soma, when i changed asio4all buffer size from xt2 setup - advanced, xt2 automatically adjusted to the same buffer size
john
idle XT2 cpu comparison (everyone please)
-
- KVRAF
- 5350 posts since 8 Aug, 2003 from Berlin Germany
True. i noticed that, but I don't think it works the other way around... Shouldn't that field be read-only in XT?1barEnuff wrote:soma, when i changed asio4all buffer size from xt2 setup - advanced, xt2 automatically adjusted to the same buffer size
john
-
- KVRist
- 344 posts since 14 Oct, 2005
im not sure exactly what you mean by read only, but if you/we adjust the samples in xt2 - advanced - soundcard its handy if xt2 adjusts automatically, i dont reckon it works the other way round with any soundcard but i can't say for certain
edit: i just checked my other computer with maudio audiophile and i can't change the buffer size with xt2 open. it says cant change buffer while audio card is being used by an app (or similar) so i changed xt2's buffer size and it snapped back to what it was (512)
so i had to close xt2 to change it
john
edit: i just checked my other computer with maudio audiophile and i can't change the buffer size with xt2 open. it says cant change buffer while audio card is being used by an app (or similar) so i changed xt2's buffer size and it snapped back to what it was (512)
so i had to close xt2 to change it
john
Last edited by 1barEnuff on Thu Jun 14, 2007 2:11 pm, edited 1 time in total.
-
- KVRAF
- 4143 posts since 7 Sep, 2001 from Melbourne, Australia
OK - an update on my testing with my desktop.
If I use ASIO4ALL on the on-board sound my CPU idles between 0-4%, but it's quite variable going from value to value.
If I use ASIO4ALL on my Edirol UA-101 I'm getting pretty much 0% most of the time with occasional fluctuations to 1%.
So there's a significant difference if you look at one result as a multiple of the other, but I'm not getting up to the 9-10% that people have been reporting.
I never talked about buffering and samplerate before, but I'm using 44.1 with 512 buffers. I don't have - use hardware buffers selected in ASIO4ALL either.
I just checked by changing the samplerate to 48k for the internal on-board soundcard and the CPU dropped - it's now oscillating between 0-2% with an occasional 3% thrown in for good measure.
I have heard that increasing samplerate does actually reduce CPU consumption though - or maybe I'm getting that confused with something else.
Regards
Caleb
If I use ASIO4ALL on the on-board sound my CPU idles between 0-4%, but it's quite variable going from value to value.
If I use ASIO4ALL on my Edirol UA-101 I'm getting pretty much 0% most of the time with occasional fluctuations to 1%.
So there's a significant difference if you look at one result as a multiple of the other, but I'm not getting up to the 9-10% that people have been reporting.
I never talked about buffering and samplerate before, but I'm using 44.1 with 512 buffers. I don't have - use hardware buffers selected in ASIO4ALL either.
I just checked by changing the samplerate to 48k for the internal on-board soundcard and the CPU dropped - it's now oscillating between 0-2% with an occasional 3% thrown in for good measure.
I have heard that increasing samplerate does actually reduce CPU consumption though - or maybe I'm getting that confused with something else.
Regards
Caleb
Happiness is the hidden behind the obvious.
-
- KVRAF
- 25008 posts since 12 Jul, 2003 from West Caprazumia
oh, something which just came to my mind:
asio4all 2.7 performed significantly worse on my system than 2.6 so I went back...
2.6 can still be downloaded from the official site and also there's a 2.8beta...
...might be worth trying for those with problems who use asio4all - Caleb?
asio4all 2.7 performed significantly worse on my system than 2.6 so I went back...
2.6 can still be downloaded from the official site and also there's a 2.8beta...
...might be worth trying for those with problems who use asio4all - Caleb?
-
- KVRist
- 344 posts since 14 Oct, 2005
now im rubbing my eyes in disbelief. on my p3 933 with a synth comp and a 1 bar loop playing my cpu is 0 - 2 all i did was change the buffer to 128,untick alsways resample to 44.1<>48, yes it crackles a tiny bit sometimes and i need use hardware buffer.
absolutely crazy!now im going to check again to make sure i'm not cracking up.
john
absolutely crazy!now im going to check again to make sure i'm not cracking up.
john
-
- KVRist
- 344 posts since 14 Oct, 2005
ok, 6 synth comps, with 6 midi parts with random notes dropped in 2 bars in length, cpu 30 - 48, while playing. i haven't tweaked the synths i just dropped them in. sweet!
john
john
-
- Banned
- 1373 posts since 5 May, 2007 from Finland
steady 2% here too
pentiumD 2.8ghz 1gb
terratec asio driver - buffer size doesnt matter
pentiumD 2.8ghz 1gb
terratec asio driver - buffer size doesnt matter
-
- KVRAF
- 5350 posts since 8 Aug, 2003 from Berlin Germany
My cpu usage went down as I lowered the buffer until I got to 240, then it started to go up again. The lowest idle cpu I got on my simtel integrated sound card was 2-4% with 48k sampling and 240 buffer, no special options or fixes enabled in ASIO4All.
-
- KVRist
- 344 posts since 14 Oct, 2005
mines set to 44.1 16bit and i just changed the buffer to 256 with only use hardware buffers ticked and its reading 0-2 sometimes flickers 3
edit: just in case its any use latency compensation in and out 32 samples, buffer offset 10ms and its the crappiest forte media 801 soundcard
john
edit: just in case its any use latency compensation in and out 32 samples, buffer offset 10ms and its the crappiest forte media 801 soundcard
john
-
- KVRAF
- 1601 posts since 24 Jun, 2004 from Australia
energyXT 1.4
- 0-2% CPU (mostly 0%), ~12mb RAM (idle, blank project, no sequencer)
- 0-3% CPU (mostly 1-2%), ~13mb RAM (idle, inserted sequencer but didn't open)
energyXT 2 (8th June)
- 0-2% CPU, ~36mb RAM (idle, blank project, no sequencer)
- 0-3% CPU, ~38mb RAM (idle, inserted sequencer but didn't open)
RAM usage is far higher (though when minimised and maximised, XT2 only uses ~11mb RAM with seq.).
Specs:
- Intel motherboard (not sure which chipset, because I never remember it)
- Intel Core 2 Duo T7200
- 2GB RAM
- Geforce 7900GS operating at 1680x1050 (come to think of it, XT2 was maximised and energyXT wasn't during this)
- Tascam US-144 USB audio interface (using their ASIO drivers, at 96kHz, buffer of .. erm .. uh .. 256 or 512, can't remember, because Tascam's driver setting thing is stupid and changes depending on sample rate)
It seems the same to me, but I realise this is a powerful system (even though a laptop). I'm pretty sure it's similar with my desktop system too though.
I could install ASIO4ALL and test them with my onboard sound, but I had nothing but trouble from ASIO4ALL on the laptop and energyXT didn't like it (starting energyXT up there would be a low level "squealing" sound), and it had huge delay and operated like general ass. In case anyone is interested, the onboard sound is a SigmaTel high definition sound chip. Onboard sound, unless Creative, are typically weak for music production things. Doesn't mean they don't work, and some are ok, but I don't think I could ever trust them to work properly, I will (I think) always use a seperate dedicated audio device. If anyone who is having problems here is using ASIO4ALL with onboard sound, it might be worth borrowing a device from a friend, or going to a store and testing a USB/Firewire device, and seeing if that makes a difference.
I also have my onboard disabled by default. Unfortunately, this is not in the CMOS setup, but in Windows, since laptops don't have much in the way of options in the CMOS setup.
- 0-2% CPU (mostly 0%), ~12mb RAM (idle, blank project, no sequencer)
- 0-3% CPU (mostly 1-2%), ~13mb RAM (idle, inserted sequencer but didn't open)
energyXT 2 (8th June)
- 0-2% CPU, ~36mb RAM (idle, blank project, no sequencer)
- 0-3% CPU, ~38mb RAM (idle, inserted sequencer but didn't open)
RAM usage is far higher (though when minimised and maximised, XT2 only uses ~11mb RAM with seq.).
Specs:
- Intel motherboard (not sure which chipset, because I never remember it)
- Intel Core 2 Duo T7200
- 2GB RAM
- Geforce 7900GS operating at 1680x1050 (come to think of it, XT2 was maximised and energyXT wasn't during this)
- Tascam US-144 USB audio interface (using their ASIO drivers, at 96kHz, buffer of .. erm .. uh .. 256 or 512, can't remember, because Tascam's driver setting thing is stupid and changes depending on sample rate)
It seems the same to me, but I realise this is a powerful system (even though a laptop). I'm pretty sure it's similar with my desktop system too though.
I could install ASIO4ALL and test them with my onboard sound, but I had nothing but trouble from ASIO4ALL on the laptop and energyXT didn't like it (starting energyXT up there would be a low level "squealing" sound), and it had huge delay and operated like general ass. In case anyone is interested, the onboard sound is a SigmaTel high definition sound chip. Onboard sound, unless Creative, are typically weak for music production things. Doesn't mean they don't work, and some are ok, but I don't think I could ever trust them to work properly, I will (I think) always use a seperate dedicated audio device. If anyone who is having problems here is using ASIO4ALL with onboard sound, it might be worth borrowing a device from a friend, or going to a store and testing a USB/Firewire device, and seeing if that makes a difference.
I also have my onboard disabled by default. Unfortunately, this is not in the CMOS setup, but in Windows, since laptops don't have much in the way of options in the CMOS setup.
-
- KVRist
- 122 posts since 20 Mar, 2005 from paris
i also noticed that changing sample rate affects the amount of RAM which falls from 30 mb to 6,78 after i have changed samplerate from 44.1 to 48.soma wrote:My cpu usage went down as I lowered the buffer until I got to 240, then it started to go up again. The lowest idle cpu I got on my simtel integrated sound card was 2-4% with 48k sampling and 240 buffer
getting back to 44.1 doesn't restore figures but keeps to 5% CPU, RAM 6,7 mb !!!
i tested ASIO2KS driver(locked buffer 192, 48 khz):
CPU 0% same RAM amount(30 mb>>6,7)when tweaking)
(Inspiron 6400,Intel T7200,RAM 2gb, ASIO4ALL 2.7, Sigmatel HD)
-
- KVRist
- 122 posts since 20 Mar, 2005 from paris
i had the same issue with ASIO4ALL until i downgraded from 2.8b2 to 2.7, chect this outdruid wrote: I could install ASIO4ALL and test them with my onboard sound, but I had nothing but trouble from ASIO4ALL on the laptop and energyXT didn't like it (starting energyXT up there would be a low level "squealing" sound), and it had huge delay and operated like general ass. In case anyone is interested, the onboard sound is a SigmaTel high definition sound chip. Onboard sound, unless Creative, are typically weak for music production things. Doesn't mean they don't work, and some are ok, but I don't think I could ever trust them to work properly, I will (I think) always use a seperate dedicated audio device. If anyone who is having problems here is using ASIO4ALL with onboard sound, it might be worth borrowing a device from a friend, or going to a store and testing a USB/Firewire device, and seeing if that makes a difference.
-
- KVRAF
- 25008 posts since 12 Jul, 2003 from West Caprazumia
DAAHOOD wrote:i also noticed that changing sample rate affects the amount of RAM which falls from 30 mb to 6,78 after i have changed samplerate from 44.1 to 48.soma wrote:My cpu usage went down as I lowered the buffer until I got to 240, then it started to go up again. The lowest idle cpu I got on my simtel integrated sound card was 2-4% with 48k sampling and 240 buffer
getting back to 44.1 doesn't restore figures but keeps to 5% CPU, RAM 6,7 mb !!!
i tested ASIO2KS driver(locked buffer 192, 48 khz):
CPU 0% same RAM amount(30 mb>>6,7)when tweaking)
(Inspiron 6400,Intel T7200,RAM 2gb, ASIO4ALL 2.7, Sigmatel HD)
that's a trap here - as soon as you load the system it will most likely go higher up than with bigger buffers - the results you are experiencing are probably not an indication of real performance - of course you can't test that with an idle application...