Archive for January, 2009
Tags: Omnisphere update impOSCar2 videos
SonicState provides a new video (high quality) featuring the Roland V-Piano, where we can hear more distinctly the sound, expression (and customization), than in the previous first official videos. Notably it seems more realistic than Pianoteq, less muded, and notably each note has its own distinct character (as on a real piano), instead just declining the same tone.
On that point we can read a very interesting comment at musicplayer forums :
I’ve got to say, I’ve been working with physical modeling since the VL1 (1995) and did heavy amounts of programming of that beast. Over the years, I think I demoed/bought nearly every PM synth (soft or hard) that I’ve come across. My conclusion is that PM is most successful when a) the waveform is simple and unchanging and b) the tone is expected to be the same note-to-note. Examples are VAs and organ clones. As the waveform gets more complex and is expected to evolve over time (e.g. brass, woodwinds, strings, pianos), PM, in my experience fails to deliver the goods. Also, with real EPs and acoustic pianos, each note has a somewhat distinct character. When adjacent notes sounds identical, my ears hear it as something is amiss, especially in chromatic passages. There is a nauseating sameness to the tone, because the notes are generated in the same fashion.
I would be more optimistic about PM if I had seen serious evolution in the past 15 years, but I can’t say I have. I do think PianoTeq’s piano model is noteworthy. I just don’t think it sounds accurate enough. When I play against QL Pianos or Bluthner Digital One there is no comparison to my ear.
I’m of the opinion the best solution might be a hybrid approach. Digidesign’s Velvet seems to use sampling for waveform accuracy but allows much of the control you’d find in a PM EP. The Trumpet is sample-based but allows the continuous control of timbre you’d find in a PM horn—but The Trumpet sounds infinitely better than PM trumpets.
A year ago at this time I was cautiously optimistic about the ‘modeled’ EPs in the new RD700GX. After working with the 700GX for a week I came to the conclusion the EPs offered nothing usable for me. Had they been on the level of Mr Ray 73 or Lounge Lizard, I’d probably own a 700GX. I am cautiously optimistic about the V-Piano and look forward to playing one (if I can find one).
I might well be in the market for a V-Piano IF it’s an outstanding instrument. It has to top my software pianos. But I’m expecting issues. Roland does have, IMO, a good track record for continuous improvements. The Fantom and V-Synth have had numerous, free updates.
Tags: NI Acoustic Refractions soundpack
Native Instruments released ACOUSTIC REFRACTIONS, a new Kore Soundpack, based on ABSynth, FM8 and Kontakt (samples) KORE modules. The 100 included sounds are morphings from base elements created with ABSynth and FM8, and from 500 Mb original texture samples : ice instruments, spinning washing machines, traffic on the Golden Gate Bridge, carpets being ripped, underground parking garages, and rain on a car windshield. Then we are close to Omnisphere concept. The provided demos are great (notably Morning Lullaby and Train Wreck, whose bass line reminds a song from Myst 2/Riven).
It is priced 79$ (through download).
GForce presented impOSCar 2, that brings le PWM (Pulse Width Modulation, as with the D-20 and D-50) on both oscillators, additional LFOs, new modulation and routing options, a pink noise generator, an enhanced arpeggiator, a revamped interface (3 sizes available), etc. Video demos are provided, and an upgrade plan from the first version of impOSCar is planned (no pricing yet, nor release date, but the MusikMesse timeframe is suggested).
Tags: CP2JavaWS Cappuccino Objective-J Java JSON web service bridge
An alpha version of CP2JavaWS 0.3 is available here. It uses Objective-J categories to encode/decode full objects graphs (any depth, with nested collections – CPArray/List and CPDictionary/Map, and with heterogeneous collection elements) to/from JSON, allowing customization and cleaner code. The server-side manages decoding of the JSON parameters (full object’s graphs support, also including nested collections and heterogeneous collection elements).
Encoding of the return (result) graph isn’t completed however (has to be modified to add the objjClass).
Now the client sends/receives objjClass information in JSON strings, so it is independant from server language, and we can plan other implementations (PHP, etc.)
The final 0.3 version is due in a few days.
EastWest provides a 1.1.13 update to PLAY. We can read the details here :
– added impulse responses for convolution reverb
– fixed doubling of drop down menus in Pro Tools 8
– fixed crash when turning streaming on/off with certain instruments
– PLAY window can be moved in Logic 7 at lower resolutions