Ear test ( audio rendering )

#2 seems to have some compression applied, and is slightly louder.

Wavefiles can be downloaded from soundcloud.

24bit , 44.1khz , sorry i forgot to mention that in the first post ;-)

Well, I wanted to play the game so I downloaded the WAV files. However it is not possible to do a real evaluation or a null test because they are not the same musical passage - the notes are different. Perhaps the differences relate to the discussions about the need to do real-time synth recording because of varying randomized synth parameters. If a couple oscillators had started at a different point on their cycle and maintained that delta throughout the recording, then the test is meaningless from the start anyway.
 
Another potential complication...
 

SoundblendWindows 10 Pro 64bit / 4Gb Ram / Radeon HD 6870 / Cakewalk Sonar Platinum / Steinberg Ci1

 
From the Voxengo web site (Italics added) :
 
"This plug-in is compatible with Windows XP computers (2 GHz dual-core or faster processor with at least 1 GB of system RAM recommended).  Note that this plug-in may not work correctly on the latest Windows versions that expose more than 2 GB of available memory to the host audio application.  Otherwise it is suggested to run this plug-in by bridging (for example, by using jBridge).  Also note that this plug-in may not work correctly with some 64-bit host audio applications via bridging."
 
We don't know what "working correctly" means - does it mean a crash, or perhaps something more subtle like distortion? I presume the former, but I don't know for sure.
 
Another issue is whether the SONAR export was fast bounce or real time. Any issues with latency, even if not immediately audible, will not occur with fast bounce. I presume Recorder is inherently real-time? 
 
For a truly scientific test, the file that's rendered should be EXACTLY the same source material. Also we do not know what the synthesizer output is supposed to sound like, so we can't judge whether one file is better than the other at reproducing the actual sound of the synth. For an extreme example, some people prefer the sound of an MP3 file over a WAV file, even though the WAV file is more accurate.
 
You would need to render the same musical passage over some other programs as well without Recorder to see if those audio engines sounded more like what's coming out of SONAR or what's coming out of Recorder. If the majority sound like SONAR, you'd likely conclude Recorder had a problem with accuracy. If the majority sound like Recorder, you'd likely conclude SONAR had a problem with accuracy. But for now, there's no way I see to draw any meaningful conclusions from this test.

bitflipper
#2 seems to have some compression applied, and is slightly louder.

( The project is the same, no changes in EQ, compression or other.. )


I did import both files in Wavelab elements 7, and the readings of  the peak's and Rms values is the same
on the meter but :

I agree that one sound's louder....

The ears are a fine instrument, but can also be fooled

Thanks for your comment ;-)

Are you going through the ProChannel on the master Bus?
 
Is it set to be pre or post the FX bin?
 
That could account for a difference in what a plugin recorder would get for its input and what Sonar would export.
 

Tom Riggs
Are you going through the ProChannel on the master Bus?
 
Is it set to be pre or post the FX bin?
 
That could account for a difference in what a plugin recorder would get for its input and what Sonar would export.
 

I use FX in Prochannel ( pre mode ), with the Concrete limiter last in the chain
At last, Sonimus Britson BUSS plugin in the master FX bin.



http://postimg.org/image/6fahvdvkd/

This is a new render i did real time :

1. Render 1 new

2. Render 2 New 


Those should be fair to each other compared to earlier versions, since both here is a real time render / Export

1. Are there any difference at all 
2. What of those do you think is from sonar and what is from the voxengo recorder
3. If there's a difference, what separates the two

Jan S

Well, they're still not the same piece of music, so you can't do a truly meaningful A/B comparison. Anyway, #2 levels are a little hotter. After normalizing the tracks to the same levels, I could hear no difference. Interestingly, with the examples yesterday, I could hear a definite difference. However, after realizing today that the levels weren't matched, I went back and normalized yesterday's files and they sounded the same after normalizing.
 
So my conclusion is...digital audio is digital audio