[geek]
Ok, I've just done some tests. I tried to be as scientific as possible... if I've made any errors in methodology or analysis, please let me know. Bear in mind that the tests were vey simple and designed to highlight any differences in sound quality using Rewire, Reason and SX native exporting.
Created a 1-bar kick-drum loop in Reason using only Redrum. Created same loop in SX using same sample and Battery. Rewired Reason to SX. Had to change level of Battery to get SX and Rewired kicks to the same level (accurate to 0.1 dB). Exported 24-bit files of:
A: Reason loop through Reason mixer (all faders at default)
B: Reason loop direct to sound card (no Reason mixer)
C: Rewire loop through Reason mixer (all faders at default)
D: Rewire loop direct to SX (no Reason mixer)
E: Battery loop from SX
Tests and comparisons performed using Sound Forge v6.0:
Tried to compare each file by inverting data and mixing to generate a file containing the differences between the two sounds being compared; however timing inconsistencies between Battery and Reason exports made some comparisons impossible. Accuracy of musical timing was determined by placing markers at the beginning of each kick drum and counting samples between kicks.
Results:
Test 1 : A vs B : The difference file contains the loop at -121.41 dB, ie. there is a minute difference in output level when using the mixer compared to outputting directly to the soundcard from Redrum.
Test 2 : C vs D : Identical-looking difference file as above. Mixing the two difference files together (after inverting) produced noise at -146.29dB where each kick had been, with digital silence in-between.
Conclusions:
There is some difference when using Rewire as opposed to exporting directly from Reason, although to all intents and purposes this difference is negligible and would not, in my opinion, be audible. Not even to you, J. Also the test was unable to answer the question of which is better, SX or Reason; however given the scale of the difference (-146.29dB) I reckon it's not worth worrying about!
Timing was consistent when exporting sounds generated by Reason whether or not Rewire was used.
SX export of Rewired Reason sounds IS NOT SAMPLE ACCURATE when exporting events which start at the beginning of the exported region, ie. if you have a note on bar 2 and you export starting from bar 2, this first note will be out of time. All other export of native-generated Reason and SX sounds IS SAMPLE ACCURATE. However there are some questions worth asking of internal timebases here, as all Reason exports (whether using Rewire or not) were slightly faster than the native SX export, ie. at 140 bpm: Reason - 18,881 samples/beat; SX -18,901 samples/beat. The length of 1 beat at 140 bpm (44,100 Hz sampling rate) is calculated at 18900 samples. This implies that Reason runs slightly faster than SX (which is pretty damn accurate!!!!), and that using Reason via Rewire slightly speeds up the internal timebase of SX to match Reason's tempo.
[/geek]
