No multiformat test at 128 or 96 kbps since we had this not very long ago (and 96 is pretty close to 128).

Does the test have to involve these specific samples or can you use any few CDs you have flying around (with "challenging" songs in classical, jazz, pop, rock, etc)? Or would this not be enough variance to accurately represent a simple random sample?

And if this is an impossibility, can all the samples be combined into one track (just for encoding), be encoded in VBR, then split again into its component sample pieces for distribution and testing?

I guess I'm just curious at this point; I don't mean to tell you how to do what you do best.

Two months of voting. How much longer till test starts?

I wanted to wait a bit for Ivan and the other Nero guys to finish the new encoder sicne they changed some paramters in the psymodel AFAIK.

The new test will be multiformat at 64 kbps. After that, either MP3 at 128 kbps or multiformat at 80 kbps (for 80 kbps, I also want to wait for Menno's test to finish).

New Nero encoder is out. So, will test start soon? Sorry for pushing I know it is hard to organize a test

Yep, it will start very soon.

Regarding a multi-format 128kbps test, I think it's best to wait til summer. Significant time (and development!) will have passed between the last test and the summer.

Apple should have released Leopard by then which will contain a newer version of their AAC encoder and LAME 3.98 final might very well be out. A test of improved Itunes AAC and LAME MP3 against constantly improving Nero AAC and Vorbis aotuv at 128 would certainly be useful then but a waste now.

Yes, IMO too 3.98 final should be in, and same is true for other encoders expected to be out til summer.

I am not going to test anything but MP3 at 128 kbps in the near future. The last multiformat listening test at 128 kbps ended with all competitors tied on first place so there isn't much room left for improvement. Things look different at 80 kbps though.

Also IMO it wouldn't be very interesting to conduct 128 multiformat test. If all encoders improve, which can be expected then they would have almost the same quality - very near 5.0. But then again you can ask only people that really can hear difference to participate. Not those that can only find a low anchor. Then again, how many people are here that can do that? 2, 3? Maybe to use only very hard samples in this test, like fatboy. Overall it would be really hard to conduct such a test.

If there is no low anchor, we can make thing more clear than ever. Quality at 128kbps is not flawless so that some of us differentiate the encoders. But I also second that we can't conduct that kind of tough test for now.