I noticed that in Sonar 8.5 the mixdown audio files with the same settings doesn't null each other. (The diffrences are around -40dBFS)
I've tried Nebula Renderer, Cubase SX 5, and Studio One 1.5 with the same settings, and each rendering null with other renderings in the same DAW. (Haven't tried null tests across different DAW)
Test environment: Win7 64bit My audio interface Echo Mona was set to 4096ms ASIO latency. Sonar 8.5 (both 32bit and 64bit tested) worked in 44KHz, 24bit signal path. DSPBUFFER of nebula reverb is 8192. Rendered in 64bit engine and 24bit engine (all shows the same result) I use a drum loop for test, always play from the start of first beat.
To duplicate my results: (You could load whatever program in nebula reverb because the latency should be fixed, I used init program to save loading time) 1. Insert one nebula reverb and render 5~6 times. The files nulls just like other plugin. 2. Insert 2 nebula reverb and render 5~6 times. The files doesn't null and the difference is around -40dBFS) And the playback cut first beat slightly 3. Insert 3~4 nebula reverb and hit play, the first beat is damaged seriously. Of course the rendered files didn't null. Use the same settings in Cubase SX5 or Studio One 1.5, when you hit play, the signal is delayed around 3-4 seconds and start play. So I guess that Sonar8.5 has problem with plugins with larger latency.
But then I tried to use a chainer plugin like Acon Digital Media EffectChainer, and load 4 instances of nebula reverb within it, hit play, Sonar played just fine. The rendering nulled each other clearly. However the EffectChainer only works in 32bit, loss of bit depth.
And I thought I might duplicate the symptom with linear EQ which is notorious for large latency. But I loaded 8 instances of PLParEQ (latency should by around 8x1024ms or less), the playback in Sonar sounds right and the renderings nulled with each other. If the problem is in Sonar 8 alone, then this kind of rendering should fail like nebula reverb.
I'm not sure where the problem lies. I would love to see other Sonar users try this and confirm the problem. Any help would be grateful. And wish the developers knows the answer. I want to load more nebula than one instances on a track.
I did more experiments to test the problem: In sonar8.5, insert in one drum track 1) 4 init NR (Nebula Reverb): null 2) 4 AlexB 73eQ with settings: null 3) 4 AlexB MTC IL1 st: difference around -46dBFS
And when using Nebula Renderer to test 3), some files shows difference around -47dBFS too. But two of them completely nulls.
So there might be some randomness in console program. (Correct me if I am wrong)
Sonar 8 tends to play urgently, so it cuts the starting beat. But backward a few samples, you got first beat and quick playback reaction while other DAW might wait for the compenstation time completely.