![]() "Chipsounds 2.0 is a complete rewrite and uses nothing from 1.0. What can we expect to see in chipsounds 2.0? What were the most common v1 user requests/feedback that you plan to address in v2? It took quite a bit of vintage hardware reverse engineering and testing, but that was the fun part!" On top of that, we had to create five different singing engines (Formant/LPC10/LPC12/Mozer and Klatt), and get the rights to vintage data files to feed them. "We knew how to make typical virtual synthesizers, but a singing synthesizer was a totally different beast there were lots of text/MIDI synchronisation issues that were tricky to solve in a way that doesn't burden the user's workflow. ![]() That being said, we had to be extra careful to preserve its original tone as much as possible. Some sources, like the infamous IBM704, who sang Daisy Bell in 1961/1963, were taken from vintage vinyl recordings and did not contain all the transitional English phonemes required - more than a thousand! So there was a lot of formant DSP voodoo done to fill in the gaps. "While each voice created its own unique set of challenges, I'd say have to say that the biggest challenge was creating voices from incomplete phonetic data. "Some sources, like the infamous IBM704, who sang Daisy Bell in 1961/1963, were taken from vintage vinyl recordings" Which was the most challenging to reproduce in software, and why? Your latest product, chipspeech, recreates seven classics 80s voice synthesis chips. It had ring mod, PWM, hard-sync, not mentioning the Hubbard-style 'wavetables' (changing waveforms rapidly) that were soon abused by most composers, and it played samples through the infamous $D418 volume trick. "The SID was designed by MOS to be sold to synthesiser companies (so it's fitting that the engineer who made it quickly left Commodore/MOS to start Ensoniq!), and was leaps and bounds ahead of all of the other stuff I'd heard at that point. Some could say the Amiga beat it with its four-channel, vari-speed, 8-bit PCM playback… but for me, it's only ever been outclassed by modern PCs running VSTis. I ended up spending four summers straight at his place in the mid 80s! In my opinion, without it's unique sound, the C64 would not have had the level of success and longevity that it did. ![]() "My old friend was the only guy in the neighbourhood with a C64. I'll let you know what else I find."Without it's unique sound, the C64 would not have had the level of success and longevity that it did" I also like the idea of being able to run other VST instruments like Vienna Instruments alongside Giga.īidule seems like a capable, flexible host for VST instruments and it's affordable at $75. However, I've heard of people getting better RAM and polyphony performance from GVI than GS3, all things being equal, so I'm going to keep experimenting with this system. So it apprears that Bidule is taking advantage of dual processors where GS3 is not. I know another guy who is running a similar test with a Core2Duo processor and reports much better performance from GVI and Bidule than GS3. (Bidule maxes at 512 samples of latency, but the system seems stable at these settings.) When it gets to 80% I hear a constant crackling unless I set the RME card to 1024 samples of latency. The CPU meter in Bidule shows 45% when idle, and reaches 70-80% when playing 80-100 voices. So if the card is set to 256, you need to set Bidule to 256 or else you get clocking problems.Įven after matching the latency settings, I still got clocking problems when running more than 100 voices through the system, something that doesn't cause problems when I'm running GS3. I'm running a P4/2.8GHz system with 1.5GB of RAM (I know I need more.) I had a lot of clocking problems until I discovered that you have to set the latency in Bidule to match the latency on your card. So far my performance has not been as good as GS3.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |