I saw very little discussion on this so I figured I'd throw a data point up and see what other people are doing with the encoder.

I grabbed the reference encoder for use for my iPod and gave it a whirl at various bitrates, intending to ABX. I tried q=0.15, 0.25 and 0.35 with -lc. Unfortunately I was a little too lazy to do a proper ABX session, but I was able tell artifacts with 0.15 somewhat easily (phew), but wasn't able to hear any sort of artifact with 0.25, even with sighted listening. So I'm encoding to 0.3 now.

0.35 clocked in at 138kbps for a set of test samples which is quite a bit higher than what is documented. But that's still far below the lame APS encodes I was doing before.

Needless to say, I thank all the Nero developers here for making this happen.

You should not use the -lc switch IMO, it makes the encoder use different quality thresholds and adds to confusion when comparing results.

I don't remember at what -q the encoder activates AAC+, but I would guess that LC bitrates lower than that are not well tuned and recommended.

But don't I need to make sure I'm using LC if I'm playing on an iPod?

No. HE-AAC is backwards compatible with LC-AAC players, they just won't play it with HE-AAC features. Just like mp3pro, non-mp3pro players will play mp3pro just fine but they won't play them to their full potentials.

However, I find that bitrates are scewed when forcing the -lc option. I encoded at -q0.45 and -q0.45 -lc, the straight -q0.45 came out to 160kbps VBR and the -q0.45 -lc file came out to 200kbps VBR. It is best just to let the Nero AAC encoder determine if SBR (used in HE-AAC) should be turned on or not.

The reason your bitrate came out higher is probably because you enabled the -lc option. Take it out and try it again, the bitrate will probably decrease by a minimum of 10kbps.

Your logic is not satisfying me.

HE-AAC obtains more "quality" at a given bitrate. HE files are playable on LC decoders, but without HE features - they are at a reduced quality. LC encodings are larger than HE encodings for the same quality level. Therefore, LC encodings at a given quality level must be at a higher quality than HE encodings played back on an LC player.

And moreover, when HE features are enabled in an encoding, the quality loss of LC is not specified. The constant-quality goal of VBR is violated. Therefore, to ensure constant quality encodings on an LC player, LC encoding must be used.

That's not true.
I have a Creative Nomad Jukebox Xtra, and it will not play mp3pro at all. It simply skips the files and says they are incompatible.

HE-AAC does have higher quality than LC-AAC for a lower bitrates, after that, they are both equal. So no, if you are targeting a 128kbps VBR bitrate then both the LC-AAC and HE-AAC will have the same quality. That is why it is pointless to use HE-AAC at bitrates over 64kbps, after that, it just starts sounding like LC-AAC.

So you don't have to enable the -lc option as the Nero AAC enocder will automatically determine if it needs to be turned on or not by the bitrate you select. Forcing the -lc option isn't even recomended by Nero. They recomend that you just pick a given quality level.

As far as mp3pro playback on non-mp3pro devices, I guess it varies. My 3G iPod (and 4G and 5G) are all capable of playing mp3pro files, they just can't play the "pro" material. I think non-mp3pro devices play mp3pro files at half the sample rate (don't quote me on this). Anyways, every device is different. My 3G, 4G, 5G, and older Rio 600 (back in 1999 with an amazing 32MB of built-in memory) all played mp3pro but they were all incapable of playing them to their full potential.

As I understand the original poster, the issue is not to generate files with a given quality level which also play on an iPod - albeit at possibly reduced quality - but to generate files with a given quality level that will be used primarily or solely on an iPod. As such it would make sense to be on the safe side and tell the encoder to not use features that are not supported on the target platform (iPod).

Axon, I think the reason for your inconsistency in bitrates is that the table of bitrates that was posted was for the default setting, without the -lc or -he flags. According to the developers, the quality scale is different for -lc, -he, and for no options.

I had similar findings in an informal test. q=0.10 was a piece of cake to distinguish -- the high frequencies were attenuated. q=0.20 was indistinguishable from the original WAV with my four arbitrarily chosen test songs. I am encoding at q=0.22, which has resulted in an average bitrate of 93kbps.

Anyway, according to the developer, if you don't specify -lc or -he, above q=0.31 the Nero encoder produces LC output. -lc guarantees LC output no matter what value you use for q, although as I mentioned the q scales are not the same. See this thread.

Since I'm only using AAC for my iPod, and the iPod documentation says that it needs AAC-LC, I figure it's safest to use -lc.

Thanks for the info. So at somewhere above 0.31 or 0.325, it will be LC anyway, but under this value it may or may not have HE features, and q-values for LC are not comparable. Good to know.

My q=0.3 LC encodes are averaging 123kbps right now so I think I will be very content with this.

Hi folx! A bit late but still want to share my story.

I've got LC compatable device (SGH-X700 mobile). I use -q 0.24 LC profile. It gives 105 kbps on avarege. For the headphones I'm using now it's ok. I tried HE with the same bitrate - sounds without SBR (probubly 22050 samplerate only). Also I tried -br 105000 LC. Sounds ok but when playing in Foobar2000 the bitrate goes up to 113 kbps when -q 0.24 LC goes to 120 kbps and higher (files of the same size). I decided to use Q-settings however I can't tell the difference. Maybe if I buy good quality headphones I'll have to re-encode with higher -q.

I just started playing with the new Nero AAC encoder so I don't know what quality settings I should go with. What -q level would ~equal a LAME -v3 fast encode. I don't mean file size or bit rates but audible listening quality on the likes of B&W bookshelf monitors. CDM1 to be exact. I also have Stax electrostatic headphones but I think my 30-16kHz 53 year old ears can't hear all that the Stax can play

Update: Based on some info I found here at HydrogenAudio, I think I'll go with -q.5 for Nero AAC ~170kbps and -V3 vbr-new for LAME mp3 ~175 kbps.

Now does the Nero AAC/mp4 have a slight edge over LAME mp3 at equal bit rates? How does Ogg Vorbis stack up?

Nero's encoder has not yet been compared against others at high bitrates. You have to evaluate it yourself.

According to the most recent tests conducted at 128kbps (can't find the link), Ogg, Lame mp3, iTunes AAC, and Nero AAC were pretty much statistically the same. Hoever, this test was conducted with a older version of Nero AAC (I think it was conducted in winter of 2005).

Still, I wouldn't imagine that Nero AAC has a huge advantage over Lame, if any at the 170kbps bitrate range.

How about at the -q.35 ~96kbps range?

Ogg Vorbis seems to have the best overall listening scores in a HA linked multiformat 128kbps ABX test I was looking at.

Of course the lossy codec needs to be mainstream so various players and mobile DVD/CD players can play compilations. mp3 seems well supported. mp4 most likely will be well supported too since the big players are behind it. Ogg Vorbis isn't listed on any players I've seen but some report there are players that support it even though it's undocumented in the literature.

I am not sure about 96kbps.

Although Ogg Vorbis attained SLIGHTLY higher results than other formats tested at the 182kbps bitrate, it wasn't anything major and it wasn't anything to warrant a statistical difference given the testing method and testing errors.

As far as hardware support, Lame is the highest, WMA is the second highest, and AAC is the third highest. AAC would be the second most supported format if you count DAPs sold (ie the iPod is the best selling) but I am counting DAPs on the market.

Anyways, you probably won't notice a huge difference in Nero AAC, iTunes AAC, or Ogg Vorbis at the 96kbps bitrates.

The only way you could ever convince yourself would be to download foobar2000 and conduct a ABX test. If you perceive Nero AAC at -q 0.35 (or -q 0.5) to be CD transparent then use it.

I using 0.9.2 now to play my hard drive music so I'll look into the ABX function and try some experiments of my own. I don't think -q.35 is going to cut it but -q.5 is good and LAME -v3 -vbr-new sounds about the same listening on my AR3a speakers via digital sound feed from my AD1980 SoundMax on board chip to RX-V450 Yamaha receiver.