Hello all

Im a confessed audiophile and im new to this whole mp3 business. I decided to give mp3s a try shying away from my preffered CD player in favour of an iRiver H320 for long distance travelling and the likes.

Naturally i was concerned about compression and the quality of my music as a result. I respect many of you have been testing mp3 LAME versions for years and have some extensive knowledge about using these codecs. Hopefully you'll not strike me down for being new

I naturally started off compressing my cd collection using the following command line params in EAC : %l--alt-preset 128%l%h--alt-preset extreme%h %s %d
I was impressed with the results and found it hard to detect any real worrying depreciation in quality taking into consideration the equipment used. I was listening through my iRiver with some Sennheiser HD 25 SP phones.

I chose extreme because i want top quality, Insane seem to be an overkill. The extra 1.5meg per file that extreme produces i can most definatly live with. A lot of people like size v quality, however with a 20gb player i could only class myself as being lazy by not deleting infrequently listened to albums for new additions if my file sizes were bigger due to using extreme.

Now what i would really like know is what technical advantages does preset extreme give over preset standard. Im finding it increasingly difficult to accuratly compare standard v extreme. My thoughts seem to indicate better dynamics and seperation between frequencies using preset extreme if you listen in with good phones. I have some Sen px200's for travel use and its harder to detect with them. I can't help but feel biased with my oppinions about extreme so i would like a technical discussion from you guys And any findings you've concluded after testing the two. I cant seem to find that many really meaty in depth topics on the use of the extreme preset.

Used v3.90.3 as the forum seems to suggest its the most stable due to extensive testing, even though 3.96.1 seems indistingsable. Again, all's i wanted was top quality and if that meant slightly larger file sizes and encoding times them im prepared for that.

READ THE FAQ or use the search function.

And by the way you answered your question yourself. If you cannot hear a difference then its truly transparent to your ears. So why bother about higher quality if its for listening purposes only.

Hope this does not sound rude, but a bit of reading before posting should help alot.

Thanks for fast reply!

Im aware of the other forum posts and ive read all the recommended settings for the LAME codecs. Ive tralled through quite a few threads on bitrate comparison and user oppinions on certain settings and command line operators.

I havent really sat down for extended periods of time testing standard and extreme ive simply chosen some songs where i think the birate/compression factor maybe tested. What im concluding myself is that extreme is better than standard to my ears becuase of better dynamics on good quality phones. This is definatly distinguisable if i play the same CD on a B&O system, the seperation becomes clear and dynamics identifyable from warm bass reflex to crisp hi-tones.

What im worried about is my biased oppinion. Im looking at the player seeing what compression im playing at and possibly convincing myself extreme is better. What im after is some of you experts that have carried out propeperly controlled tests with respectable results. Most importantly to me, the solid reasons behind the enhancements if any that extreme will give over standard.

Im very picky about audio quality im now im hassling you guys about it

Should be "--alt-preset extreme %s %d"

The reason you did not find much info is that almost no one can differentiate between the two. And "dynamics" and "separation" have nothing to do concerning the differences between -aps and -ape.

The difference, IIRC, is a little less artifacting on problem samples. Such "castanets" and pre-echo issues.

Search the forums a little bit.

I really think you could go much lower on your encodings and never notice a difference. Not even on better equipment. You have to train yourself to listen to artifacts. And you will find that the actual differences are not what you would think.

Welcome to the forums.

Hi OldSkoOL,

when you're on the move, you'd be hard-pressed to notice any benefits from using the extreme preset over standard. Especially not with Senn PX-200s. If the extra size does not bother you, stick with extreme.

(Personally, I feel extreme is overkill for portable use. I would also try "--preset standard -Y" or "--preset medium" with L.A.M.E. 3.96.1 and compare them with the 3.90.3 extreme MP3s. If you cannot distinguish between them, the former is a better choice. Also, use an ABX program to compare the files).

Cheers,
CD

Excellent response AtaqueEG!

Thanks for that insight. Indeed i am new to detecting "digital" artifacts as ive not really listened to a great deal of mp3s. I can easily detect imperfections that exists on the origonal CD's anything from incredibly poor mixing/mastering distortion to the studio echo and vocal mic isolation. I guess my ears are trained to picking up on the origonal faults.

Also i guess im worrying about compressing my music in terms of quality as its taken me this long to buy an mp3 player and compress my collection.


Also thanks to Gabriel for correcting me on the command line options

Not to mention its impact on battery life

Thank you



Why dont you encode to lossless AND to MP3? That is what I am doing right now. FLAC encoding is super fast and with programs like FLACATTACK and MAREO (again, search the forums) you can do both encodings at the same time.

This way, you can change your mind. And even encode to other formats if they get supported on the iRiver (a possibility now that Rockbox is being developed). And also you can try the newer versions of LAME, which are very promising (there is some evidence that they could already be better than 3.90.3 on alt preset standard) without fear and worries.

Thank you! I have been typically listening with the px200's in my car on the way to work. I do sometimes get imersed into some songs but only really become critical when listening on my HD 25's in bed when all is silent. I dont like to trust my own judgement when i have tested and experieneced so little with mp3s.

As pointed out im sure the only differences between standard and extreme appear to be the way it deals with specific instruments, most of which my ear has never listened out to before therefore becoming transparent.

I have read up about ABX testing and im interested in carrying out some measured tests eventually. Any input though is greatly appreciated!


Thanks for all the responses

I visted the FLAC web site then to my horror realised i couldnt play them on my iRiver. I will also look into Rockbox as i think i skimmed over that as being a hacked/third party firmware for the iRiver series.

Both shall be explored!




I initially thought not to bother even reading the alpha testing thread due to my inexperience and lack of input i could give but i will definatly read that thread 2moro when i get bored at work.


Thanks for the pointers

it is ok ... it means

low:
--alt-preset 128

high:
--alt-preset extreme %s %d

...which is impossible to use (together)...what are you traying to say?

Are you speaking from personal experience or do you know of available evidence of some sort? I'm really interested to know. I chose mp3s over aac because I read somewhere in this forum that aac drains the ipod battery faster than mp3. i realize, of course, that could still be the case, except that, if i understand it correctly, preset extreme uses more battery life than preset standard.

I've been using preset extreme too, because of what I thought were some ringing I've encountered in my files. Upon reading OldSkoOL's posts, however, I'm beginning to think it could simply be due to poor mastering of the source CDs. I'll have to go back to my CDs and do some tests.

edit: formatting

APX will drain your battery faster due to bigger files ... bigger files resulting in more HDD access activity.

So as i understand it now the differences between standard and extreme are rather minor and especially hard to detect on portable plays. Is it safe to say that extreme wouldnt provide any real noticable improvement in clarity/clearness/sonics compared to standard. However it may produce slightly better clarity on certain instruments and sounds where audio is naturally weakened as a result of compression and lower bitrates.

One thing i would like to ask is, Compact Dick suggested to use --preset standard -Y

I read up on the usage guide for LAME. I see -Y is experimental. I tried searching but the forum wouldnt allow anything less than 4 characters. Could someone briefly explain what the experimental -Y will do for my encoding?


TIA!

http://www.hydrogenaudio.org/forums/index....=ST&f=15&t=5803

http://www.hydrogenaudio.org/forums/index....t=ST&f=16&t=703



edit: added URL and quote

Perfect, thank you!

it is true that AAC drains the battery much faster than mp3, however on iPods with hard drives this can be compensated when using smaller files. Therefore, 160kbps mp3 will play as long as 128kbps AAC. It does not matter what codec, bitrate or setting was used since the CPU power needed is always the same. Therefore, the CPU power needed depends on the format only...and AAC just needs more than mp3, period. If you want more information on this search the forum

IMHO this is not detectable by everyone and especially not on portables. --preset standard (aka -V2) may be already overkill for portable players even if you use very good headphones. If you use the crappy headphones that come with iPod and stuff and if you listen in a noisy environment (for instance outside) --preset medium (aka -V4) is overkill as well...

I own an iPod mini + Sennheiser PX100 headphones (which are way too good for an iPod) and I encode to --preset standard. That is because I use the same files for listening at home and therefore don't have to re-encode. Bitrates ~196kbps that LAME 3.96.1 and --preset standard produces work for me...

Try a blind ABX listening test and see if you can tell the difference between --preset standard and --preset extreme with headphones on your home system in a quite environment...I doubt it that you'll be successfull...not even on known 'problem samples'.

Also, see my post I recently made about this topic: http://www.hydrogenaudio.org/forums/index....ndpost&p=276150

Yeah, but AAC is l33t!

Apple's site stimates a runtime of 15 hours on the newest iPods, using 4min average, AAC 128k files

Obviously they would recommend using AAC, but I hightly doubt you could get better runtime AND acceptable quality going below 128k on MP3.

Jojo thanks very much This post was exactly what i wanted to read.

You made a good point here Ive grown up with music and listen to it day in and day out. Luckily i have some pretty good equipment as its a hobby too. I can hear so many imperfections in origonal cd copies which are obviously there on the encoded copies too. The artifacts you mention i guess are mp3 artifacts. My ears aren't trained to pick up on them atm so your argument and bloating and extreme seems very valid in my situation.

I just hope i dont start picking up on those digital mp3 artifacts just yet. I have enough of a hard time dealing with poorly mastered cd's

Im definatly interested in the -Y parameter now as i understand it cuts off some higher frequencies which normally cannot be heard. Will experiment with that too. Just encoded with preset medium, did also reduce the filesize with slightly lower bitrate and noticed the base bitrate dropped down to 80 too.

I did listen to my girlfriends iPod and found it sounded really flatt to my ears not much seperation and quality in the sonics but then again those nasty white headphones were probably to blame. I'm even thinking of encoding some AAC files now to compare.

Im quite into this now its pretty interesting. Some great posts on this forum. I'll give the new alpha a whirl too not that i'd probably see much difference. I mainly wanted some advice because i'll be encoding a lot of cd's over the next week. Im glad i asked the question now rather than after.

All's i want it perfect quality within the boundaries of mp3 compression.

That is why I suggest encoding to lossless also. That way, if you change your mind later (and believe me, you will, I have a couple of times) you will not have to re-rip and tag all those CDs. And you can experiment with the new alphas without much more work than a batch-convert script.



Then again, encode to lossless. Try the newest alphas. Test them. Looks like MP3 is somewhat interesting again.

Oh, and if you encode to lossless, you can later encode to AAC for your girl. That is one of the major advantages of the iPod.

Ive already got FLAC installed and have tried encoding it too it looks great and i can create ogg files which i can play on my iRiver which is a definate bonus!!

Im unsure about trying the alphas. Sadly the testing in the currently alpha thread isnt as intensive as i'd like therefore im un-decided on what to make of the quality produced by the latest alpha and preset standard. If it is the best yet i'd propbably encode the rest of my cd collection with it.

Batch testing is attractive though i'd like to explore that!

Definatly, although im writing this comparing/listening to some alpha encoded files im probably best not to encode my collection with it. Im still thinking of sticking to 3.90.3 but choosing standard over extreme and possibly using -Y.

Tonight shall be an eventful comparing adventure.

We're going a bit OT, but...

What he said means that with that command line options it is possible to choose between --alt-preset 128 and --alt-preset extreme depending on which radio button you have checked in EAC compression settings, either "High quality" or "Low quality".

Everything between %l applies to "Low quality", while everything between %h applies to "High Quality".

I'm not sure (as I've never used this feature) if the "%s %d" part, common to both, can be "factored out" of the %l%l and %h%h or if it must repeated inside each of them.

Sergio

Yes that command line was used with EAC settings which explains the two command strings in there

what are you talking about? There's no need to go below 128kbps mp3 to achieve better battery life! In fact, you could use 160kbps mp3's and would get equal battery life to AAC 128kbps. Since 128kbps mp3 LAME performed almost as good as iTunes AAC 128kbps you have the choice...you could encode in LAME 3.96.1 -V5 --athaa-sensitivity 1 (which produces 128kbps files) for slightly worse quality over AAC 128kbps but therefore end up with longer battery life - or you could go with LAME 3.96.1 -V4 or --preset 160 (which produces 160kbps files) in order to achieve equal battery life over AAC 128kbps but would end up with better quality.

Now, the iPod shuffle is a different story. I'm not sure if it has the ability to use lower CPU for mp3's or not like the other iPods...but if it does it would mean that a 32kbps AAC file would still drain the battery faster than a LAME mp3 320kbps file...

So to recap i take it preset extreme doesnt have any clarity advantages over preset standard. All's it may* have is some rare improvments in certain instruments where birate can distort the sound...

Also with regards to the -Y parameter. I understand it drops all higher frequencies that aren't detected by the human ear therefore reducing the birate in those areas resulting in smaller files. However for anyone that has tested this experimental parameter, does it have any downsides particulary with regards to quality depreciation / introductuion of artifacts...

Thanks in advance! Just trying to build up a solid basis before i encode the rest of my cd collection Probably going to go with aps standard with 3.96.1 for smaller file sizes and same quality as 3.90.3 and possibly the -Y parameter dependant on the result of this discussion

Right. And you should never drive above the speed limits.

That's what I suggested FLAC encoding also. To test. What do you have to lose?

AAC is equally "unfinished" and no one suggests not encoding current versions.

I use extreme, the result: Less Paranoia. I feel it sounds more effortless for some songs. Rip and listen... aaaaah. You won't have to worry that youre not getting the most out of your music. The battery life tradeoff is worth it to me. You can always charge your DAP. Insane is much bigger in size and this time not worth the extra size cause the sound differences compared to extreme is minimal if none in a lot of songs to my ears.
Also, I am not a fan of the 3.90.3 lame. I feel it sounded dry and slow compared to the snappier and at the same time better decay in the sound of 3.96 and .1 This is after comparisons with my own songs ex. although the two lame versions sound similar, I prefer wma192 over the older lame and like the latest lame the best. Hmmm

Dry and slow? I think I hear the sirens of the TOS8 Police.