Excellent point Chris...and if I were entrusting this to a mission critical application, I'd be on board. I'm playing with this using a $140 rebuilt dual core 3.4ghz machine I bought on Ebay. It handled BBP with ease, but ST seems to have more going on under the hood and therefore is hungrier. It's using a Marian Trace Alpha into a Broadcast Warehouse 1 watt exciter into a few inches of wire and is monitored by an Inovonics 530 monitor. It's my playtoy & if it worked reliably on this fairly modest system, then I could have total confidence that it will be worthy of a better system to support it in a "real" application where there's something more than my own entertainment on the line. Thing is, it's not able to work at full potential on this system so I'm unable to make that call with confidence. I've tried to get someone to tell me the minimum processor speed needed to ensure that it can run ST under the harshest conditions (BBP clearly spells that out on their site) but I'm getting "you can't judge it just by the processor speed" answers which make me unwilling to invest in something better at this time.chriscollins said:I'm interested in playing with this program. I must profess something, though.
I am seeing a lot of comments on CPU load. If I were to put this in a working environment, it would be built on a new machine, running from an SSD. You can put together a killer machine for <600, so I suppose I don't understand the complaints...
I would never trust station processing to an old computer I had laying around, so I just don't see why he should be expected to code for that. I'd rather him hit the CPU harder to get me a better product. Just my opinion, though.
What are the specs of your machine in terms of processor type & speed, RAM ,etc?oldiesstation said:Really need to have a robust CPU and Ram. Solid state drive.Like you say,build a great machine,tweaked for audio,under 600.00.I'm running 9 bands and cpu is just under 45.Just ain't gonna work great on an older bloated machine.
Thanks Oldies...at least now I have an idea of what's needed if I want to give this thing the ability to rock to the max of it's potential.oldiesstation said:I5-3330 quad 6gb ram. win 7. i've seen systems by dell and hp under 600 that would work fine or just roll your own.
Yes...that does help. That's still a fairly modest machine...much appreciated.Storm905 said:Bob, my test pc for ST is only a Core2 Duo 2.93 GHz E7500 with 2GB RAM on XP with a Julia card for MPX output. Latest ST beta is rock solid on that pc @ ~55% reported peak cpu usage with most ST modules enabled.
I've run three instances of ST ok on a Win7-64 I7 3.4 GHz pc with 4GB of RAM with <45% reported CPU use. Hope that helps.
BobOnTheJob said:Excellent point Chris...and if I were entrusting this to a mission critical application, I'd be on board. I'm playing with this using a $140 rebuilt dual core 3.4ghz machine I bought on Ebay. It handled BBP with ease, but ST seems to have more going on under the hood and therefore is hungrier. It's using a Marian Trace Alpha into a Broadcast Warehouse 1 watt exciter into a few inches of wire and is monitored by an Inovonics 530 monitor. It's my playtoy & if it worked reliably on this fairly modest system, then I could have total confidence that it will be worthy of a better system to support it in a "real" application where there's something more than my own entertainment on the line. Thing is, it's not able to work at full potential on this system so I'm unable to make that call with confidence. I've tried to get someone to tell me the minimum processor speed needed to ensure that it can run ST under the harshest conditions (BBP clearly spells that out on their site) but I'm getting "you can't judge it just by the processor speed" answers which make me unwilling to invest in something better at this time.chriscollins said:I'm interested in playing with this program. I must profess something, though.
I am seeing a lot of comments on CPU load. If I were to put this in a working environment, it would be built on a new machine, running from an SSD. You can put together a killer machine for <600, so I suppose I don't understand the complaints...
I would never trust station processing to an old computer I had laying around, so I just don't see why he should be expected to code for that. I'd rather him hit the CPU harder to get me a better product. Just my opinion, though.
If you buy something new, please keep in mind that the CPU demands will probably keep increasing over time when I release new versions. The current version will run perfectly on one if it's fast enough - and better than it has for a long time once I've added multicore support to the multiband section - but I have some things planned that may require more cores. If you don't need a low latency mode, just ignore what I just wrote.Storm905 said:consider a Core2 Duo pc
Adding more latency would allow for some tricks that could improve things a bit, such as stopping the AGC level from rising when you know that some loud sound is coming soon. And I could improve the highpass filter (make it steeper with less artifacts). But except for things like this, the improvement would be very small - if I don't add any extra intelligence, the step from latency 4096 samples (85 ms) to an infinite number of samples would be equal in size to the step from 2048 (43 ms) to 4096. Basically, every reduction of the latency by half will increase artifacts and other effects by a factor 2-4. The difference between latency 512 (11 ms) and 1024 (21 ms) is quite big - in part due to a bug in my code that I'm planning to fix in the future; steps after that have far less impact. (Note that you need to add ASIO latency to all these numbers.)chriscollins said:A question for you. How much better could you improve the sound, if you allowed 3-5 seconds for latency? Would there be any true benefit to the extra processing time?
hvz said:Adding more latency would allow for some tricks that could improve things a bit, such as stopping the AGC level from rising when you know that some loud sound is coming soon. And I could improve the highpass filter (make it steeper with less artifacts). But except for things like this, the improvement would be very small - if I don't add any extra intelligence, the step from latency 4096 samples (85 ms) to an infinite number of samples would be equal in size to the step from 2048 (43 ms) to 4096. Basically, every reduction of the latency by half will increase artifacts and other effects by a factor 2-4. The difference between latency 512 (11 ms) and 1024 (21 ms) is quite big - in part due to a bug in my code that I'm planning to fix in the future; steps after that have far less impact. (Note that you need to add ASIO latency to all these numbers.)chriscollins said:A question for you. How much better could you improve the sound, if you allowed 3-5 seconds for latency? Would there be any true benefit to the extra processing time?