Hi!,
I am currently developing my AAC encoder and testing it with pqevalaudio tool. It gives ODG measure.
I have observed that when I, for example, use only short windows the ODG is much better than encoded file with only long windows (same parameters etc) but in subjective listening (it wasn't professional test but with me and several friends) gives the opposite results (sequences encoded with long windows give better quality).
Can anyone explain why PEAQ gives wrong results? Is there any mistake in tool? I've tested it also with EAQUAL.exe and results are similar to pqevalaudio.
Is there any sense to use ODG measure if it gives so strange results?
No. Always trust your ears, and those of a few other listeners, using double-blind tests on a lot of test items.
Chris