|
 Kyma Forum
 Kyma Support
 Sample Rates
|
| next newest topic | next oldest topic |
| Author | Topic: Sample Rates |
|
Denis Goekdag Member |
posted 17 October 2009 08:33
 
I'm kind of scatching my head here how sample rate is handled in Kyma. I do most of my work in 192khz so there's plenty of potential for hassle considering the sample rate of my Capybara is limited to 96khz. What's weirder is that behaviour is not consistent --- if I knew that SR > 96khz was just not possible *at all*, fine, I'd downsample anything I want to Kyma-tize, but in some spots 192khz seems fine.... So far I've observed the following: - I can preview from disk at any sample rate --- does this invoke any realtime SRC, perhaps by quicktime? - I can load 192khz data into the sample editor but it will play back at Capy SR (so an octave lower when running at 96k) - I seem to be getting tons of weirdness if I try to RE-analyse 192k files ? - I *can* do a spectrum analysis on a 192k file and the resulting spectrum will play back at the correct pitch (at least if I remember correctly, it was a looong night's work....) Would there be any way to render to disk at 192khz even if the Capy's I/O maxes out at 96k (well, 100 actually)? Cheers, IP: Logged |
|
cristian_vogel Member |
posted 20 October 2009 12:06
out of curiosity.. Why do you work at 192K? IP: Logged |
|
Denis Goekdag Member |
posted 21 October 2009 07:19
Mainly because it gives me an extra octave of possible downward transposition to play with. Also a lot of plug-ins perform considerably better at 192k (read: the often poorly designed anti-aliasing filters can be placed outside of the audible range and thus have little to no impact on the audible band). IP: Logged |
|
cristian_vogel Member |
posted 21 October 2009 11:50
interesting, thanks! IP: Logged |
|
SeanFlannery Member |
posted 22 October 2009 08:31
I've been working mostly at 32k on my 8 processor capy mainly to get as much performance out of the processors. I have experimented with higher sample rates and really noticed the difference. The higher the better as far as synthesis goes. AUD $0.02 IP: Logged |
|
shennelly Member |
posted 01 November 2009 06:55
quote: Dan Lavry would disagree: http://www.lavryengineering.com/documents/Sampling_Theory.pdf IP: Logged |
|
Denis Goekdag Member |
posted 02 November 2009 11:46
No, he doesn't. Dan Lavry simply points out that, similar to the time/frq trade-off in FFT, or even the uncertainty principle found in quantum physics, the higher the sampling rate ( = time resolution ) the lower the statistical precision of the amplitude measurement will be. As the additional information gained by high sampling rates is beyond the audible range AND most consumer delivery formats are incapable of reproducing this information anyway, Mr Lavry concludes that the net loss in amplitude measurement accuracy is not worth the net gains. This is a relevant point IF your intent is to sample a signal and reproduce it without any alteration, as in the final step of a mastering chain. Now the kind of stuff we DSP nerds do *greatly* benefits from the added resolution. IP: Logged |
All times are CT (US) | next newest topic | next oldest topic |
  |
|
Contact Us | Symbolic Sound Home
This forum is provided solely for the support and edification of the customers of Symbolic Sound Corporation.
Ultimate Bulletin Board 5.45c