Here are the final results for the IS/PNS 80 kbps LC AAC listening test:

http://www.audiocoding.com/results/image003.gif

http://www.audiocoding.com/results/image004.gif

Since the main reason for this test was to determine whether PNS and/or IS gives any advantage, results where only low anchor was identified were not counted.

Chunky was used with option -p 0.05 to remove results with a wrong grading.

Key, encoders used, results used and ratings are in FinalResults.zip.

Results that were not used because no relevance for this test are in NotUsed.zip.

Thanks to everyone who helped us with this test!

So, would you conclude there is no reason to use either IS or PNS or both at bitrates as high as 80kbps in your codec?

Yes, the results show that our current implementation of IS and PNS have no benefit over not using it. Given the cons of using these tools (read: iPod can't decode it properly) it is probably not a good idea to enable them.

OK. Are you thinking of enhancing the current implementation or starting a new listening test at a lower bitrate to determine the threshold?

I agree on the cons, try to keep it LC as much as possible/desirable.

Would LC even make sense at a lower bitrate?

According to FCD 14496-3 IS and PNS are part of the MPEG4 "AAC low complexity" profile.

The test results are interesting. All confidence intervals overlap so you can't say for certain that one setting performs better than another. I guess it's just trading different kinds of artefacts for Ahead's current implementation. I wonder if there's room for improvement. I expected the IS / PNS tools to be more useful, actually.

Cheers!
SG

I personally wouldn't think so especially if PS/INS cause compatibility problems. Not even sure what players can decode LC-PS/INS but not HE. Rockbox?

Thanks for posting the results, menno. Glad I could help a little.

Why not if you have a device which supports only LC and quality is acceptable? It is always better to have more songs on a device

Sort of reminds me of all those people who wanted to encode to mp3pro when it came out at 128 Kbps or above...

About the ipod, are there patent fees for decoding PNS / IS?

Nothing extra, if you have an AAC license you can do PNS and IS, it's part of LC AAC.

If it is more complex to decode it may be a battery issue.

If it was intentional, it wouldn't sound like it does now.

Complexity should actually decrease for IS as less MDCT data should be decoded

Same goes for PNS, too.

I wouldn't go into speculating why Apple iPod has no PNS support, but my best guess would be that it was just not tested because iTunes Encoder does not use PNS... but could be any other reason as well.

Hhmm... I see.

Well, considering there is a new Apple AAC codec coming with OS X Leopard and native FLAC support, who knows what else they'll change in the audio department.

I am more or less expecting HE-AAC, at least for streaming playback, since iTunes is starting to lose out on more and more radio stations by not supporting it.

when you guys do the next test, throw an email my way. i can encode samples using AAC on my PS3... uses the 3gp container format, tho o_O

Quick Answer SONY, the psp could only playback aacs in that back when!