart. At least they were closely related.
    • Like
    Reactions: PearlStarLight5 and Ryleigh Goldner
    pico
    pico
    I don't think it's a coincidence. I have met extremely talented mathematicians and software developers that began their lives as music majors or as members of bands. I really struggle to intuitively understand both math and music despite sincerely enjoying them both, so I've always felt they were closely related somehow for that reason.
    IO+
    IO+
    When i start to look at it a little deeper i find that music are similar to math. People around the world use 12 note system but some people usually use 7 note out of 12 for convenience and rhythmic we base on a grid line of 1/16 and pitch of sound in nature is very mathematical (ratios) like a simple whole note or half note (semitone), even when i mix stuff, i have to do some math too but that's more like a muscle memory or burning image that happen on the fly.

    There’s much more to music than time signatures, note durations, scales, sound frequencies, harmonic related and etc. I do agree but still... We build a whole structure of the song by using math and emotion.
    • Like
    Reactions: pico
    IO+
    It seems like i can use Miku the same way as i use Luka, there not much of a difference between them. If i use project files that specifically tune for Luka and then switch to Miku, i may hear a funny sounds but nothing break, i can fix it with minimal input.

    Miku pitch-correct between phonemes transition make tuning predictable with no sanity lost.
    Some voicebanks have tendency to pitch deviated(small dip) when do a fine note blending or... it stop making sound altogether.
    IO+
    A rough mix wip, i don't have Miku english so i'm force to use Luka instead. The hard part is to make percussion-ish voice i have no idea how to make it at the moment, if you have any idea please let me know.


    This song was adapted into Japanese in 1997 and was used in Initial D anime later in 1998. It's hard to miss this anime if you are a car guy.
    • Like
    Reactions: MillyAqualine, WyndReed, DigitalVintageTunes and 2 others
    View previous comments…
    mobius017
    mobius017
    Maybe clip off the end of appropriate syllables by automating DYN down to 0?
    AddictiveCUL (Add)
    AddictiveCUL (Add)
    I'm not very good at tunning yet, but every time I tried to make vocaloid speak more than sing I did it by varying the pitch more radically. You can also do like Mobius if you want a stutter sound. Tho I did twice a kinda of stuttered sound. The first one I just made bunchs of the same notes consecutively and the second time I made it by cuting part of the extended vowel and made little "tracks" inside of the main track consecutively so it sounds like, you know, it's cutting.
    IO+
    IO+
    I think i found the way.
    I find that keeping the pitch a constant straight-line can make it sound like percussion so i use Pitch Snap Mode and then use [-] Symbol to keep the sustain note going until the end of the note. It may not working sometime if note are so close at each other, so i have to clip dynamics off as mobius suggested and it's actually work as i wanted.

    I wanted to do a faithful cover of the original song but English part make thing difficult like try to run uphill on a barrel.
  • Loading…
  • Loading…
  • Loading…
  • Loading…
  • Loading…
  • Loading…
Top Bottom