which is more efficient (better quality for the same bitrate)? ran out of space on my ipod (used to encode with apple lossless), and im looking for transparent audio on shure e4cs at the lowest bitrate possible. which is better for the job? and if AAC, which encoder (nero or itunes) and at what settings?

I am quite sure the general consensus are that AAC are more efficient than MP3. But in the end you might not be able to hear a difference, using the same/comparable bitrate. Both Nero and iTunes are safe choices for AAC.

Why not give it a try?

Both can achieve transparency in the ~128 kbit/sec region, depending on the samples used. However, with AAC you might be able to go down to ~90 kbit/sec without hearing any difference. Use ABX testing to find out what is transparent for you.

Keep in mind that iPods don't support HE-AAC (used in low bitrate modes by Nero) and that some new iPod models don't play Nero AAC at all.

My opinion:

At or above 160 Kbps most people can't tell the difference between LAME MP3 and iTunes AAC.

Below that rate, iTunes AAC is better.

I did a personal listen test on 128 cbr and 128 abr about 2 years ago (i did ABX those but lost the logs), and found that Ogg at q2 and iTunes at 128 ABR sounded better then Nero AAC at q0.35 and Nero only sounded a tad better then LAME -V 5 --vbr-new.

A year later i tried out Nero AAC again at q 0.35 with a track and also with iTunes AAC at 96kbps ABR. I could not ABX the iTunes encode, but the Nero encode was easy since there was a nasty precho artifact on it. Then a few months later i tried a newer version of Nero AAC and performed worse then the Feb 12 07 version.

I week ago i tried out Nero AAC Feb 12 07 again, and I was shocked that I could manage to ABX at q 0.60.

Serj Tankian - The Unthinking Majority [Nero AAC Feb 12 07 q 0.60]



There is a precho on this someware but i forgot were it was, I will have to listen it again sometime. The artifact also appears on the August build aswell along with a 20kbps bitrate increase.

I also got 10/10 on Metallica's For Whom The Bells Tolls, which had warbling at 0:48, which i think it is really unexpectable to have artifacts like that at 220 kbps VBR.

From my personal listen test in the past and today, I find Nero AAC to only be a tad better then LAME, but i find Nero AAC has the most annoying artifacts i ever heard while the artifacts on LAME are lot more tolerateabale.

/mnt... can you abx OGG (aotuv5, libvorbis or lancer) -q4 with these samples?

PS: Do you run linux? (eg. /mnt)

I can't ABX the difference (on problem samples) between iTunes AAC at 160kbps (VBR) and Lame MP3 at -V2 --vbr-new (around 192kbps). Right now i'm using the MP3s just because they'll stream to my xbox 360 as well... will probably switch later when my iPod starts to get full.

(I have lossless backups of all my CDs, so i can switch lossy formats whenever i wish).

edit: removed false statement; apparently Nero is now gapless on iPods.

http://www.hydrogenaudio.org/forums/index....st&p=505251
http://www.hydrogenaudio.org/forums/index....mp;#entry496270
http://www.hydrogenaudio.org/forums/index....mp;#entry476797

ok...so it appears iTunes or Nero AAC at around 192 is the safest bet. does anyone know for sure if nero is compatible with an ipod touch? silversight made a comment about newer ipods not playing nero AAC. also, how is the iTunes AAC VBR? and is it worth using Nero to encode vs. itunes?

I have a 16GB iPod touch and now encode my music using the February 2007 build of Nero's AAC encoder. The files work fine on my iPod touch as well as my other AAC compatible devices. Nero AAC is 100% gapless on the 5G, 5.5G, iPod classic, 3G iPod nano, and iPod touch models.

My ABX results are actually different from /mnt's in that I tested the latest iTunes AAC encoder against the February build of Nero's encoder at 128kbps VBR (-q0.4 for Nero and 128kbps VBR for iTunes). I actually found the Nero encodings to be a whole lot harder to ABX. In fact there were many samples that I couldn't ABX so I encode my audio at -q0.45 just to be safe. The files produced are around 150-160kbps VBR and are transparent to me. I also like that you can use EAC, dbpoweramp, or MAX (for Mac OS X users) in order to get secure and accurate rips. You can do this using iTunes unless you use iTunesEncode which I don't really like.

I suggest ABXing for yourself as 192kbps VBR AAC is actually overkill for me and for many other people. The default setting for Nero AAC is -q0.5 which produces files around 170kbps VBR. Download foobar2000 and conduct for yourself. After all, you will be the one listening to your music with your ears using your equipment. No need for us to tell you which bitrate/format to encode at.

iPods and iTunes started to support gapless playback on Nero AAC a few months later later after the 5.5 iPod and 2G Nanos were released. I can listen to Pink Floyd's Dark Side Of The Moon and Nine Inch Nails' The Fragile and The Downard Spiral at 100% gapless on Nero AAC on my iPod Classic and 2G 4GB Green iPod Nano. Also iTunes does not take ages to pick up the gapless info on Nero and iTunes AAC files, unlike LAME header mp3s.

Yeah am also starting to rip to lossless for archive to a DVD and still to stick to lossy for desktop and portable use, and have lossless for future use.

at the current level of improvement in LAME, its only slightly worser than AAC. But there are other factors like battery usage, and as far as I am aware Vorbis and AAC use lots of battery power. So it might be a wise option to concider lame. the problem is there in compensations. You need to compensate one thing for the other.

On a personal note, I have found Nero or Apple 80 - 96kbps vbr to be quite clear and near transparent. You might want to use it over 128 kbps mp3, but don't replace 128kbps lame with 128kbps nero, as the difference in quality will be little compared to power consuption.

Do you think that the difference of battery life is still noticeable for the same quality (AAC 128 kbps =+/- MP3 160 kbps)?

Yeah i use Linux, mostly use Debain or Ubuntu used to use Red Hat but Red Hat it went downhill after Fedora Core 4 it was dependency hell so i now avoid all rpm based linux distros. But most of the time I still use Windows for foobar and EAC and for PC Gaming.

Anyway I will try The Unthinking Majority and For Whom The Bell Tolls on Ogg aotuv5, when I have some free time. Anyway I finding that The Unthinking Majority seems to be hard to encode track since it is not transparent at LAME 3.97 -V 2 --vbr-new.



A random precho at 3:09.

The track is transparent on Nero AAC at q 65, I have not tried it at -V0 --vbr-new on LAME yet.

I have a Sansa e270 6GB with rockbox for half a year now and I play mp3, aac and ogg, all between 120 and 160 kbps. I use Nero aac and noticed it is clearly very powerhungry, after 8-9 albums I'm out of battery but with ogg (aotuv) and mp3 (lame vbr) battery last always 10-12 albums. But my surprise was with FLAC, after 14 albums the Sansa still hadn't shutdown allthough it was close to the limit. I will try mp3 and ogg at about 320 kbps because I think the higher the bitrate the less battery is used.

My listening habbit with my dap is always the same. I just listen full albums, one after an other, with ReplayGain and volume between 0 dB and +3dB. Browsing is very minimal.

I want to remark that my experience isn't just a result of one time usage. Weekly I travel by train 6 hours and I always listen music and repetitively I see the same battery life. This is why I now use aac only occasionally.

Rockbox has poor support for AAC, and also Rockbox uses software decoding on AAC files on most DAPs.

Rockbox uses software decoding for AAC for all DAPs on which AAC is supported. (For all supported DAPs except the Archos DAPs, Rockbox uses software decoding for all codecs.)

What about MPC ~ 175 k ?? Proven and robust for many years. Seems like a great choice for size / quality / battery life.

See below.

Not very VBR-ish in some respects.

AFAICT, there is now a "true VBR" mode in Core Audio, but it's not accessible from iTunes's GUI. (Whether it's present and accessible at all on Windows, I don't know.) You should be able to find more details on that from old postings:

Google search: itunes "true vbr" site:.hydrogenaudio.org/forums.

(Perhaps you may want to go into the "Advanced Preferences" and restrict the time frame on that search to the last few months.)



I think both encoders are fine. I've got plenty of files encoded with both encoders, and derive much enjoyment from what I've encoded with them.

If space is an issue for you -- as your original post implies -- you may prefer to use Nero's encoder. For example, if you were to encode with Nero's encoder at the default setting (q 0.5) the average filesize would be smaller than if you were to encode at the nearest equivalent setting in iTunes -- which would be 160kbps with the VBR box checked.

That's perhaps the issue (assuming you've no preference as to ripping software.)

This seems to arise because you can't set a true quality level in iTunes, only an explicit bitrate (although you can also ask for VBR-ish behaviour together with that bitrate). When VBR is set iTunes will range a fair bit above that bitrate when it needs to, but not, apparently, dip much below it when actually there'd be no call for it not to and it arguably should. That explicit bitrate seems to act like a "floor". Therefore, it seems, iTunes will waste bits.

But if space is not as much of an issue as all that, I think either encoder is fine.

I've been thinking about trying mpc. I've played 2 mpc albums but that of course doesn't lead to a conclusion. But this is kind of offtopic.

I think the power consumption issue is irrelevant if one uses the built-in, factory default firmware for their devices. I have run tests using Nero AAC, iTunes AAC, and Lame mp3 at various bitrates (128kbps VBR, 160kbps VBR, and 192kbps VBR) so see the differences between battery playback. Well, for all bitrates, both my 5G 60GB iPod, and 16GB iPod touch would get the same battery life for both AAC and mp3 files. The AAC encoder didn't seem to affect the battery life either as I figured that Nero's AAC encoder would eat up more juice since it is more of a true VBR encoder. My 30GB Zune had the same results as well. The only player where there was a slight decrease in battery playback time when using AAC was with my 4GB Creative Zen. It is rated for 22 hours of playback and it got 21 hours when using Lame mp3 at 190kbps and 21.5 hours when using Lame mp3 at 128kbps VBR. Both Nero AAC and iTunes AAC would get between 19.5-20 hours depending on the bitrate. I wouldn't concider that "lots of battery power."

So it is definitely safe to use AAC on a portable player without Rockbox as the battery playback time differences were slim to none. If you must use Rockbox then stay away from AAC as it isn't very efficient on that firmware.

Which Ipod models do not play Nero? That is a disturbing development for me. Is there a fix?

It is possible that I mixed two topics in my head, as I can't find the corresponding thread for my claims anymore. Sorry for the confusion. I hope somebody can lighten this up a little bit.

Pretty much all iPods will play Nero AAC files. I think some people had issues with the 1G/2G/2.5G iPod shuffle but I have both a 2G and a 2.5G iPod shuffle that play Nero AAC files just fine. I have had positive experience with Nero AAC playing on my 3G 40GB, 4G 40GB, 5G 60GB, 2G 1GB Shuffle, 2.5G 2GB Shuffle, and 16GB iPod touch along with my 30GB Zune and now I found a simple fix for my 4GB Creative Zen. So yeah, there aren't really any issues. HE-AAC support still isn't here (it would have been nice with the 16GB iPod touch since that isn't a whole lot of storage space) but LC-AAC files created by Nero work fine.

It's not something I've paid much attention to, because I don't use very low bitrates, but I can find a few references by googling "ipod he-aac".

According to Wikipedia:



However, there's a thread here (HA) that comes up, where posters indicate that they have played these files on iPods. But there's some question about how much they're getting out of them:



http://www.hydrogenaudio.org/forums/lofive...php/t32153.html

I recently stumbled on the glory that is HE-AAC - I was shocked when I heard how good they sounded at low bitrates, and they played very nicely in Foobar2000. Then I loaded one up into iTunes - it reported the track length incorrectly (approximately twice as long), and when it played it didn't sound like it did in Foobar. Somewhere I read that what it is playing is actually the LC part of the file, but I can't find a source for that right now. In any case, I would imagine iPods would probably handle these files similarly - it will probably play (possibly buggy), but you won't get the full quality of the HE-AAC file.

It is true that iTunes, QuickTime, and iPods can play HE-AAC files. They will only play the LC portion of the audio file though so a 64kbps HE-AAC file will have the same audio quality as a 64kbps LC-AAC file when being played through iTunes and from an iPod. HE-AAC support has been something that people have wanted back in 2003 when Apple released the 1G iPod mini. People have signed petition after petition and, 5 years later, we still don't have anything.

I doubt Apple will add HE-AAC support now given that flash memory capacities keep drastically increasing while their prices decrease. Before there was a need for HE-AAC as the iPod mini's held 2GB and 4GB. Now the iPod touch is up to 32GB and I wouldn't be surprised if Apple releases a 32GB iPod nano this fall along with a 64GB iPod touch. It would be nice if Apple added HE-AAC support especially since my iPod touch has a 16GB capacity but I don't see that happening.

Actually it will have worse quality than a 64kbps LC-AAC file because of the space occupied by the HE information. If I am not mistaken, the audio is resampled to half the rate it would normally have at 64kbps and the HE information is used to simulate the lost high frequencies.

I believe that AAC-HE files when decoded as LC only contain 1/2 the sampling rate of the source file and effectively a reduced bitrate, so the quality is generally quite a bit worse then the equivalently sized LC file.

Oops, yeah, my mistake. I meant to say that a 64kbps HE-AAC files will have less audio quality than a 64kbps LC-AAC files. Actually I originally wrote "a 64kbps HE-AAC file will have the same audio quality as a 64kbps HE-AAC file" so I edited it on the spot to say LC-AAC. I am listening to the new Nine Inch Nails album and for some reason, one of the songs made me a little dizzy.

Sweet!! I had no idea. I'm going to use it now, for sure! Sorry about the out-of-date info; post corrected.

Care to name a DAP which doesn't use software decoding for AAC? I can't think of a single mainstream DAP of the last years which doesn't use a general-purpose CPU or DSP to decode all audio codes in software.





Nero AAC 320 benchmarks out on Rockbox on your hardware @ ~155% realtime for decoding. 128Kb/s=180%
Lame MP3 320 benchmarks out on Rockbox on your hardware @ ~200% realtime for decoding. 128Kb/s = 212%
Vorbis 350 = 200%. 128Kb/s=257%

FLAC 8 decodes @ 600% realtime on your hardware.

http://www.rockbox.org/twiki/bin/view/Main...manceComparison

I'm actually a little surprised you're seeing such a direct correlation between codec and runtime lengths. While the CPU is the largest consumer of power, it is hardly the only.

Your conclusion is wrong, Alexxander, though.
The higher the bitrate the more battery is used - in two ways.
1 - more CPU to decode. (Since Rockbox dynamically clocks the CPU, less time @ idle = more battery used.)
2 - more disk access to refill RAM buffer. (Yes, I know you have a flash player, but reads (while not needing to spin-up a HDD) are not w/o their costs.)
What you are seeing is simply that FLAC is very efficient to decode.

I think that actual battery life numbers are a better indication of real world performance. Sure, it is nice knowing the real time decoding speed but what does that tell us in the end other than how fast the firmware decodes? From my experience with Rockbox on a 5G 60GB iPod, it doesn't get the same battery life (on a single charge) as the iPod's default firmware when playing either AAC or mp3 files. Hence, that is why I made that comment. Now things might have changed as I did that test about 4-5 months ago. I have no need for Rockbox so I don't want to go through the testing process again.

Things have gone from 55% to 80% of Apple runtime on the iPod Video, and from 66% to >100% on the Sansas.

But I wasn't trying to convince you, kornchild2002, of anything. I simply quoted you since I was bringing some real numbers in which supported something you had said earlier. My main point was to Alexxander, that high bitrates didn't cause his good runtime,FLAC caused his good runtime.

Just to clear things up...

I haven't trained myself. I encode at .45 because the .50 of nero adds a lot of the filesize.

nero is gapless on my ipod classic. You have to make sure the files are .m4a, not .mp4

Again, if you train yourself, you can make any music miserable. I couldn't detect anything on nero aac so I'm happy with it.