• 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.
I
Reaction score
609

Joined

Profile posts Latest activity Postings Usernames About

  • Have any of you used an AI image generator? I'm wanting to use one to create images for an upcoming music video, but I have no experience. All I know is that they exist. Any recommendations?
    Here's a free and newly updated VST2 saturation plugin for Windows users:

    I've given it a bit of trial run and I do like the results. Subtle but effective.
    Out of curiosity, I rendered a short test file using both Eleanor AI and Solaria, then mixed them together using Solaria as the mid channel and Eleanor AI as the side channel. Here is the result:

    Note that Eleanor AI is also turned down -3 dB.
    Hmm... I may have found a quick fix for SynthV's painful ~2 kHz issue AND the excess sibilance. Recently I've been getting rid of the ~2 kHz problem by rendering the pitched and aspiration sounds as two separate files, then filtering out the unwanted ~2 kHz pitched sounds in a sound editor, and finally merging the pitched and edited aspiration files manually. This worked to get rid of the painful ~2 kHz, but not the sibiliance, which I got rid of with regular de-essing. However, it dawned on me that perhaps the excess sibiliance was caused by an issue that was the mirror of the ~2 kHz pitched info leaking into the aspiration info. Perhaps some aspiration info was leaking into the pitched info? So I tried rendering the the pitched and aspiration files as usual, then phase inverting one of them before manually merging the two. The end result seems to be an improvement! But I need more free time to test this further.
    Does anyone else get intermittent clicks and pops with SynthV AI voicebanks?
    • Wow
    Reactions: Leon
    Cluemily
    Cluemily
    I got this for the first time the other day whilst using SOLARIA, but every time I've used SV since I've never been able to recreate it. But if you're still regularly experiencing it, it might be worth posting about it on the Product Feedback section of the SynthV forum. It might be something they can look into.
    I
    inactive
    I have to admit that I avoid their forum because it's such a hot mess. To steal an expression, it has a "last one in the shop" look.
    Cluemily
    Cluemily
    Ah, that's fair. I'd suggest an email, but their website itself suggests the forums be used for feedback. Their e-mail is for recruitment and business. :kaito_lili:
    If it's only a SOLARIA issue, they're currently working on her vocal modes, so maybe when the update comes through it may change something? Or it can be something reported to her devs?
    As of today Reaper officially supports per-FX instance and per-FX chain oversampling. This allows users to oversample plugins that don't support oversampling. Of course, results may vary...
    I just finished spending a good chunk of time trying to tame the harshness in Solaria's higher notes. Mildly successful...

    Is it just me, or have all of SynthV's voicebanks sounded harsh starting at around C5? And by all, I'm also including every voicebank released for the original pre-Studio version of SynthV.
    I
    inactive
    A ha! Good to see corroboration on the frequency area. This might be the problem, with the frequencies in that range being doubled up unnecessarily thereby creating nasty resonances.

    (I don't where to submit bug reports...)
    I
    inactive
    I know I'm supposed to be doing school work, but an answer suddenly occurred to me:

    Export the pitched data and aspiration data as separate files. (I use 32 bit floating point so I don't have to worry about dithering and whatnot.) Then mix them in your DAW (both at unity gain), but with a narrow notch filter placed at around 2.3 kHz on the aspiration file. Mixing this filtered aspiration data with the untouched pitched data will create a more pleasant listening experience!
    lIlI
    lIlI
    Awesome, I'll try this!
    Uh oh. I've run into a Solaria pronunciation bug.

    My lyrics say "out of sight, out of mind", but she sings "out of sit, out of mind".

    The phonemes are correct, and if I make a new note elsewhere to test "sight" it sounds fine. But in the context of my music it always comes out as "sit". Doesn't seem to be a note pitch or note length issue, just a context issue. Hmm...

    And it's the last line of the song!
    • Wow
    Reactions: Ronny777
    Ronny777
    Ronny777
    I don't have SynthV to Test around in so I'm not certain about how the phonetic symbols react with one another so, out of curiosity, do you still have the same sound problem if you keep (s ay t) in "sight" but change the (aw) in "out". It looks like the (aw) is messing with the timing, forcing the (ay) to pronounce faster than it should. If I'm looking at the waveform correctly it looks like the (t) is programmed to annunciate a complete stop between vowels.. It seems to also be doing the same thing with the words "out of". If its the phonemes conflicting with each other (dx) instead of (t) might be the only work around.

    I hope my rambling made sense.
    I
    inactive
    I tried what you suggested and changed the [aw] in "out" to a couple of different vowel sounds, but it made no difference.
    Nokone Miku
    Nokone Miku
    Just wanted to say that I often run into the same sort of thing with Miku English. I usually fix it by adding a [j] phoneme into the word. So "sight" would be [s aI j t] and "pain" might be [ph eI j n]. In extreme cases I end up needing to add an extra vowel like [s aI][- j V t] or [p eI][- j I n]. When the note is short it still sounds like one syllable.
    Solaria seems to work well with my music. But some of the pronunciation is from a type of USA accent that is foreign to me.
    Prism
    Prism
    Emma is from Nashville, Tennessee so that makes sense. She would probably have a slight country twang because it's basically the Hollywood of country music. Probably the closest accent would be early Taylor swift or Carrie underwood.
    I
    inactive
    Most were fixable dictionary issues, such as "decorative" being parsed as a three-syllable word instead of a four-syllable word. (I guess in Nashville they don't pronounce the "o"? I mean, it's schwa-y even in my corner of the globe, but it's still there.)

    However "block" sounds like something between "black" and "block", which I couldn't fix.
    So I walked to the University campus this morning to get my booster shot. I pretty much had the entire trip to myself, partly owing to people staying indoors to celebrate the omicron season, and partly owing to an air temperature of -29°C with a wind chill of -36°C. But on my way back, just as I was leaving the campus property, I spotted some guy walking on the sidewalk that spanned perpendicular to my direction of travel. He was wearing a grey hoodie, his hands in the pockets and his head snugly in the hood. In addition to the hoodie he also wore sneakers, socks, and shorts. And by shorts I mean shorts. As in the sort of shorts one would wear in mid-summer. Certainly a lot less than my two pairs of long johns and thick winter pants. What's extra crazy is he didn't seem to be in a hurry, and was instead moseying along quite nonchalantly.
    Has anyone managed to locate a manual for Synthesizer V Studio Pro? The help option just takes me to a YouTube link, GUI translations, and scripting information.
    Shiver me timbers! I just received my Eleanor Forte AI download code and serial number. Expect a new video in a day or two! (Unfortunately I doubt anyone on this site will enjoy the upcoming video because I'm heading down a different path with my self-authored English lyrics.)
    lIlI
    lIlI
    Congrats! Looks like they shadowdropped her.
    Älfa Dröttning
    Älfa Dröttning
    I got her code too! So far I love her vb! I have some time off coming up so I'm going to work on a cover with her that I started with her lite vb. I'm looking forward to your video!
    morrysillusion
    morrysillusion
    same! looking forward to installing her asap!!
    For the one other person on this site that uses Reaper, a cool new "delta solo" feature has been added. Of course, some plugins already have a delta solo function, but many don't, making this addition super-duper handy.
    • Wow
    Reactions: WyndReed
    WyndReed
    WyndReed
    Woah, this is so cool! Thank you so much for sharing!!
    I
    inactive
    I knew there was at least one other Reaper user on this site!

    Because Reaper is updated so often it's easy to miss when major new features are added. But at least the frequent updates are teensy-weensy to download and quicky-wicky to install.
    Recently I've been messing around with mic IRs to help colour some synthesized sounds that I use. (Modo Drum to be specific.) But as an experiment, I decided to try a mic IR on Eleanor Forte Lite and got some good results.

    Below is a short song sung twice, the first time without a mic IR (100% dry) and the second time with a mic IR (100% wet). They have been level matched so you won't be fooled by a change in volume.

    No EQ (apart from the IR itself), no compression, and no effects. The difference is subtle, but the mic IR does clear up some mud. Pay particular attention to the held vowel in "song."

    There's a bunch of free mic IRs available here:
    Microphone Impulse Response Project

    The above collection is all "vintage" mics, complete with a smattering of broken mics, which could be interesting for special effects.

    Unfortunately, the mic IR that I used in the above demonstration is not free. I bought it from PastToTheFuture. It's an AKG C12 IR using a cardioid pattern placed at 2 inches. You can get it as a single mic package, or in a collection package with a bunch of IRs.

    Note that IRs only capture a static frequency response of a microphone, rather than a real-world dynamic response. Additionally, all voicebanks have either been recorded with a microphone (concatonative) or trained with a microphone (AI), so they've already been coloured by a microphone. By using a mic IR, you're essentially recolouring a precoloured signal.
    So, yeah. Here's a song from 1968 by an all-girl "group sounds" band called Pinky Chicks. The track is supposed to be a joke, but the vocal effects are so silly that I spent more time laughing at the song rather than laughing with the song.

    MillyAqualine
    MillyAqualine
    Sounds like a cartoon song mixed with a yé-yé-styled (no wonder since the date of release) tune =o

    It's strange because it's kinda like Cendrillon from Disney's adaptation with Gusgus and the other mice x'D Shame though because they have a nice voice

    Kagami Kawaiine was here
  • Loading…
  • Loading…
  • Loading…
  • Loading…