Within the next month, I shall be using a new computer, an iMac G5, and plan to re-encode my entire music collection in iTunes on this machine. For the last several
months I have been diligently encoding my collection to Apple Lossless using EAC to rip and iTunes to encode. All songs have been properly tagged with included artwork and archived in ALAC to a DVD. Thus the re-encode will consist of simply right clicking a list of Apple files on the DVD in iTunes and converting to a desired format. All tag information will transfer over on the conversion.

The question begs, what should I convert to? I will be listening to digial audio on the following equipment in decreasing order of use:
1) iPod with Sennheiser PX 200 and 100 headphones normally, Apple ear buds for working out.
2) mixed audio CDs burned from the digital audio files on a Ford Mach stereo system
3) iPod through a Sony DTS surround receiver with Onkyo 5.1 speaker system
4) iMac G5 with Creative iTrigue L3450 speakers

I am considering 3 alternatives with a small chance at a 4th
1) LAME 3.90.3 --alt preset standard
2) LAME 3.96.1 --preset standard
3) Apple AAC 192
4) Apple AAC 160

For those that do not know, on a Mac, Blacktree offers a program to use LAME within iTunes instead of Apple's crappy MP3 encoder.

Each format has slight disadvantages and advantages (98% of all music sounds transparent to me at the these settings)

I, and am sure many other iPod owners, would love to hear the knowledgable digital audio people at Hydrogen Audio weigh in on their preferences and why they feel so.

For more info on the debate between 3.09.3 and 3.96.1 - here is a forum where the debate rages: http://www.hydrogenaudio.org/forums/index....showtopic=27650

Ezra, that seems like a great set-up.

I use iTunes for playing music and synching my iPod.

On the side I have my songs ripped by EAC and stored in FLAC.

In order to find out what to use when converting the songs for the iPod I have made a program ipod-abx, that allow me to ABX songs on an iPod.
iPod-ABX

I use the Sennheiser PXC-250 and was supprised about my diffuculties ABX'ing even 128 kbps AAC and LAME. Apples MP3 encoder fell through though.

Here are the results:


I was expecting to be able to better distinguish between the codecs, but no.

You could try this on your iPod combined with the Sony DTS receiver. You might be supprised about your findings.

Why re-rip just because your getting an Apple? Apple computers play MP3s, IPODs play MP3s...I don't understand what your gaining by doing this.

My files are in a variety of formats and compression settings as they have been ripped and compressed (or downloaded) over the years as I have learned more and more about digital audio - mostly through this forum. Many are in LAME 3.92 at 192 CBR or LAME 3.90.3 at 160 CBRApple AAC 128 or other.....

With a new PC, its a chance to start fresh and take advantage of the latest and greatest. Rather than just copy my files over, I can create fresh compressions of them. If I had not been tagging and archiving, I would not take the time to do so. But its really a matter of loading in 3 or 4 DVDs, select all files, convert to ______ and go to sleep. In the morning, delete the lossles files from the computer, leaving only the tagged MP3/ or AAC copy. Not a whole lot of work. Now re-creating the smart playlists is another matter........

The question is what to compress to......(topic of the forum)

in my opinion, from the 4 options, i'd pick LAME 3.90.3 --alt preset standard. if it doesn't fit the ipod, i'd pick LAME 3.96.1 --preset standard.

Thanks for the reply Twostar. 3.90.3 APS does fit the iPod quite fine. What are you basing your decision on? That's kind of the point. I have my own opinions for and against each of the 3 codecs but wanted to hear from others.

well, it's just me but i pick mp3 over aac because:

1) i trust lame preset standard quality
2) i can play my files on just about anything, not just on an ipod

i'd choose aac only if i was conserving space and have to use low bitrates.

I'll have to agree with the poster above: there's not a huge reason to re-rip everything to AAC if you already have a bunch of --aps encodes in your possession. You don't gain much in terms of space by going to 192kbps AAC, and with anything lower, you'll risk losing transparency.

I myself migrated to OS X a little more than a year ago. Although I've played with AAC and ALAC, the vast majority of my stuff is --aps encodes. Why? My Mac plays them, my iPod mini plays them, my car plays them, and my new DVD/SACD/DVD-A player will read MP3's right off of DVD+R's. Try that with AAC!

Of course, for a person starting out with ripping, and not needing the compatibility of MP3, AAC might be a wiser choice. It rips much faster on a Mac than LAME, and most likely will give very good quality at 192kbps, though there is no objective data (that I'm aware of anyway) that shows it to be more or less comparable to --aps; that's just conjecture at this point.

edit: nonsense deleted, nothing left
to myself

Yes, that seems to be the main point of confilct:

The compatability of LAME --alt presewt standard MP3 vs. the less burden on the iPod battery of AAC.

My iMac ships next week so it is decision time, QT 192 AAC or LAME 3.96.1 APS MP3 to encode my music collection.

The approx. 10% file size reduction and better iPod battery efficiency of AAC or the multi-device compatability of MP3? Both codecs at the above mentioned setting produce transparent results to my ears so sound quality is not the issue.

However, is the general consensus now that iTunes 4.7 has fixed the AAC bug that existed in 4.6?

Decisions, decisions - and give the HOURS that will go into compressing 10k fully tagged songs from Apple Lossless to "x" - this decision will have to last!

In terms of CPU power, mp3 seems to require less than AAC. On a 128 kbps AAC, my 4G iPod's clockspeed hovers around 30 to 32 MHz. I played a 256 kbps mp3 today and it seemed to hover around 28 to 29 MHz. Not sure if that correlates directly to battery life though.

That is interesting. I would assume the reason AAC is easier on battery life is because the processor does not have to work as hard. Your evidence seems to conflict AAC being easier on the processor.

Yes, I found it interesting too. I'll do some more tests (ie. play some more 256 kbps mp3 files) and have a look at the processor speed on average. I only played one song yesterday so it's not conclusive.

this my recent decoder test with dbpoweramp (winXP, P4 1.4 GHz)
they are just normal compile, not ICL

decoded using libVorbis 1.1..0
OGG 092kbps VBR 69X
OGG 122kbps VBR 62X
OGG 166kbps VBR 55X
OGG 180kbps VBR 56X
OGG 309kbps VBR 43X

decoded using FAAD2
AAC 096kbps ABR 50X
AAC 128kbps ABR 50X
AAC 160kbps ABR 48X
AAC 320kbps ABR 42X

How nice if Ipod supports OGG Vorbis

On an IPod, AAC is using more power than MP3.

I've read just the opposite in other posts. However, if what you say is true - why would anyone use AAC right now? Roberto's listening tests have shown LAME holds its own with AAC at 128 kbps and at transparency settings (192 AAC and APS LAME), AAC offers about a 10% advantage in file size over MP3 which is close to meaningless. MP3 offers uncomparable device flexibility.

Note: transparency settings reflect transparency to MY ears. They do not reflect the results of proven listening tests. Rule #8 junkies - relax.

Be careful: you're talking about the overall results of the test, including various kind of listening abilities. For some people, lame 128 is clearly inferior to AAC 128 - and for them, AAC is much preferable to MP3.

I encoded a song into an mp3 at 320 kbps. I encoded the same song into AAC at 128 kps.

For the mp3 file at 320 kbps:

Max = 30 MHz
Min = 26 MHz
Avg = 27 or 28 MHz most of the time

For the AAC file at 128 kbps:

Max=33 MHz
Min=27 MHz
Avg = about 30 MHz

The AAC file fluctuated a bit more than the mp3.


EDIT: I should note that the iPod dynamically changes the processor speed so it may be programmed in such a way to use more MHz than is actually needed for AAC than mp3. Hence it may be to do with the iPod's speed changing algorithm rather than actual raw speed required for decoding.

How do you check the iPod processor usage?



The "official" word is that AAC is 30% more efficient than mp3, so a 160kbps AAC should be roughly similar to a 192kbps mp3.

However, there are some areas where even 128kbps AAC can outperform mp3's at 320kbps, such as in stereo depth and high frequency accuracy (thats according to a sound & vision article some time back).

Since AAC is an non-proprietary format, there's nothing stopping it from reaching the same level of compatibility as mp3 has today. Also worth considering is both that the AAC encoder is a lot better optimized for the Mac than what LAME is, and that there is lots of room for quality improvements the coming years..

Yes, I have read that multiple times as well. However, LAME APS file sizes for the music I listen to average around 210. A 30% improvement would indicate a comparable AAC file size of around 160. I know I can still ABX from the original at 160. I cannot at 192. However, 192 is only a 10% improvement.



Optimized for MAC? Other than encoding speed, how does this translate into improvements?

I agree that AAC will get better. I should have said "why would anyone use it as it exists TODAY over LAME MP3?

Yeah, I'm only talking about encoding speed. LAME has yet to be optimized for the vector processing units on the G4/G5.

http://www.hydrogenaudio.org/forums/index....ndpost&p=249073

if you re-encode mp3 you'll lose quality...yes it is as simple as that...so why bother re-encoding? you'll just lose quality...would be just as smart to rip all your cds in 128 then break the cds and burn the 128 rips to cds...lol

From Lossless

I conducted a file size comparison (via bitrates) between LAME 3.96.1 preset standard and iTunes 4.7 AAC 192. I used Oasis "Definitely Maybe" as the CD of choice. I used this CD because: 1) I can ABX the original at AAC 160, making AAC 192 a requirement for transparency, and 2) it is one of my more difficult CDs for LAME preset standard to encode at a resonable bitrate. With 3.90.3 alt preset standard my previous encoding average is in the 220 range. I think this must be due to the high frequencies prevalent throughout the CD.

The end result is that AAC 192 produces only a 6% size saving vs. 3.96.1 PS. The average bit rate acorss all songs using 3.96.1 preset standard is 205 kbps vs. (surprise!) 192 AAC. The bit rates vary from 222 to 183. 3 songs encode below 192, with the remainder all above 200. Both encodings are transparent at the aforementioned settings to my ears.

Therefore, for purposes of most of the rock music I listen to, I think I can reasonably conclude that AAC, as it stands TODAY, offers little advantage over LAME encoded MP3 if transparency, or very near transparency - some will contend they need lossless for true transparency - is a requirement. Maybe death metal offers a greater file size savings, but I do not listen to death metal.

Hopefully iTunes 5 will change this equation, but for now I think LAME MP3 is a better option for encoding my collection than 192 AAC. Am I missing anything?

~192 kbps lame/aac seems to be a rather high bitrate to use with the iPod.

Have you considered keeping both the lossless AND lossy files on the iMac?

Change of strategy

Doing that should allow you to decide on an optimal file size for the iPod, without sacrifying the quality you get on the computer or when burning CDs.

that's right! It even says it on Apple's website, that AAC will drain the battery faster! I have an iPod too and I use Sennheiser PX-100 headphones as well. I always use LAME 3.96 --preset standard ; overall, LAME 3.96 (--preset standard) is better than LAME 3.90.3 (--alt preset standard)...the only reason why LAME 3.96 isn't the recommended Lame version is, because there might be some issues in other modes other than --preset standard.

Lame 3.96, --ps averages around ~196kbps (based on my music-collection, including many different styles)...so I'd just use this setting since it gives you a longer battery-life on your iPod as well...

From Apple's site:

"When compared side-by-side, AAC proves itself worthy of replacing MP3 as the new Internet audio standard. Take a look at these AAC advantages over MP3:

...Improved decoding efficiency, requiring less processing power for decode"

This is just wrong.
Looking at the AAC and MP3 standards, it seems to me that AAC is absolutely nothing that could make it faster to decode, rather the contrary.

I haven't read anything that suggests AAC drains the iPod battery noticeably more or less than mp3. I have lots of AAC's and I still get the standard 8 hours battery life.

If you want long battery life, turning off the light and not changing songs often is probably the best thing you could do..

I'm pretty sure that I read on the Apple's Website that AAC drains the battery faster (I even have a post here somewhere about this). They must have removed it...however, here is what I found: http://docs.info.apple.com/article.html?artnum=61475



so, why would they suggest to use 160kbps mp3's and 128kbps AAC's to test the battery life? Shouldn't it be at least the same bitrate? I use mp3's on my iPod mini and I get a battery life of 10 hours...

Perhaps because those are the default bitrates?

3.96 PS was found to be no better or worse than 3.90.3 APS. 3.90.3 still has the most extensive testing, so even if 3.96 didn't have issues in other modes it still wouldn't be the recommended encoder version.

The 250 GB hard drive is big but not THAT big. 12,000 song files.

I'm probably something missing but the only advantage of 3.96 (.1) I see in encoding speed and that it consumes a little less space than 3.90.3...

That statement certainly contradicts the iPod's own speed changing algorithm, where I've seen it uses less MHz on average for mp3s (at any bitrate) than AAC at 128 kbps.

I don't think so. It's for the battery test and if you get less than 8 hours something is wrong with your battery...so I think apple just wants to be save...and as I said...it used to say it on their website that a 160kbps AAC file needs aprox. the same power as a 192kbps mp3 file...if you are bored you could run a test



from the FAQ:





and that is not enough? What else do you expect?

Yes, I agree. It seems to me that most of the power draw would be from spinning up the hard drive itself. At a higher bitrate, buffers would need to be filled more often, and so at higher bit rates more reads would take place. Power draw differences due to any processing differences in the formats almost seems negligable.

EDIT: I suspect this is a subtle ploy by Apple to push the AAC format by making it seem superior in every way to MP3.

from the FAQ:


Improved preset system means that it is easier to choose a quality level using the V switch. ( http://www.hydrogenaudio.org/forums/index....showtopic=18091 ) It doesn't mean that the encoding quality is improved over 3.90.3.

The decoding power is quite important. For small sizes, hard drive access is not that important. That is why AAC is using more power than mp3.

But for big files (is non compressed or lossless files) the files are huge. In this case, hard drive access become the most important factor regarding power use.

that's true, however, there was a thread were both, LAME 3.96.1 and 3.90.3 were compared and overall LAME 3.96 was better (--aps)...

Alright, show us the link to the thread where it is clearly proven that 3.96.1 is superior to 3.90.3...

Here are the two keywords of my post --aps and overall