Hello, I'm a french user of Nero 7.

Since some versions, the quality settings for Nero digital encoding (Internet, Streaming, Transparent...) have disappeared. Now, there's just a cursor and default setting is 142 kbps. When I rip my cds now, I don't know if I make MPEG2 LC or MPEG4.

I saw on Ahead's website: "Transparent quality at 128 kb/s". Does it mean that if I set the cursor to 128 instead of 142, I achieve a "Transparent quality"? Or do I have to change a other setting somewhere? It's quite incredible for me to have that quality with a so small file...

I would really appreciate an answer from a Nero's developper, but others "specialists" are welcome.

Thanks to all and excuse my poor english.

edit: Oh dear; what a kerfuffle. For the purposes of those looking for information, I take it all back.

You should test for yourself.

http://www.maresweb.de/listening-tests/mf-128-1/results.htm
You sure about that?

MPEG2 and MPEG4 are standards suites. They contain a description of an audio codec, called Advanced Audio Coding (AAC), together with a set of profiles. The only one widely used in practise is the LC profile. This is also what Nero AAC uses.

MPEG4 adds some improvements over MPEG2 to this codec, and later on some new profiles were added as well. These are HE-AAC (using SBR) and HE-AACv2 (using SBR+PS) and they perform well at low bitrates. Sometimes people erroneously call these aacPlus and aacPlusv2.

The Nero encoder will use either MPEG 4 LC AAC, MPEG 4 HE AAC or MPEG 4 HE AAC v2, depending on the bitrate and encoder settings. Normally, high bitrates (>= 80kbps) use LC, lower bitrates use HE AAC, and the very lowest ones (<= 32kbps) use HE AAC v2.

You can find threads here with information on what bitrates and quality settings cause the encoder to use which mode. But this is subject to change as the encoder is improved. You can also force the encoder into a specific mode with commandline settings.



What is transparent is mostly dependant on the listener. Recent tests on this forum have shown that even for trained listeners, in the majority of cases (>85% if I remember correctly), people are unable to distinguish an encoding from the original at 128kbps.

If you are not a trained listener, the odds are that 64kbps would be transparent too, but we err on the side of caution.

Thank you for all these answers.

Specially to Garf for the clarifications.

I understand that AAC quality has become very subjective due to the quality of the encoders (Nero in particular). I mostly listen to pop, rock and metal and I was pretty satisfied with the Streaming preset.

So I have a last question. For a given album ripped with Nero 6 at the "beginnings" of the encoder, if I ripp it today with Nero 7, would I achieve a better quality with the same bitrate or file size?

You would the best judge of that . I do believe that your results will be better compared to the older encoder.

whoops, the average bitrate of all encoders seems to be ~140kbps not 128!

Yes. You might want to use the standalone, free commandline encoder, since it's even newer than wat is in Nero 7.




You are wrong. The settings correspond to an average of 128kbps for a wide variety of music. Just read the threads about the tests. A particular set of samples may produce a different average, but this doesn't change the fact that the overall average is still 128kbps.

OK, everything's clear to me.

I'd like to use the free commandline encoder, but I'm used to the GUI of Nero 7. It's not possible to use the free encoder through Nero 7?

No. You can either take a look at the frontends various people made, or wait until it getst included in a future Nero 7 update.

Thanks for all

Ok, so I was both incorrect and typed at least two words that I didn't mean to, but those results don't prove in any way that 128kbps is enough to be transparent for any audio samples (of course they don't disprove it either). What that test shows as that 137+kbps is enough cases.

I'm not claiming it is (warnings, eh?), but from those results alone, it's entirely possible that an actual VBR that averaged to 128kbps for any one of those samples would not be transparent. Why is it that all the bitrates are so high? That's not a criticism of the test, but the only explanation I can think of is that these are intentionally problematic samples, so they use more bits. In that case, wouldn't it make sense to also test "simpler" files (i.e. files which fall beneath 128kbps at the same setting)? Apologies if this has been asked before, but I haven't read the topics (obviously, else I would have known the result..!)

tgoose, the idea is that those settings will provide 128 kbps on average when you encode a lot of music in various genres. If those samples are higher than that, then by deduction, you can assume they are problematic samples since the psymodels adjusted for them. If they didn't adjust, then that is a knock against the encoder, isn't it?

Anyhow, you made a claim, we refuted it, and now you need to provide evidence to counter ours. Otherwise your claim that 128kbit/s isn't enough is unsubstantiated at best. The listening test results specifically indicate that a lot (most?) people found the samples indistinguishable. We couldn't even get statistically significant results proving there was a difference, for ANY of the codecs!

Those samples where about 30 seconds long and many of them had spots where the bitrate was under 128 kbps. The whole sample was tested and one could look out the audible difference anywhere in the sample. When you look at the scoreboard and the results which are close to 5, you have to think, how many of the all (more or less trained) listeners didn't hear any difference. Those results show's that it's very highly a likely that the sample is transparent to it's casual listener. You should take a look at those samples yourself to see how close to transparency they actually are.

The exact numbers and percentages are in one of the post-test threads.




Maybe, but then it wouldn't be an 128kbps test, because the real average of the encoders would then be LOWER than 128kbps.



Correct. This would have been very obvious if you had even bothered to read around.



What is the point of making the test simpler if a) the encoders are already performing so well most people can't hear differences b) they are all good enough that it's almost impossible to determine who's better.

Using easier samples would indeed lower the bitrate, but won't cause any problems for the encoders either. Think about what happens when the encoder mis-guesses the difficulty of an easy sample, and what happens when he mis-guesses the difficulty of a hard sample.

I don't doubt that they are (when I get a working audio interface I'll listen, for sure; At the minute I have nothing but terrible internal laptop speakers, so I'd be lucky to ABX 64kbps!) I'm not questioning that those samples are transparent because I have no reason to suspect they might not be, and no way of checking. Since I hadn't read this test before, my first statement was not based on it whatsoever and thus doesn't attempt to comment on it at all.


I don't say make the test simpler, I say test the lower bitrates of that setting; if it misguesses an easy sample and thinks that it's easier than it really is, surely that's just as likely to cause problems as if it misguesses a difficult sample and thinks it's easier than it really is?



I've admitted I was wrong and if I didn't before, now I apologise for misleading the original poster; I'm almost certain that the last time I looked at such tests (it may have been some time ago...) they were nothing like so positive, and I was surprised when I saw that one; not keeping up with later tests may not be an excuse, but it's certainly a reason. By admitting being wrong, surely I'm taking back the statement I made? Since I have seen tests that go some way to contradict it, there's little point in my trying to continue. What I will do is continue questioning until I find out why samples with bitrates falling toward the lower end being tested.