Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Today, which AAC encoder gives the highest quality? (Read 49001 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Today, which AAC encoder gives the highest quality?

It's in the title and description, but it's not specific enough. When I say "high quality", I mean close(est) to the audio quality of a good lossless file.

I'm on Windows 7 Ultimate 64-bit. My plan is to turn my FLAC collection into AAC (and still keep the FLACs, of course). I'm looking to make them as close to the quality of lossless music as possible.


1. These tests were performed 3 years ago (2011) by IgorC. Are these tests "accurate" about the quality of those encoders today?

2a. If you said YES to #1, then which one of these pictures shows the best quality AAC that iTunes will dish out from an ALAC/WAV/AIFF file? pic1 pic2

2b. If you said YES to #1, then what settings do I use to get the highest quality AACs out of foobar2000's "AAC (Apple)" qaac encoder?

2c. Why does iTunes not use CBR even when I uncheck the "Use Variable Bit Rate Encoding (VBR)" option?


3. If you said NO to #1, then which encoder(s) will create AAC files that are closest to the quality of a "good" lossless file?


Today, which AAC encoder gives the highest quality?

Reply #1 –
I doubt the encoders have changed much, if at all, since 2011.

IgorC's tests were for ~96 kbps. You must not extrapolate them to other bitrates. The results can be quite different as the settings change. So don't even look at those graphs unless you are looking to encode at right around that bitrate.

Understand what you are asking for. "The quality of lossless music" is whatever sounds the same as lossless to you. So any settings that produce transparent results (i.e., you can't tell the difference) are of equally high, maximum quality. For most people, transparency happens at far lower bitrates than you would expect. Turning all the knobs to 11 to make you feel like it is of higher quality has no audible benefit; it merely wastes space, even though objectively the quality may be higher (less quantization noise, for example).

For pic1/pic2 you're asking about a choice between VBR and CBR, with the bitrate set to maximum. Variable bitrate (VBR) means allow the bitrate (frame size, really) to vary while maintaining constant quality (as measured objectively, e.g. the amount of quantization noise, regardless of whether you can hear it). Constant bitrate (CBR) means allow the quality to vary while maintaining a constant frame size. Which one is better is impossible to say. A fluctuation in quantization noise or the overall performance at a particular VBR setting may be audible in one case, but change the encoder, the listener, or the piece of music, and it's no longer an issue.

You might think maximum bitrate and CBR is maximum objective quality, since with those settings every frame is as large as it can be. But as we know from LAME, the encoder may use different algorithms for VBR than for CBR, producing different results. Maximizing settings can also cause the encoder to try to preserve more of the upper frequency bands, which may not be a good thing (it may not go so well, or the lower bands may suffer, for example). That said, I don't recall anyone proving with an ABX test that they can hear the difference between max.-bitrate CBR and highest quality VBR, be it MP3 or AAC.

For my portable players, I use the FhG Winamp encoder at VBR q3. It comes out to about ~100 kbps (so, IgorC's test is of interest). So far I can't tell the difference between that and the lossless originals. Sure, I could encode at a different q level to have "better" quality, but why? I won't hear the difference, and I want to save space. Maybe I could go with even smaller files and a different encoder, but I'm happy with what I've got.

But since you sound paranoid, you should just choose pic2.

Today, which AAC encoder gives the highest quality?

Reply #2 –
Quote from: mjb2006 on
But since you sound paranoid, you should just choose pic2.


"Paranoid"? This is my first thread & first 24 hours on these forums, and I get nervous around experts.

Please don't assume I know nothing just because I'm new.  I posted this thread here because its topic is one thing I don't know about. However, you (and my ~5 months of lurking these forums) helped me with that, and I thank you for it.


Anyway, I tested some ALAC files just an hour ago...ran them through the iTunes AAC encoder using pic1 settings, then with pic2 settings.

The files with pic 1 settings are larger and have 320kbps bitrates according to foobar2000's "Selection Properties" UI (under General). The files with pic 2 settings are (obviously) smaller and have ~300-310kbps bitrates (according to foobar2000 again). The odd thing is that the pic 2 (CBR) file fluctuates in bitrate just like the pic 1 (VBR) file does.

I see now that the difference is pretty much null...but I'll be going with the pic 1 settings because I don't care about the slightly (0.2 MB) larger size, and the slight bitrate increase is a nice bonus.

Thank you, again.

Today, which AAC encoder gives the highest quality?

Reply #3 –
Quote from: bigjoegamer on
The files with pic 1 settings are larger and have 320kbps bitrates according to foobar2000's "Selection Properties" UI (under General). The files with pic 2 settings are (obviously) smaller and have ~300-310kbps bitrates (according to foobar2000 again). The odd thing is that the pic 2 (CBR) file fluctuates in bitrate just like the pic 1 (VBR) file does.

I see now that the difference is pretty much null...but I'll be going with the pic 1 settings because I don't care about the slightly (0.2 MB) larger size, and the slight bitrate increase is a nice bonus.
That's a bit contradictory. Since bitrate is a synonym for file size divided by duration, it makes no sense to not care about file size increase but to care about the bitrate increase. How is it a "nice bonus" to get a higher bitrate? Your aim seems to be to use as much space as possible by using extremely high bitrates while maintaining the disadvantages of lossy audio encoding (e.g. no guarantee of transparency).
It's only audiophile if it's inconvenient.

Today, which AAC encoder gives the highest quality?

Reply #4 –
Quote from: Kohlrabi on
Your aim seems to be to use as much space as possible by using extremely high bitrates while maintaining the disadvantages of lossy audio encoding (e.g. no guarantee of transparency).


My aim is to accomplish all of these things at the same time:

1. lower file size than flac/alac (1/4th the size so far)
2. greatest "transparency" possible for AAC

With 320kbps AAC , I get both, don't I? 

Today, which AAC encoder gives the highest quality?

Reply #5 –
Quote from: bigjoegamer on
...
With 320kbps AAC , I get both, don't I? 


As said, transparancy is a personal thing, it depends on your hearing capabilities in combination with other factors like listening environment, speakers, etc. Most probably also you can go much lower than 320kbps AAC to achieve transparancy. Do some tests to find out for yourself.

Today, which AAC encoder gives the highest quality?

Reply #6 –
Quote from: Alexxander on
Quote from: bigjoegamer on
...
With 320kbps AAC , I get both, don't I? 


As said, transparancy is a personal thing, it depends on your hearing capabilities in combination with other factors like listening environment, speakers, etc. Most probably also you can go much lower than 320kbps AAC to achieve transparancy. Do some tests to find out for yourself.


Transparency/storage-saving isn't my goal. My goal is highest quality possible while staying lossy.

Even after saying that, I will admit...192kbps AAC sounds the same as 320kbps AAC to me. According to my ABX tests, at least.

Today, which AAC encoder gives the highest quality?

Reply #7 –
As for CBR showing varying bitrate while playing, it's true. It's more correct to call it constrained VBR (CVBR). This is true of most popular 'CBR' formats, including MP3, which uses bit-reservoir to briefly increase the bitrate allocated to specific frames. Either so-called CBR or VBR at 320kbps are likely to far exceed your transparency threshold.

In another thread, some very good ABX testers have mentioned that with iTunes AAC 150-160 kbps settings solve practically of their worst problem samples. You're double that rate and should be pretty damn confident you won't notice any artifacts. iTunes Store uses 256kbps "CBR" encoding presumably chosen to ensure very little in the way of customer complaints/refunds (and because it's about double the old 128 kbps "CBR" setting which was already very good, and double feels like a worthwhile "upgrade" to the layperson).
Dynamic – the artist formerly known as DickD

Today, which AAC encoder gives the highest quality?

Reply #8 –
Quote from: Dynamic on
iTunes Store uses 256kbps "CBR" encoding presumably chosen to ensure very little in the way of customer complaints/refunds (and because it's about double the old 128 kbps "CBR" setting which was already very good, and double feels like a worthwhile "upgrade" to the layperson).

iTunesPlus is 256 kbps constrained VBR.
CBR has more difficult to encode samples even at such high bitrate as 256 kbps.

Today, which AAC encoder gives the highest quality?

Reply #9 –
Have you considered lossyflac? You can boost the bitrate to just below that of a regular flac and still be lossy! This way you can save minimal amounts of space and still stress out about potential artifacts. It's exactly what you want.

This post may contain traces of sarcasm.

Today, which AAC encoder gives the highest quality?

Reply #10 –
Quote from: Dynamic on
As for CBR showing varying bitrate while playing, it's true. It's more correct to call it constrained VBR (CVBR). This is true of most popular 'CBR' formats, including MP3, which uses bit-reservoir to briefly increase the bitrate allocated to specific frames. Either so-called CBR or VBR at 320kbps are likely to far exceed your transparency threshold.

In another thread, some very good ABX testers have mentioned that with iTunes AAC 150-160 kbps settings solve practically of their worst problem samples. You're double that rate and should be pretty damn confident you won't notice any artifacts. iTunes Store uses 256kbps "CBR" encoding presumably chosen to ensure very little in the way of customer complaints/refunds (and because it's about double the old 128 kbps "CBR" setting which was already very good, and double feels like a worthwhile "upgrade" to the layperson).


Very good explanation.

I've made an observation...You can choose to use "CVBR" in foobar2000's Apple AAC encoder. This is the equivalent of the iTunes AAC encoder with VBR checkmarked (like in pic 1 at the first post). How do I know? @320kbps, the resulting AAC file will be the same size with both encoders. This has been the case with every lossless file I have transcoded so far.

Today, which AAC encoder gives the highest quality?

Reply #11 –
Quote from: Gecko on
Have you considered lossyflac? You can boost the bitrate to just below that of a regular flac and still be lossy! This way you can save minimal amounts of space and still stress out about potential artifacts. It's exactly what you want.

This post may contain traces of sarcasm.


You say that like 320kbps VBR AAC is a bad thing. But, yeah, I know..."sarcasm". 

Today, which AAC encoder gives the highest quality?

Reply #12 –
Quote from: bigjoegamer on
You say that like 320kbps VBR AAC is a bad thing.


For you it is, because by your own admission:

Quote from: bigjoegamer on
Even after saying that, I will admit...192kbps AAC sounds the same as 320kbps AAC to me. According to my ABX tests, at least.


If 192 and 320 kbps sound the same to you, 320 kbps is only wasting space for no audible gain in quality; therefore, it is worse than 192 kbps for your purposes.

Today, which AAC encoder gives the highest quality?

Reply #13 –
Quote from: Aleron Ives on
If 192 and 320 kbps sound the same to you, 320 kbps is only wasting space for no audible gain in quality; therefore, it is worse than 192 kbps for your purposes.



No. No, it's not.


Quote from: bigjoegamer on
Transparency/storage-saving isn't my goal. My goal is highest quality possible while staying lossy.


To be slightly more specific: the quality of the file's data must be as high as possible and stay lossy at the same time. Yes, the data itself, not the sounds I hear.


Today, which AAC encoder gives the highest quality?

Reply #14 –
You're posting on the wrong forum to get advice on your goal, then.

Quote from: bigjoegamer on
the quality of the file's data must be as high as possible and stay lossy at the same time. Yes, the data itself, not the sounds I hear.


As stated in TOS #8, measuring the quality of "the data itself" is not a valid way to quantify the quality of lossy compression. The only thing that matters when considering the quality of lossy audio is how it sounds, and the only way to demonstrate how lossy audio sounds is through an ABX test, which you have apparently conducted. Why are you so set against trusting the information your ears have given you? Your goals are more in line with those of lossless compression, i.e. to achieve the smallest possible file size while not losing any information contained in the original file. It would seem that the only answer you would find satisfies your criteria is to leave your FLAC files the way they are.

Today, which AAC encoder gives the highest quality?

Reply #15 –
Quote from: Aleron Ives on
You're posting on the wrong forum to get advice on your goal, then.

Quote from: bigjoegamer on
the quality of the file's data must be as high as possible and stay lossy at the same time. Yes, the data itself, not the sounds I hear.


As stated in TOS #8, measuring the quality of "the data itself" is not a valid way to quantify the quality of lossy compression. The only thing that matters when considering the quality of lossy audio is how it sounds, and the only way to demonstrate how lossy audio sounds is through an ABX test, which you have apparently conducted. Why are you so set against trusting the information your ears have given you? Your goals are more in line with those of lossless compression, i.e. to achieve the smallest possible file size while not losing any information contained in the original file. It would seem that the only answer you would find satisfies your criteria is to leave your FLAC files the way they are.


Then I hope it isn't ban-worthy (it would be my first and hopefully last "offense"), because after we're done here, I'll want to learn more from you guys on these forums. Since those rules are that relative to your advice, they really must be something to take heed to...
My lame excuse is this: Laziness/That "Hydrogenaudio Forum Terms Of Service" thread isn't under "important topics" in that forum (http://www.hydrogenaud.io/forums/index.php?showforum=25).

Anyway...I've already achieved the goals in line with lossless compression before I created this thread, i.e. I wouldn't post in AAC - Tech if this was about lossless encoding.

I could have easily turned my FLAC files into MP3 CBR 320kbps (I know, I can still do that), but after reading everywhere that "AAC is higher quality than MP3 at the same bitrate", I figured that, for the same size as MP3 320kbps (in other words, less than half of compressed lossless codecs like FLAC), I could have better quality than the best of MP3s.

But, you have helped me to think more about it, more about changing my methods...and all of this,  really. 


Another reason for not staying lossless is that I might want to have a copy of my FLAC collection in AAC and/or MP3 for compatibility reasons. Also, for uploading those lossy files to the "cloud" instead of FLAC, thereby saving the small cloud space (also, spending less time uploading/downloading them than I would with lossless files).

Today, which AAC encoder gives the highest quality?

Reply #16 –
Quote from: bigjoegamer on
Anyway...I've already achieved the goals in line with lossless compression before I created this thread, i.e. I wouldn't post in AAC - Tech if this was about lossless encoding.
Thanks for reminding me that this is the wrong subforum. Moved.

Quote from: bigjoegamer on
I could have easily turned my FLAC files into MP3 CBR 320kbps (I know, I can still do that), but after reading everywhere that "AAC is higher quality than MP3 at the same bitrate", I figured that, for the same size as MP3 320kbps (in other words, less than half of compressed lossless codecs like FLAC), I could have better quality than the best of MP3s.
Those statements are meant for bitrates which have a high likelyhood of audible artifacts. Most people will be hardpressed to hear any deficiency with (modern) MP3 encodes at 320kbps, outside of problem samples. Using AAC can solve these last 0.01%, but it has its own problem samples. No lossy encoder will be perfect regardless of how many bits you allow it to use. Did you experience any problem with 320kbps (LAME) MP3 encodes?

Quote from: bigjoegamer on
Also, for uploading those lossy files to the "cloud" instead of FLAC, thereby saving the small cloud space (also, spending less time uploading/downloading them than I would with lossless files).
That's contradictory to what you said before, that you don't care about the filesize/bitrate increase. If you want to make the most of your limited space you should use a much lower bitrate target than 320kbps.
It's only audiophile if it's inconvenient.

Today, which AAC encoder gives the highest quality?

Reply #17 –
Quote from: Kohlrabi on
Thanks for reminding me that this is the wrong subforum. Moved.


Whew..I actually thought I'd get in serious trouble...! I'll be more careful when making any more threads.

Quote from: Kohlrabi on
That's contradictory to what you said before, that you don't care about the filesize/bitrate increase. If you want to make the most of your limited space you should use a much lower bitrate target than 320kbps.


I was talking about cloud storage (but I guess you knew it). Right now, it is more about how the highest quality AAC files are still less than half the size of most FLACs (i.e. lossy storage vs. lossless storage in the cloud), not about how lower lossy bitrates = smaller file size.

Quote from: Kohlrabi on
Those statements are meant for bitrates which have a high likelyhood of audible artifacts. Most people will be hardpressed to hear any deficiency with (modern) MP3 encodes at 320kbps, outside of problem samples. Using AAC can solve these last 0.01%, but it has its own problem samples. No lossy encoder will be perfect regardless of how many bits you allow it to use. Did you experience any problem with 320kbps (LAME) MP3 encodes?


I understand it even more now!    Thank you for explaining. No, really. No wonder you are a mod.

A big reason why you make sense is because of my experience with all of those 320kbps LAME MP3 encodes...
My experience is that I've heard pretty much no difference between my recent 320kbps (iTunes encoder) AAC encodes and their 320kbps LAME MP3 counterparts.

But I'm still confused...Is there any other disadvantages to using 320kbps AAC instead of 320kbps MP3? These days, they're both gonna work on everything, right?

Also, I thought AAC was far more "advanced" than MP3 and would live longer than MP3. Guess I'm wrong again on both of those...

Today, which AAC encoder gives the highest quality?

Reply #18 –
Quote from: bigjoegamer on
Then I hope it isn't ban-worthy (it would be my first and hopefully last "offense"), because after we're done here, I'll want to learn more from you guys on these forums.

Judging from your responses thus far, it doesn't really seem like you want to learn from us, though. Multiple people have tried to tell you that your goals don't make sense (keeping your lossy files as large as possible when the whole point of lossy compression is to make your files as small as possible as long as there is no audible quality degradation), but you don't seem to want to listen.

Quote from: bigjoegamer on
These days, they're both gonna work on everything, right?


Nope. MP3 is still far more compatible than AAC when it comes to hardware players. PC software usually supports both, though.

Quote from: bigjoegamer on
Also, I thought AAC was far more "advanced" than MP3 and would live longer than MP3.


The thing about MP3 vs AAC is that both formats can achieve transparency on much of the music that exists. The difference is that because AAC has improvements over MP3, it can sometimes achieve transparency at lower bitrates than MP3 can, e.g. if MP3 can achieve transparency at 192 kbps on a certain sample, AAC may be able to achieve transparency at 128 kbps, instead. Yes, that saves space, but the savings aren't that great: it isn't even a 2x reduction in size, whereas the difference between CDDA (1411 kbps) and 192 kbps MP3 is already more than a 7x reduction in size. As such, the extra space you can save with AAC is not necessarily all that useful when storage is so cheap these days, especially since MP3 is much more compatible with hardware players. Of course the numbers in this example are entirely arbitrary; the transparency threshold (in kbps) for MP3 vs AAC will vary between samples, thus giving AAC a bigger or smaller advantage in space savings depending upon the music in question.

The real point that we're trying to make is that you don't need to use 320 kbps with either MP3 or AAC. You already said that 192 kbps AAC is transparent to you, and transparency is the only thing that matters, because it is the only thing that you can test objectively (with a double-blind test).

Today, which AAC encoder gives the highest quality?

Reply #19 –
Quote from: Aleron Ives on
Quote from: bigjoegamer on
Then I hope it isn't ban-worthy (it would be my first and hopefully last "offense"), because after we're done here, I'll want to learn more from you guys on these forums.

Judging from your responses thus far, it doesn't really seem like you want to learn from us, though. Multiple people have tried to tell you that your goals don't make sense (keeping your lossy files as large as possible when the whole point of lossy compression is to make your files as small as possible as long as there is no audible quality degradation), but you don't seem to want to listen.


Check the post above yours where I replied to/quoted Kohlrabi. I've changed for the better, even though I had some questions left. I understand if you didn't see it; your post was posted only 5 minutes after mine.

Now I'm wondering under what conditions is it a good idea to use 320kbps lossy files? I'd love to know, so I know when to use them, if EVER again.

Today, which AAC encoder gives the highest quality?

Reply #20 –
See my edits above, too. 

Quote from: bigjoegamer on
Now I'm wondering under what conditions is it a good idea to use 320kbps lossy files? I'd love to know, so I know when to use them, if EVER again.


The only reason you'd want to use CBR is if you're in a situation where VBR won't work, e.g. trying to stream the audio or trying to make it play on something that doesn't support VBR. VBR lets you achieve your desired quality level at the lowest bitrate possible, which is ideal for lossy codecs.

Today, which AAC encoder gives the highest quality?

Reply #21 –
Quote from: Aleron Ives on
See my edits above, too. 

Quote from: bigjoegamer on
Now I'm wondering under what conditions is it a good idea to use 320kbps lossy files? I'd love to know, so I know when to use them, if EVER again.


The only reason you'd want to use CBR is if you're in a situation where VBR won't work, e.g. trying to stream the audio or trying to make it play on something that doesn't support VBR. VBR lets you achieve your desired quality level at the lowest bitrate possible, which is ideal for lossy codecs.


Thank you. I read the edits just now...excellent statements/expression. 


So, anyway...CBR seems to be slightly more compatible then, right? I'm thinking that's the way to go, because I want compatibility in case I ever copy an MP3 file and share it with someone, or play it on older hardware/software. It feels "safer" that way, for lack of a better word.

Today, which AAC encoder gives the highest quality?

Reply #22 –
It's always the same here on HA: the 'don't waste any bit' group is so strong that you'll be attacked when having another attitude.

But there's also relief: from the encoders' usage polls we know that there's also a significantly large group of users who obey to the results of listening tests and/or their own experience but who also allow for a certain safety margin.
And there's good (though not necessary) reason to do so: When looking at say the current listening test you'll find that usually AAC behaves very well or even perfect even at a low bitrate like 96 kbps. But there are exceptions to this: With at least one sample AAC quality at 96 kbps is really poor (and I know samples where it's even poorer). And this isn't weird electronic stuff.
The 'don't waste any bit' group's attitude towards this situation is 'ignore it, it's very rare'. This is correct, and there's nothing wrong with this attitude. But it's not the only way to be. The safety margin group responds to these situations with, sure, a safety margin. After all there will never be experience with the universe of music, and significantly many people just want to make sure (more exactly: want to have a higher degree of safety).

If you use AAC @320kbps you have a really extreme safety margin which also in my opinon is overkill. I suggest to use Apple AAC --tvbr 91 (~193 kbps on average) which has already a very strong safety margin, or Apple AAC --tvbr 100 (~227 kbps on average) if --tvbr 91 doesn't bring you peace of mind.
But if you feel like having to use AAC @ 320 kbps: go ahead, there's nothing wrong with it in case you don't get into trouble with storage sapce.
lame3995o -Q1.7 --lowpass 17

Today, which AAC encoder gives the highest quality?

Reply #23 –
Quote from: halb27 on
It's always the same here on HA: the 'don't waste any bits' group is so strong that you'll be attacked when having another attitude.

But there's also relief: from the encoders' usage polls we know that there's also a significantly large group of users who obey to the results of listening tests and/or their own experience but who also allow for a certain safety margin.
And there's good (though not necessary) reason to do so: When looking at say the current listening test you'll find that usually AAC behaves very well or even perfect even at a low bitrate like 96 kbps. But there are exceptions to this: With at least one sample AAC quality at 96 kbps is really poor (and I know samples where it's even poorer). And this isn't weird electronic stuff.
The 'don't waste any bits' group's attitude towards this situation is 'ignore it, it's very rare'. This is correct, and there's nothing wrong with this attitude. But it's not the only way to be. The safety margin group responds to these situations with a safety margin. After all there will be never experience with the universe of music, and significantly many people just want to make sure (more exactly: want to have a higher degree of safety).

If you use AAC @320kbps you have a really extreme safety margin which also in my opinon is overkill. I suggest to use Apple AAC --tvbr 91 (~193 kbps on average) which has already a very strong safety margin, or Apple AAC --tvbr 100 (~227 kbps on average) if --tvbr 91 doesn't bring you peace of mind.
But if you feel like having to use AAC @ 320 kbps: go ahead, there's nothing wrong with it in case you don't get into trouble with storage sapce.


You make good points and I agree with them; (especially this -->) there's nothing "wrong" with 320kbps AAC, except storage space (and MP3 is more compatible, even today).

EDIT: added stuff

FYI, for AAC I use either iTunes itself, or the AAC Apple encoder in foobar2000.

For MP3, I use the LAME encoder in foobar2000.

Today, which AAC encoder gives the highest quality?

Reply #24 –
CBR 320 is an obsolescent holdover from the early days of MP3, and even then I could ABX a fair proportion of samples against WAV.  VBR sounds better than CBR at any bitrate (or at least never worse) and you'll have to work hard to find something that won't play it.  AAC sounds better than MP3 (unless you pick a really dumb encoder) at any bitrate but especially at lower bitrates.

Nobody here knows what your ears can hear on your equipment, but I'm betting you'll never spot the different between lossless and qaac around 160, and almost certainly not even at lower bitrates.  Find some known hard samples, see what bitrate you can spot them at, knock it up a notch, and forget about it.  Encoders aren't going to get worse, but your music library will only ever grow

As for "hi-res", whole different story, not going there, but you're not going to find anyone here telling you it is a good idea.