iMac 2GHz Intel core 2 duo OSX 10.5.8
Macbook 2GHz Intel core 2 duo OSX 10.5.8
Pro Tools LE 8
In course of that I could hear some 10+ crackles, each one with a spike on the Options / Perf page, Audio Load panel. BAD.
So I went to Options / Devices and changed the Sample Buffer from 512 samples (11.6 ms) to 576 samples (13.1 ms) and changed the ASIO Buffer to match.
A reboot of Pianoteq to make the rematch "take", then a replay of "Zuber04XP.MID" got ZERO crackles (and spikes) from beginning to end. GOOD.
To me this indicates a direction, at least when listening to Disklavier XP files. After all these files are several times larger than the equivalent MIDIs, yet played in the same time. The CPU Load simply must rise. For other cases, like when playing, my previous Sample Buffer size (established I forget when) was 384 samples (8.7 ms). That's less than 1 / 100 of a second, a delay I doubt I can hear or react to. And I would never in a fit play so aggressively as these e-piano competition types, even if I could, so where's my expectable huge CPU Load?
Thus I've got good hopes of managing crackle in future, for all situations likely to bug me.
]]>I remamber when I was a member of Beta tester team, that Niclas posted a very CPU hungry piano midi file, from a famous piano piece. He asked everyone to test it, and if pianoteq could hold it with no cracks, and uisng a a latency least as short as 10ms, pianoteq could hold everything, cause the piano piece was the most intense of all classic pieces in terms of polyphony.
I vaguely remember the Grieg piano concerto mentioned in an old Pianoteq FAQ as an example of high polyphony.
]]>I don't remamber if it was a Rachmaninov intense piece or a Liszt piece. But it probably was a Liszt.
I supose would be nice to get everyone again and test their systems using this ultra CPU hungry midi.
I uplouded the Liszt midi, since I believe it's the must CPU hungry:
CPU usage 30-60% depending on used effects, length of IR sample, and polyphony.
]]>I had no problem with ptq 3 on this machine, but convolution and a much more detailed model take their toll on cpu usage, even after extremely efficient optimizations by Julien. My Tascam US122's buffer doesn't go below 256 samples (5.8 ms), so I would suggest using that value also. I have another card, and at 192 samples I get crackles. Latency is inaudible for me at that value.
]]>I am using a dual core that is seven years old. Occasionally I would get some crackling, and the in the audio load graph (Options/Perf), I had some red lines.
A few months ago, I completely cleaned the HDD, re-formatted, and re-installed Win XP Home. What a difference - my Performance Index went from about 23 or so to 40.
This solution is too often over-looked because of the work involved. Truth is, I didn't do it myself - I took it to a computer shop, and paid them to do it. I'm still using Win XP Home, but it's become a screaming machine again that never stumbles on anything. Internet access speed is increased too.
Glenn
It uses ASIO for the Terratec Phase 22/28 (same as my other Audio software uses)
The ASIO Control Panel of the Terratec Audio device shows the same 192 sample buffers.
]]>I'll plead in my defence that the previous incarnation of PTQ had NEEDED me to drive the amp full belt, different beast altogether. In sum, I now can position Pianoteq's volume slider sensibly, so that a punch into yellow's a warning and into red's to be avoided, then turn the amp's volume knob to suit (to avoid crackle, that is).
Which leaves me with overquiet performance from my amp still - but that's not a problem, it's some shopping.
]]>