HydrogenAudio

It will be interesting and maybe usefull to see what bitrates people use mostly for different generation codecs like MP3, AAC, Vorbis etc.


I understand I post this poll in wrong subforum as I haven't rights to create poll in Poll section. Please, don't remove this interesing poll.

I use FLAC mostly, but in order to get stuff playing and fitting on my iPod, I convert to MP3 V5. Used V4 but that didn't fit, and V5 I really cant hear the difference anyway.

It is maybe confusing to put MPC with AAC and Vorbis, I would use about 200+ with MPC but lower with AAC or Vorbis.

It is maybe confusing to put MPC with AAC and Vorbis, I would use about 200+ with MPC but lower with AAC or Vorbis.

I thought the same when I read the poll.

Anyway, my vote for lossless and Aoyumi's tuned ogg encodes @ -q 5 (~160 kbps) for portable.

The first question can be answered easily but not the second one. What's the definition of a "modern" codec (in this poll if you want)? Grouping AAC, Vorbis and MPC suggests these three are modern and similar and on the other hand indirectly suggests mp3 is not modern. And what to answer to the second question when using mpc at 250+ kbps and also aac at around 128 kbps? (Just an example). Sorry but I can't participate in this poll.

I use FLAC mostly, but in order to get stuff playing and fitting on my iPod, I convert to MP3 V5. Used V4 but that didn't fit, and V5 I really cant hear the difference anyway.

Yeah, me too. I rip all my CDs to FLAC but also encode them to LAME -V 5.5 (voted 5) and neroAacEnc -q 0.25 [~80 kb/s] for my portable devices.

MPC, Vorbis and AAc should be voted seperately.



I use
MPC: ~170kbps VBR
AAC: ~96kbps VBR

I'm surprised this thread hasn't been hosed by users upset about MP3 not being termed a "modern" codec. It's true that MP3 lacks some of the features of Vorbis, AAC, WMA, etc. but current original music releases (except DVD Audio releases, which exist but are vanishingly rare) generally don't take advantage of those newer features anyway ... just my opinion.

Voted V0 for MP3 and 320kbps+ for WMA (back when I still ripped to that codec, I no longer do)

Voted MP3 and V0. With my tin can ears I could probably get by with less but I'm a believer of always building in a little insurance. And with hard drive space so cheap and the iPod and Zune with models larger than my music collection I don't much care about storage.

 "MP3" vs "modern codecs like MPC"  I thought it's "no trolling" here?

However... Vorbis @ q6, MP3 @ 192 or 256 kbps (depends) ...

LAME V5.

I only use mp3 for my portable with settings -V5 for music and -V8 -mm for speech.
I use modern lossy codecs for movies. Nero AAC -q 0.35 (~96kbps).
For my main music library on my pc i use FLAC. In the past i used MPC -q5 (~180kbps) though.

I tend to want all my music to be the same bitrate, can't really come up with a good excuse why, just makes me feel warm and fuzzy inside.

With LAME I got to the point where I used -V 0, so it matched anything I bought off the Amazon MP3 Store. Recently though I purchased a Zune and a Zune pass, so I've been using 192kbps WMA, that way my ripped CDs and DRMed Zune pass content are all the same bitrate.

I can't ABX the difference with either option (-V 0 or 192K WMA), so it works for me.

I used aac for my nokia and the setting I use is vbr -q0.5.
all song sounds transparent to my ears....
i know i don't need to use that high but I'd love to be on the save side.... 

for my pmp, i used mp3, fhg vbr-160kbps average (about -v3 for lame...).

Lame mp3 -V3 (siggy could have told you that  )

And I do not use more modern codecs. I've never had a need to and don't record multi-channel audio.

Lame V0

I rip to AAC @ 320 VBR via Max (http://sbooth.org/Max/). It yields some pretty tasty bit rates...

MP3 -Vx or 320 kbps? That's all the options I get? Am I being banned from stating that I encode to MP3 @ 192 kbps CBR?

Btw, I'm planning to rip my music collection to AAC LC at a high bitrate of 256 or 320 kbps.

Chris

MP3 -Vx or 320 kbps? That's all the options I get? Am I being banned from stating that I encode to MP3 @ 192 kbps CBR?

Why do you use cbr??

After a very long exhaustive search for my sweet spot, I rip everything to LAME 3.97 -V5 (default; vbr old).  I like its default lowpass of 16kHz.  I don't archive my CDs to lossless.  If ever something bad happens to my original CDs, I'll just make do of what was left.

Clarity is in the ears of the beholder!  Contentment is the final step towards perfection. = )

My current settings are Wavpack Hybrid with the 160 kbps bitrate for most of what I encode, especially the master copies.
my longer website audio files are Ogg-Vorbis in the 60 kbps bitrate (-q -0.21) range, which works perfectly for older mono recording 60-70+ years old.
offtopic edit: almost to 200 posts!

Why do you use cbr??

Why not?? Doesn't matter. The question was, which bit rates to use for MP3 or modern codecs, not: which bitrates for LAME and modern codecs, and not: which VBR modes for LAME. Lame Ain't an MP3 Encoder, right?

Yes, LAME is a great encoder, but please don't forget there are others which do not have a -Vx switch (actually I don't even know by heart which bit rate range each switch maps to), and in some cases you might want/need to use CBR.

That's all I wanted to clarify.

Chris

Quote from: PaJaRo on 2009-06-09 23:05:54

Why do you use cbr??

Why not?? Doesn't matter.

Yes it does matter, since if you use vbr you will get either smaller files or higher quality files.

Yes, LAME is a great encoder, but please don't forget there are others which do not have a -Vx switch (actually I don't even know by heart which bit rate range each switch maps to), and in some cases you might want/need to use CBR.

Do you have any special reason to use a codec that does not support vbr?
What are the cases when you  might want/need to use CBR?

There are some cases where people come across hardware limitations.  Older DVD/home theater devices, car CD decks, etc. may only be compatible with CBR mp3 files.  These cases are rare these days and are more prominent in older hardware.  So it can happen if someone is still using older hardware (it works so why change it out?) and doesn't feel like upgrading.

My lossy music library is encoded with Nero's latest AAC encoder at -q0.5.  I found that the -q0.45 setting was generally transparent for my needs but I decided to go up to the next setting (in 0.05 intervals) as I listen to music in the metal genre.  I have found that Ministry and Nine Inch Nails songs often produce more artifacts (or is it artefacts?) with lossy encoders.  I didn't want to chance coming across an audible artifact in the future so I just went with -q0.5.

the second poll result looks like a bell curve to me:)

the second poll result looks like a bell curve to me:)

Yes that also struck me, the way the non-mp3 results seem to follow a simple "normal like" distribition whereas the mp3 distribution tends to cluster around three particularly popular levels (V5, V2 and V0). I suppose these more or less correspond to "portible", "standard" and "best" (vbr) quality respecitively.

V2

I don't use newer codecs.

---

According to the quality vs. file size chart on the wiki page for LAME, V2 is not much lower quality than CBR 320 but massively smaller file size. In fact, it seems to be the best you can do before you start sacrificing more quality for smaller file size reductions.

As for newer codecs, the advantages they offer are fairly dubious, not well supported in most hardware, and they seem to mainly be trying to lure you into more heavily patented codecs (WMA/AAC) or into a religious war over licensing. (Ogg Vorbis).

Really I don't care about newer codecs for those reasons, Lame is a great MP3 encoder, there's plenty of open source decoders, and all hardware supports it, MP3 is probably the most open and universally supported codec there is despite the few lingering patents that have already started expiring that only *some* countries recognize.

People that use "black box" codecs like WMA just bewilder me. 

As for newer codecs, the advantages they offer are fairly dubious, not well supported in most hardware...

I don't want to start a war or anything but this statement is technically false.  Why?  Most portable hardware (that is currently out there and being sold) is taken up by iPods and iPhones.  They are 100% AAC compatible.  Even Microsoft and Sony are jumping on the AAC board which means that only a few companies are left behind (mainly SanDisk as they don't support AAC in all of their portables and Insignia, Creative has made AAC compatible players for awhile now).  Now, once iPods stop taking up over 70% of the market share (ie less than 50%), then that statement will become true.  Additionally, support for AAC is growing in other hardware areas.  There are over a handful of DVD players at Best Buy that support them, many Blu-ray players, car CD decks, every current generation home console, and every current generation portable console (ie DSi and PSP) all support AAC files.

Sure, AAC may not offer the same "across the board" compatibility that mp3 does but I wouldn't say that the AAC community is being "dubious" especially considering the constantly growing range of devices that can playback AAC files.

I agree.  "Dubious" is a poor word choice.

Drawing conclusions from the wiki's chart about sound quality was not a very good idea either (a terrible one, actually).  That chart is simply intended to show that bitrate affects size and quality.  To what degree depends on what is being encoded and must be determined through blind testing from which the results will certainly vary depending on the individual.  Personally I think that chart should be scrapped.

EAC 0.99a5 -> Wavpack 4.50 lossless for home play and archiving CDs.

foobar 2000 -> LAME V3 VBR for the occasional portable trans-coding.

Quote from: Daemon7 on 2009-07-14 19:29:19

As for newer codecs, the advantages they offer are fairly dubious, not well supported in most hardware...

I don't want to start a war or anything but this statement is technically false.  Why?  Most portable hardware (that is currently out there and being sold) is taken up by iPods and iPhones.  They are 100% AAC compatible.  Even Microsoft and Sony are jumping on the AAC board which means that only a few companies are left behind (mainly SanDisk as they don't support AAC in all of their portables and Insignia, Creative has made AAC compatible players for awhile now).  Now, once iPods stop taking up over 70% of the market share (ie less than 50%), then that statement will become true.  Additionally, support for AAC is growing in other hardware areas.  There are over a handful of DVD players at Best Buy that support them, many Blu-ray players, car CD decks, every current generation home console, and every current generation portable console (ie DSi and PSP) all support AAC files.

Sure, AAC may not offer the same "across the board" compatibility that mp3 does but I wouldn't say that the AAC community is being "dubious" especially considering the constantly growing range of devices that can playback AAC files.

From what I've seen, the current Lame 3.99 tree has already tied up a lot of the loose ends and bugs left in 3.98.x.

MP3 is as resilient as it is because of exactly the reasons that "should" kill it off, you're right that it's not the "newest" or "best", but from what I've seen, the guys behind MPEG-4 AAC are being far more litigious over it than they have been over MP3, the patent issue is even more of a minefield, Apple still to the best of my knowledge tries to sell inferior quality files with proprietary DRM on their standard price level that won't work on anything *but* an iPod.

Also, every device supports MP3, and while technically AAC support is growing, there's not that many legal ways to buy files from online stores in it. I think iTunes is still the only ones selling it, and at a 30-40% pricing premium over what the typical music store charges per track if you want it in the same quality without DRM.

The fact remains that AAC is not widely accepted among users, that M4P may as well be a DRM'd WMA, that even Apple themselves admitted that 97% of the files on an iPod were MP3 files that didn't come from iTunes. (And I'd really like to know how they got those numbers without spying on their customers), and for whatever reason the market is still demanding MP3.  WMA, AAC, and Ogg Vorbis are fighting over the scraps, and in that order.

A true ISO-standard AAC file that was acquired legitimately is still very much a rarity unless you ripped them yourself, then you come across the fact that there are no good options for encoding them. You can choose FAAC which is sub-par, or you can use Nero or iTunes and depend entirely on the whims and licensing of the people that make that encoder. With WMA, there is only one thing that can encode them properly.

The fact is that the industry has made a damned mess, Microsoft has a totally non-standard black box format, Ogg Vorbis may very well violate patents in some countries, and there is no heir-apparent to MP3, only a tower of babel, with MP3 as a lowest common denominator.    Maybe if Apple, Microsoft, Real, the MPEG group, and the RIAA had been just a tad less greedy, instead of acting like sharks and trying to stab the end user in the back and then go after each other once there was blood in the water....

MP3 is as resilient as it is because of exactly the reasons that "should" kill it off, you're right that it's not the "newest" or "best", but from what I've seen, the guys behind MPEG-4 AAC are being far more litigious over it than they have been over MP3, the patent issue is even more of a minefield,

That's not true. AAC licensing is straightforward, it's a one stop shop.

Apple still to the best of my knowledge tries to sell inferior quality files with proprietary DRM on their standard price level that won't work on anything *but* an iPod.

Not true. DRM-free is the standard for quite a while. Perfectly spec compliant AAC.

Also, every device supports MP3, and while technically AAC support is growing, there's not that many legal ways to buy files from online stores in it.

Well just the biggest online music outlet in the world.

I think iTunes is still the only ones selling it, and at a 30-40% pricing premium over what the typical music store charges per track if you want it in the same quality without DRM.

0,99 per DRM-free track and 8,99 for an album is a 30-40% premium? Over what?

The fact remains that AAC is not widely accepted among users,

Got a reference?

A true ISO-standard AAC file that was acquired legitimately is still very much a rarity unless you ripped them yourself, then you come across the fact that there are no good options for encoding them. You can choose FAAC which is sub-par, or you can use Nero or iTunes and depend entirely on the whims and licensing of the people that make that encoder.

Well besides buying ISO compliant AAC at the iTMS directly, what is it that is so enslaving about Nero's free command line encoder? That it doesn't run natively on Puppy Linux?

And BTW, the reason I prefer AAC is the really high number of LAME problem samples, that could never be fixed, even at 320kbit/s. Because the devs should not be less talented than Nero's, which could fix about anything in the past, I conclude that the format itself has too many constraints to compete with modern AAC.

Both greynol and rpp3po pointed out some very good points that I won't go over again.  I would like to touch on this statement though.

The fact remains that AAC is not widely accepted among users, that M4P may as well be a DRM'd WMA, that even Apple themselves admitted that 97% of the files on an iPod were MP3 files that didn't come from iTunes. (And I'd really like to know how they got those numbers without spying on their customers), and for whatever reason the market is still demanding MP3.  WMA, AAC, and Ogg Vorbis are fighting over the scraps, and in that order.

Do you have any links to news articles, Apple quotes, and studies (that aren't older than 2 years) to backup these statements?  It just comes across, to me, that you have these negative grudges against WMA, AAC, OGG Vorbis, etc.  You pretty much dislike anything that isn't mp3 yet you are either relying on false information, opinionated articles from 2003, or just don't care.  I give you this graph:
(http://img43.imageshack.us/img43/5915/ipodsalesquarterlylg.gif)

Those are quarterly iPod sales up through March 2008.  It is outdated but it represents the sales of new iPods.  Apple sold a little over 21 million units, world wide, in the month of December in the year 2006.  They then sold over 22 million units during December 2007.  That is over 43 million iPods that they sold.  Yeah, that doesn't look like wide adoption of the AAC format.  Most people actually download iTunes, plug in their iPod, and start ripping CDs without changing settings.  They don't realize that you can change audio import settings.  The average person just wants things to work and they don't focus on the small details such as AAC, Lame mp3, WMA, FLAC, Apple Lossless, whatever.  The majority of the portable audio community uses AAC files whether they intend to or not.  Case in point, Apple has sold over 6 billion songs (http://en.wikipedia.org/wiki/ITunes_Store#Market_share_and_milestones) since its launch back in 2003 and is the number one music retailer (beating out CD sales from Wal-Mart, Target, Amazon.com, Best Buy, ANYONE!) in the U.S.

Nope, people just aren't buying 4 million + songs a day in DRM-free AAC files.  Nope, they sure aren't adapting the format.  I don't mean to sound so negative but the iTunes Store sales alone are enough to show that people are in fact adopting the mpeg-4 AAC audio format with the majority players being sold in stores (and sitting in homes), across the world, working with AAC files.

Edit: I am not trying to say that Lame mp3 (or mp3 in general) is teh sucksorz or anything like that.  I am just trying to get across that AAC support has drastically grown and that this is mainly due to high iPod sales and the iTunes Store becoming the number one music retailer in the U.S. (while also being extremely popular in other countries).  AAC encoders are no longer these "black boxes" that you speak of.  iTunes gives people a bunch of different way to use its AAC encoders (not as much as QuickTime Pro but you get the picture) and Nero even offers a free, high quality, command-line AAC encoder.

Quote from: Daemon7 on 2009-07-15 22:49:41

MP3 is as resilient as it is because of exactly the reasons that "should" kill it off, you're right that it's not the "newest" or "best", but from what I've seen, the guys behind MPEG-4 AAC are being far more litigious over it than they have been over MP3, the patent issue is even more of a minefield,

That's not true. AAC licensing is straightforward, it's a one stop shop.

Quote from: Daemon7 on 2009-07-15 22:49:41

Apple still to the best of my knowledge tries to sell inferior quality files with proprietary DRM on their standard price level that won't work on anything *but* an iPod.

Not true. DRM-free is the standard for quite a while. Perfectly spec compliant AAC.

Quote from: Daemon7 on 2009-07-15 22:49:41

Also, every device supports MP3, and while technically AAC support is growing, there's not that many legal ways to buy files from online stores in it.

Well just the biggest online music outlet in the world.

Quote from: Daemon7 on 2009-07-15 22:49:41

I think iTunes is still the only ones selling it, and at a 30-40% pricing premium over what the typical music store charges per track if you want it in the same quality without DRM.

0,99 per DRM-free track and 8,99 for an album is a 30-40% premium? Over what?

Quote from: Daemon7 on 2009-07-15 22:49:41

The fact remains that AAC is not widely accepted among users,

Got a reference?

Quote from: Daemon7 on 2009-07-15 22:49:41

A true ISO-standard AAC file that was acquired legitimately is still very much a rarity unless you ripped them yourself, then you come across the fact that there are no good options for encoding them. You can choose FAAC which is sub-par, or you can use Nero or iTunes and depend entirely on the whims and licensing of the people that make that encoder.

Well besides buying ISO compliant AAC at the iTMS directly, what is it that is so enslaving about Nero's free command line encoder? That it doesn't run natively on Puppy Linux?

And BTW, the reason I prefer AAC is the really high number of LAME problem samples, that could never be fixed, even at 320kbit/s. Because the devs should not be less talented than Nero's, which could fix about anything in the past, I conclude that the format itself has too many constraints to compete with modern AAC.

http://www.tuaw.com/2009/04/13/billboard-i...-up-sales-down/ (http://www.tuaw.com/2009/04/13/billboard-itunes-prices-up-sales-down/)

"The iTunes Top 100 chart has 40 different songs with a new price of $1.29, and one day after the changes, those songs dropped an average of 5.3 places on the chart, while cheaper songs moved up on average. And on the second day of the price change, ten of the tracks that saw their prices rise within 24 hours dropped a huge 12.4 chart positions on average."

The last time I used iTunes was in like 2004, back when they *only* offered crippled 128-bit CBR, non-standard AAC, but I have played around with Nero. Anyway, last I heard the normal pricing for an iTunes track which wasn't crippled was $1.29, with 99 cents for crippled files being their norm. I just won't have it on my computer for a multitude of reasons including the fact that I've seen spyware infestations that didn't damage Windows as badly. (Or start as many system services)

I can't really tell the difference between a Nero AAC file and a Lame MP3 file of the same bitrate (CBR) or around the same size (VBR) until you get to ridiculously low bitrates and Nero HE-AAC kicks in and goodbye to most of the hardware that plays AAC.

AAC for me just kind of seems like a headache to fix a few trivial design oversights in the MP3 format. If you *have* to pull out a spectrogram to tell the difference between two files, two formats, two encoders, or two bitrates, chances are that you've gone into overkill somewhere.

By my estimation, precisely every legal store that isn't iTunes isn't selling AAC anyway. AAC is what? Almost 10 years old now and it has one backer?

Both greynol and rpp3po pointed out some very good points that I won't go over again.  I would like to touch on this statement though.

Quote from: Daemon7 on 2009-07-15 22:49:41

The fact remains that AAC is not widely accepted among users, that M4P may as well be a DRM'd WMA, that even Apple themselves admitted that 97% of the files on an iPod were MP3 files that didn't come from iTunes. (And I'd really like to know how they got those numbers without spying on their customers), and for whatever reason the market is still demanding MP3.  WMA, AAC, and Ogg Vorbis are fighting over the scraps, and in that order.

Do you have any links to news articles, Apple quotes, and studies (that aren't older than 2 years) to backup these statements?  It just comes across, to me, that you have these negative grudges against WMA, AAC, OGG Vorbis, etc.  You pretty much dislike anything that isn't mp3 yet you are either relying on false information, opinionated articles from 2003, or just don't care.  I give you this graph:


Anything that isn't MP3 has ultimately had the fate of becoming an "also-ran". People tend to choose MP3, people that don't know what the hell they are doing buy whatever the store sells them.

As for WMA, I just don't like the idea of "Put your data in....*magic*...out pops your file.", MP3's backers never went to lengths to hide how the format works. (AAC also more open than WMA, but with FoulPlay DRM that's a whole different story)

Why are you guys even going at it with this guy?  It's getting lame quick. 

And the MP3 settings should've been dropped from this poll.  We do one every year.  No need for a mid year "what V setting do you use" poll.

A strictly AAC bitrate poll would've been interesting, and without the "I don't use modern codecs" option.  I don't care what you don't use.  To know how the people who exclusively use AAC, use AAC, would be interesting.

Hardware support for AAC is growing mainly because of Apple's support, which kornchild2002 mentioned above and also its support by major mobile phone manufacturers.

But IMO i think most of the average iPod user's music is likely to be transcoded from 128kbps WMA to 128kbps AAC, by migrating from WMP to iTunes.

ATM Mp3 is still a very competitive codec at mid and high bitrates e.g LAME V2 - V0 due to it being more mature then AAC. IMO i can see future AAC encoders performing alot better, while Mp3 encoder devs are stuck with design flaws i.e. block usage and sfb21.

"The iTunes Top 100 chart has 40 different songs with a new price of $1.29, and one day after the changes, those songs dropped an average of 5.3 places on the chart, while cheaper songs moved up on average. And on the second day of the price change, ten of the tracks that saw their prices rise within 24 hours dropped a huge 12.4 chart positions on average."

If you're going to use quotations to make a point, be sure they actually make your point (bolding mine).

AAC for me just kind of seems like a headache to fix a few trivial design oversights in the MP3 format. If you *have* to pull out a spectrogram to tell the difference between two files, two formats, two encoders, or two bitrates, chances are that you've gone into overkill somewhere.

Who said anything about having to look at a spectrogram?  What makes you think the shortcomings of the MP3 format are only trivial?

No, and it doesn't *matter* how many devices can play AAC if nobody is using it.

That's a big "if".  I think notion that nobody is using AAC has been thoroughly discredited.

I'm assuming that since most of the less than legal music floating around, and all the music in less expensive stores that don't hose Windows is all in MP3 format, that anything out of the 97% of music that lands on the "average" iPod that is not purchased from Apple, and that the user did not encode with iTunes, is not AAC.

How much of that number do you thing is music that was encoded with iTunes?

Apple marketing likes to play numbers games

They're not the only ones. 

Hardware support for AAC is growing mainly because of Apple's support, which kornchild2002 mentioned above and also its support by major mobile phone manufacturers.

But IMO i think most of the average iPod user's music is likely to be transcoded from 128kbps WMA to 128kbps AAC, by migrating from WMP to iTunes.

ATM Mp3 is still a very competitive codec at mid and high bitrates e.g LAME V2 - V0 due to it being more mature then AAC. IMO i can see future AAC encoders performing alot better, while Mp3 encoder devs are stuck with design flaws i.e. block usage and sfb21.

Well, in 2000 an AAC encoder would give you horrible quality, in 2009 it's on par.

You could argue that sfb21 bloat is a problem with MP3, and you'd be right. It's becoming less of one as storage expands. I could maybe buy into this particular argument if you were talking about those little 4 gigabyte "chewing gum pack" $30 MP3 players (or the same thing from Apple for about $90, minus a screen), but I really don't see your point when a decent hard drive player has almost 200 gigabytes and nothing you ever do will fill it. Hell, even flash players have 32 gigs now and a micro-sdhc expansion port. At any rate you're still arguing about how many weeks or months worth (if played nonstop) of music you can pack onto a device that will be obsolete next year by a model with twice the space that costs half as much. If you did skimp on storage, the -y switch should unbloat the file quite nicely at the expense of a very small amount of quality.

It's getting to the point where file size matters less and less as codec makers are still brutally trying to compete to be the world's tallest midget at absurdly low bitrates. I think that when MP3 has overstayed its welcome, it will be because storage has gotten to where you will think nothing of using lossless formats. I'd expect lossy codecs in total have maybe another 5-6 years left in them, further I have to seriously doubt the relative sanity of anyone who actually buys any lossy files at CD prices.  To each their own.

Frankly, what concerns me more than anything is how people trust binary-only encoders and decoders that are put out with extremely restrictive terms of use. (And especially with DRM, nothing can guarantee that they'll even be able to get any of their data back out later.)

You could argue that sfb21 bloat is a problem with MP3, and you'd be right. It's becoming less of one as storage expands. I could maybe buy into this particular argument if you were talking about those little 4 gigabyte "chewing gum pack" $30 MP3 players (or the same thing from Apple for about $90, minus a screen), but I really don't see your point when a decent hard drive player has almost 200 gigabytes and nothing you ever do will fill it. Hell, even flash players have 32 gigs now and a micro-sdhc expansion port. At any rate you're still arguing about how many weeks or months worth (if played nonstop) of music you can pack onto a device that will be obsolete next year by a model with twice the space that costs half as much. If you did skimp on storage, the -y switch should unbloat the file quite nicely at the expense of a very small amount of quality.

It's getting to the point where file size matters less and less as codec makers are still brutally trying to compete to be the world's tallest midget at absurdly low bitrates. I think that when MP3 has overstayed its welcome, it will be because storage has gotten to where you will think nothing of using lossless formats. I'd expect lossy codecs in total have maybe another 5-6 years left in them, further I have to seriously doubt the relative sanity of anyone who actually buys any lossy files at CD prices.  To each their own.

The main problem is that everyones music collection grows, and most dap customers do not want to upgrade their daps every year.

There is alot of people that might a want a flash based dap (e.g iPod Nano or Touch) to carry around their whole music collection instead of a hard drive based dap. Which atm flash storage still playing catch up with size and it's still more expensive then hard disc storage. Also if everyone adopts lossless for music, then theres still be some good use with lossy codecs for streaming media to save bandwidth.

I have to admit paying full CD price for a lossy album is pretty lame, but it does the job for very rare albums that you cannot find.

Frankly, what concerns me more than anything is how people trust binary-only encoders and decoders that are put out with extremely restrictive terms of use. (And especially with DRM, nothing can guarantee that they'll even be able to get any of their data back out later.)

You are running in circles. You cannot even buy DRM'ed AAC anymore. And what's DRM got to do with AAC, anyway? Is paper bad because you can print toxic securities on it? And you are still owing an explanation what those "extremely restrictive terms of use" would be. Freely encode what you want as much as you want for personal or buy a license for commercial use? Yes, what a rip-off!

You could argue that sfb21 bloat is a problem with MP3, and you'd be right. It's becoming less of one as storage expands.

It's not mainly a bloat problem*. It's a quality problem. You could have noticed that already just by reading this thread.

We have just recently discussed a sample here, where LAME wasn't even able to encode a single wind instrument without artifacts up the the maximum bit rate. Nero was able do it transparently < 200kbit/s. There's no single known sample, yet, where it would be the other way around and MP3 beat (Quicktime or Nero) AAC .


* The efficiency gain comes free.

And you are still owing an explanation what those "extremely restrictive terms of use" would be. Freely encode what you want as much as you want for personal or buy a license for commercial use? Yes, what a rip-off!

No, in WMA's case it's more like "Encode as much as you want, all you want (and then find out that playback on anything that's not Windows requires reverse engineered codecs,  maybe can't be burned, maybe can't be transferred to a DAP, and you certainly can't make more of the damned things without Windows. But maybe you're just clueless now and we can spring this on you when you've had enough of Windows.)"

After that argument, every other gripe about WMA is pretty much gravy.

As for AAC, you require a codec that says you can encode as much as you want, all you want, and if you play it back at a nightclub or something then you have to pony up. The way the encoder operates is a secret, and the only open source AAC encoder is light years behind. (Which is definitely a bigger problem than just ignoring the personal use only legalese)

With MP3, the personal use only crap still applies (again, you can ignore it easily), but both encoding and decoding have high quality open source and publicly documented implementations.

Ideally yes, Ogg Vorbis would come save us from the personal use crap (every other format) and the "black box codec" (WMA and to a much lesser extent AAC) crap, but since Microsoft and Apple refuse to support it, it's going nowhere.

My argument for MP3 has never been that it was the best, it's a least common denominator with acceptable quality. It beats having your data held hostage in a format that is not well understood or not widely implemented.

"Yes, go ahead, use it all you want, pile up GIGABYTES of data in our format, we'll turn things around on you later" It's kind of easy to see the "credit card mentality" of proprietary file formats.

We have just recently discussed a sample here, where LAME wasn't even able to encode a single wind instrument without artifacts up the the maximum bit rate. Nero was able do it transparently < 200kbit/s. There's no single known sample, yet, where it would be the other way around and MP3 beat (Quicktime or Nero) AAC .

I'm beginning to have my doubts that many such samples actually exist. At least every so-called "mp3 killer" sample I've ever ran across sounded pretty much like the source file when I've encoded it. Most of the "mp3 killer" samples just means your particular encoder is not all that great and needs tweaking. LAME has been notorious for some tracks that other MP3 encoders (or later versions of LAME)have handled fine.

I could show you plenty of things that AAC encoders used to mutilate, still do (FAAC), or need a higher bitrate than the MP3 for. I've never blamed the bitstream format though.

There's been plenty of worse things done to audio than MP3, cassette tapes and vinyl records spring to mind. I still facepalm when I see where someone has encoded a vinyl record, made 100 megabyte tracks in FLAC format, and you go to play them and it's like popping and clicking, and the record has obviously been in a 100 degree attic for 30 years, and they still insist they need FLAC because the record is better than a CD.

Back to the subject of sfb21, most MP3 encoders, including Helix and the "Fruenhofer IIS" official MP3 encoder both just throw away *everything* in that band just to avoid having to deal with it. You have to opt-in to have it even attempt to mess with it. Kind of shows you how important "they" think it is.

LAME is the only MP3 encoder that I've seen that really wants to have anything to do with sfb21 in VBR mode, and I'm wondering if there's any rhyme or reason to it, or if it's more of a "Shot in the dark, probably wasting bits" thing.

FLAC for archives, iTunes True VBR @ q100... yields around 135-140 kbps (I voted 160).

Flac for archiving and Nero true VBR AAC q 0.95 for everything else (listening on pc,ipod,mobile phone)

No, in WMA's case it's more like "Encode as much as you want, all you want (and then find out that playback on anything that's not Windows requires reverse engineered codecs,  maybe can't be burned, maybe can't be transferred to a DAP, and you certainly can't make more of the damned things without Windows. But maybe you're just clueless now and we can spring this on you when you've had enough of Windows.)"

Nobody is advocating WMA here. Do you even read the other posts in this thread? Unlike AAC it is a proprietary codec. One could get the impression that you have to resort to that, because you realize what BS you were serving up the whole time regarding AAC.

The way the encoder operates is a secret, and the only open source AAC encoder is light years behind. (Which is definitely a bigger problem than just ignoring the personal use only legalese)...

With MP3, the personal use only crap still applies (again, you can ignore it easily), but both encoding and decoding have high quality open source and publicly documented implementations.

AAC encoding and decoding is not a secret, it's a well documented standard. Just like MP3. LAME is just an implementation, no additional public documentation. There are also high quality open source decoders for AAC. You can't blame the standard that FAAC isn't taking off in face of such excellent free encoders.

You can use LAME binaries without having paid for a license and you can use the binary-only encoders without a license. The evil AAC empire is not going to erase the en- and decoders remotely from your disc. LAME's open sourceness has got nothing to do with the licensing situation.

Ideally yes, Ogg Vorbis would come save us from the personal use crap (every other format) and the "black box codec" (WMA and to a much lesser extent AAC) crap, but since Microsoft and Apple refuse to support it, it's going nowhere.

Again, AAC is not a black box codec! You somehow confuse open standards with the availability of one implementation's source code. Ogg Vorbis is an idealistic project. Nobody is there to guarantee that it is patent free. It's just supposed to. It can make more sense for a company to drop some money for licensing in a thoroughly researched patent environment.

My argument for MP3 has never been that it was the best, it's a least common denominator with acceptable quality. It beats having your data held hostage in a format that is not well understood or not widely implemented.

Yes, people often project personal feelings onto stuff like that. "I rather stick to what I know and what is good enough than falling over and over again for the newest fancy!" kind of things... Nobody is going to hold your property hostage with AAC, nobody is able to. Talk about paranoia. At least not a single bit more than with any other format.

I'm beginning to have my doubts that many such samples actually exist. At least every so-called "mp3 killer" sample I've ever ran across sounded pretty much like the source file when I've encoded it. Most of the "mp3 killer" samples just means your particular encoder is not all that great and needs tweaking.

Well some people may have better ears than others. Plenty of people at HA do hear differences and post ABX logs.

I could show you plenty of things that AAC encoders used to mutilate, still do (FAAC), or need a higher bitrate than the MP3 for. I've never blamed the bitstream format though.

There we are again, in the good old times, when everything was better. Present just one sample that LAME would handle better than Quicktime or Nero with current versions. You'll have a very hard time. The other way around there are plenty.

Back to the subject of sfb21, most MP3 encoders, including Helix and the "Fruenhofer IIS" official MP3 encoder both just throw away *everything* in that band just to avoid having to deal with it. You have to opt-in to have it even attempt to mess with it. Kind of shows you how important "they" think it is.

That's flawed reasoning. Something that's fucked up beyond repair might not be worth the time spent commercially. Nobody ever said that the >16k band was unimportant. Much music I own sounds muffled when you low-pass it at 16k.

Ogg Vorbis is an idealistic project. Nobody is there to guarantee that it is patent free. It's just supposed to. It can make more sense for a company to drop some money for licensing in a thoroughly researched patent environment.

Monty claimed that Ogg Vorbis is throughly researched by patent lawyers. Also, AAC is not immune to  submarine patents either. Witness what happened to the html standard in the recent Eolas case.

Quote from: rpp3po on 2009-07-16 11:56:38

Ogg Vorbis is an idealistic project. Nobody is there to guarantee that it is patent free. It's just supposed to. It can make more sense for a company to drop some money for licensing in a thoroughly researched patent environment.

Monty claimed that Ogg Vorbis is throughly researched by patent lawyers. Also, AAC is not immune to  submarine patents either. Witness what happened to the html standard in the recent Eolas case.

But why would anyone need submarine patents on AAC when there's an entire surface fleet? 

No, and it doesn't *matter* how many devices can play AAC if nobody is using it.

The same can be said for ANY format.  My point was to show you that there are plenty of devices out there that support AAC.  Additionally, as I previously stated, most people don't change the CD import setting for iTunes.  They simply download the program, rip their CDs, and plug their iPods in.  This means that they are using the AAC format without know.  The quote from Jobs was in the beginning of 2007 and he was talking about their data from the year 2006.  4 billion songs have been sold from 2007-2008 so his original numbers no longer apply.  Additionally, the iTunes Store has become the #1 music retailer in the U.S. since then.

I'm assuming that since most of the less than legal music floating around, and all the music in less expensive stores that don't hose Windows is all in MP3 format, that anything out of the 97% of music that lands on the "average" iPod that is not purchased from Apple, and that the user did not encode with iTunes, is not AAC.

Well, we all know what happens when people assume things.

So if they sell 500 billion iPods and 10 billion AACs, but 100 billion MP3s get on them "somehow", AAC is still comparatively tiny. Apple marketing likes to play numbers games because if you list numbers rather than percentages, it looks bigger. (And this is with me *trying* not to facepalm over that stupid "Don't Steal Music" sticker on every new iPod)

Apple has sold over 206 million iPods and over 6 billion songs.  That means that there are definitely more AAC files on iPods (and computers) than you like to think.  Additionally, Apple isn't the only company who likes to play the number game.  Pretty much every company out there does it.

I'm surprised that Jobs managed to go to percentages to show that almost nobody buys more than 1 or 2 albums before they give up on iTunes. That's a pretty big flub.

When did he do this?  Are you talking about that quote (which was not linked to) from back in 2007 before iTunes sold over 6 billion songs and became the number one retailer in the U.S.?  You just can't use older numbers as the iTunes Store continues to sell over 4 million songs per day.  That is no small feat and, with each purchase, the mpeg-4 AAC format is gaining in distribution/amount.

Anything that isn't MP3 has ultimately had the fate of becoming an "also-ran". People tend to choose MP3, people that don't know what the hell they are doing buy whatever the store sells them.

What people are you talking about?  Over 6 years of experience on a different website tells me that people, in general, tend to just go with what works.  They will buy an iPod, download iTunes, rip their CDs, download songs, and move on.  They don't care so long as it works.  The same goes for people who buy other DAPs.  They will use the software that comes with them without really changing settings.  They don't care, they just want things to work.

As for WMA, I just don't like the idea of "Put your data in....*magic*...out pops your file.", MP3's backers never went to lengths to hide how the format works. (AAC also more open than WMA, but with FoulPlay DRM that's a whole different story)

I guess that is why we don't have open source WMA decoders and even an encoder or two.  We just don't know how WMA works and must rely on Windows Media Player for everything...

Also, with your post above mine, I hope that you aren't trying to imply that mp3 doesn't have a series of patents surround it while other codecs do.

I guess that is why we don't have open source WMA decoders and even an encoder or two.  We just don't know how WMA works and must rely on Windows Media Player for everything...

If you want no decoding support for WMA Pro 9 or 10, no decoding support for WMA lossless, and a poorly hacked together ffmpeg encoder that produces a "valid WMA bitstream" that sounds like utter crap and is missing pretty much all the features of the MS encoder, then yes, you have *marvelous* WMA support outside of Windows. 

it's amazing how a thread could start off asking what each user's preferred format/bitrate is and somehow manage to turn into a heated debate between those loyal to a dated format and those who prefer to use a newer, more efficient one... and yes, I am referring to MP3 and AAC, in that order.

but since the debate came up, my $0.02- if I'm going to take the time to convert my entire music library to a portable, lossy format, the first thing on my mind is which format will allow me the highest quality vs. filesize... and considering I'm in the Apple camp and use a device with full capability of playing back both MP3 and AAC (i.e. an iPod), it doesn't seem all that advantageous to use MP3 when in my experience the quality gains of using AAC at a portable filesize/quality far outweigh any compatibility-based advantages to MP3.  don't get me wrong, I'm all for listening to arguments toward either side, but I have to say I'm a little annoyed that I opened this thread wanting to read about people's preferred formats/bitrates and instead had to trudge through a slew of arguments about them just to read a few posts actually related to the original topic.

meanwhile, back to the topic at hand... I've been using QuickTime AAC (True VBR) for just under a year now, and while I've gone back and forth between -q 95 and -q 120, I've eventually settled on -q120 since I was able to ABX roughly 5% of the files encoded @ -q 95, whereas everything @ -q 120 is completely transparent to my ears.  plus it's still moderately portable with an average bitrate of ~163kbps (my library, at least), so that's a perfect quality/filesize point for me.

it's amazing how a thread could start off asking what each user's preferred format/bitrate is and somehow manage to turn into a heated debate between those loyal to a dated format and those who prefer to use a newer, more efficient one... and yes, I am referring to MP3 and AAC, in that order.

It was never a debate regarding mp3 vs. AAC.  This thread turned sour whenever Daemon7 started to spread false information (and bad logic) regarding mpeg-4 AAC audio.  My replies were not to say that mp3 sucks as a modern format, it was just to disprove Daemon7's statements.

Funny that after all of that, he decided to pick out my open-source WMA statement instead of actually examining my statements regarding AAC adoption and AAC hardware that is out there.  They also managed to completely ignore my iTunes Store sales numbers which are a lot newer than the 3 year old ones (which they never linked to).  Oh well, I guess that is why they haven't really come back.

This happens a lot at hydrogenaudio where someone will ask a simple question and another person will stir things up.  The end result will be a few pages of debate going back and forth between both sides (or, in this case, one person and a slew of others).

- Flac for archiving

- Otherwise Lame V0 or Nero AAC Q0,60 (Q1.0 for Classic -my mobile player lacks flac support)

- FLAC for archiving.
- MP3 V4 for my portable player (transparent in most cases)

This question is a bit hard to answer, because  currently  I use  FLAC and convert to whatever is my mood when transferring to my mp3 player.  If I don't have a lossless version, I try to  have at least mp3 vbr with profile v2 or if cbr at 160kps (or I delete from my hard drive).  i avoid "modern codecs", unless I transfer to my ipod, that supports aac.

- For archiving I use Flac -8 (could be 5, too; but my machine is fast enough to have a few more bytes saved)
- For use on mobile players I encode to mp3 with ABR 160..192 and switch -q0. I found this to deliver the optimum quality with a reasonable size.
  (See also this post: Posting in forum about Lame (http://www.hydrogenaudio.org/forums/index.php?showtopic=66649&view=findpost&p=662072).)
Peace Gene

FLAC for archive here too.

For modern codecs I chose 256 kbps, only cos I occasionally get stuff from iTunes, so I use that sometimes as well I guess!

But I encode to mp3 (with Lame v0) for my iPod archive. I can honestly say I'm too lazy to explore AAC any further, and as I prefer to use Linux, mp3 is just easier. 

wavpack for archive/PC library.

nero aac ~125kbps for the ipod.

I rip to flac and listen to my flac's on the computer but I encode to LAME mp3 v0 for my Creative ZEN and other portables.

320 kpbs CBR mps.  Cos I'm insane.  :-)

All listening is on a HTPC over a reasonable amp and speakers, so storage not too much of an issue and may as well get the near-CD quality. 

Meanwhile, working on sorting out how to play and tag flac on windows media centre.  Once I sort that out will prob switch to flac V5 or something for the rolling ripping programme (this seems to be one of those background tasks that's never complete).  Meanwhile, of course, there are still many 128kbps MP3s in the collection from my younger days when it all needed to be ripped quickly prior to emigration and being without CDs for a few months...

320 kpbs CBR mps.  Cos I'm insane.  :-)

All listening is on a HTPC over a reasonable amp and speakers, so storage not too much of an issue and may as well get the near-CD quality. 

Meanwhile, working on sorting out how to play and tag flac on windows media centre.  Once I sort that out will prob switch to flac V5 or something for the rolling ripping programme (this seems to be one of those background tasks that's never complete).  Meanwhile, of course, there are still many 128kbps MP3s in the collection from my younger days when it all needed to be ripped quickly prior to emigration and being without CDs for a few months...

OK... so after posting that, I had a bash at ABX-ing (in foobar) original CD against -V0, -V2 and -V4.  11/13 with -V4, but without even getting past the first one, I knew I was guessing for -V2 and (obviously) -V0!

So... time to actually think about the encoding strategy.  In my defence... I only discover HA about a fortnight ago, signed up 2 days ago, and began to actually think about these issues - up to now it's just been getting stuff working that has preoccupied me (media center incl HDTV recording, ripping vinyl, cover art and so on - all in a situation where the family expect it to just work!)

FLAC for new CD Rips since I don't have the space concerns any more that I had a couple years back.

When I do use MP3, it's V9 (@32khz ~ 75-96kbps for my little cheapie portable, or V7 for MP3 CDs for my car) using lame 3.93.1 -k --nspsytune

I can't group the "modern" codecs under the same answer:
-For videos, and the music on my laptop I use vorbis q0 (no lowpass).. so ~64-70kbps.  I also use it for multichannel ~260kbps.
-I don't use AAC at all, but for stereo I would be using something around 115kbps if I had the need to use it.

So in the poll, I picked the two lowest bitrate options. 

I tend to want all my music to be the same bitrate, can't really come up with a good excuse why, just makes me feel warm and fuzzy inside.

It's called OCD, and we all do it a little

LAME v0 here.

Ogg on occasion to play around with it.

V2 for most material, V6 for some for which quality isn’t as important.

I wish the big players would support Vorbis – I love the concept and am really happy with its performance around 128kbps… I just don’t have any use for it and most of my friends wouldn’t know how to get Vorbis files to play.

I mostly use V0 or 320
and FLAC on modern codecs

I put down Lame -V3 for my portable mp3 player.  At work, I use Vorbis q0.0 (64kbps) because I would otherwise fill up a computer with music and they wouldn't like that .  Plus I've only a Dell 2.0 speaker system that really can't do justice to any music, so why waste space when vorbis is good enough at that bitrate given the equipment I have to listen to it on?  Win-win for me I'd say.

I have never used MP3 in my life. From the very start it was WMA, then OGG, and now I'm using the HE-AAC encoder in Winamp to rip CDs to .m4a files, always 64kb/s, and it suits me just fine
For my MP3 player, I still use OGG ~96kb/s

FLAC -8 / TAK -p4m on PC

LAME 3.98.2 -V2 / NERO AAC 1.5.1.0 -q 0.55 / OGG libvorbis 1.2.3 -q5 (if supported)  for portable devices

FLAC -5, some in -8 for archiving purpose (Ripped with EAC).

For audio player portable : LAME 3.98.4 320 CBR or NeroAAC 1.5.4 Q0.79.

Archiving / Transcoding: Wavpack lossy @ 415 k -fx3  (with correction file for very rare albums)

Portable: GOGO @ 160k

LAME V1 forced stereo (unless i think the music is particularly synthy or distorted). my forced stereo bias has been reinvigorated due to hooking up an old Quad Sansui (QRX-3500) i rescued from mum's basement of dust-gathering objects and listening to it synthesizing the surround as well as Windows Vista doing it (opposed to the classic Winamp 5.1 output plugin). (the primary use for surround was for gaming. and maybe movies)

and speaking of movies, AAC 128kbps is what i use for most (x264) "encodes of minimal size". i'd probably use higher (or even original audio) if i were shooting for visual transparency in an encode.

Nowadays, I use AAC with VBR ~256 kbps (on average) for just about everything, because compatibility isn't a concern to me and neither is the space taken by ~256 kbps music.

Mostly I use Apple Lossless,but occasionally I encode MP3 or AAC (mostly MP3) and then I use 320 kbps. Space is not really an issue anymore, so why not go for the best - even if the difference between 256 (vbr) and 320 is very small.

I think that ~200 kbps is best choice for psychoacoustic lossy encoding, it gives acceptable quality + reasonable filesize.

For my music I used:

AAC: Nero AAC Encoder 1.5.4.0  -q 0.55 or QTAACEnc 20100307 --tvbr 90 --highest
Vorbis: OggEnc libvorbis 1.3.1 -q6
Musepack: mpcenc 1.30--quality 5 (--standard)

or if I need best compability I use LAME 3.98.4 -V 2
All this codecs/presets give ~200 kbps at average for my kind of music.

And as I hear (I spent many ABX tests on Trance music), the best codec for my ears/equipment/music is Musepack standard - I can't hear any audible artefacts on any problem (for other formats) samples. I'm sure it's because the Musepack doesn't have pre-echo effect that I hear clearly with all other codecs on my killer samples....

My first post !!!   
Well... I use Flac for Archiving  thru EAC
I was using Mp3 v0 for my Sansa Fuze But recently switched to Vorbis Ogg at about 192.... because it seems i get better battery life AND less artifax MP3 will in time grow old and fade...like 8 tracks!!!(tapes) Guess there will always be die hards though.. I wish EVERYTHING supported VORBIS and Everyone. I mean whats not to like.....Can YOU HEAR a difference at v5-> ???? I use V6 just to be sure but if I had someone tell me that the difference between V5 and V6 is nil I may drop down to save alittle space...

oggenc -q 4,5 --advanced-encode-option impulse_noisetune=-15

i use V0 mp3 or v2 sometimes , because i think that this bitrate has the best analogue between quality - size. i still prefer mp3 because the most players are compatible with them.

Nowadays i use FLAC -8. Extracting takes more time than encoding, so i see no reason to use -5. For my car and portable i reencode my flacs to LAME -V4. For audio in movies i use Nero aac 0.35 (with h264 CRF 18 for video). And I encode my jpegs in q86 without subsampling .

loosy, using nero aac under 128kbit/s or aoTuV above 128 and flac mode 4 (normal) for ripping

MP3 V0 & Flac -8

I use mainly MP3 because of it's compatibility but also use OggVorbis once in a while. Both on the highest quality possible...

MP3 (LAME) = 320kbps CBR
Vorbis (aoTuV) = -q10 ≈ 500kbps

But I try to use lossless where possible (I even buy lossless music) and transcode them (where needed) to FLAC -5

Vorbis (aoTuV) = -q10 ? 500kbps

That's really overkill. Vorbis isn't optimized for such high bitrates, so you probably only waste a lot of bits without gaining any audible benefits over, say, -q6. You should go for lossless at such bitrates or maybe LossyFLAC.

Nowadays, I'm using Ogg Vorbis, q5.0. Sounds fantastic at home and on my Sansa Clip+.

FlaCuda -11 for archiving
LAME VBR V0 for daily usage
LAME VBR V5 for car stereo

I must seem like the oddball here.  For my LP transferring project, I've been using Apple Lossless for long term storage.  When I actually listen to something on the go (read: transfer to the iPod), I convert to AAC 256.

I live in a Mac household, so I have not found a way to make FLAC a convenient storing and streaming solution for my house.  With Apple Lossless, I can store, catalog, and stream everything around the house.

ogg vorbis LancerMod (SSE3) -q10 VBR

I mostly use FLAC