Im curious to find out various peoples preferences as far as their choice of switches for Lame.
currently im using:

--alt-preset standard --lowpass 19.5 --nogap -m s

because I find that has a nice warm bass but also with good clarity in higher frequencies , so much so that with my hifi surround sound setup I typically cannot tell the difference between the original CD and the mp3 convert (with few exceptions).

EDIT: While I'm at it, what does -Z do exactly? I see several people using it but cannot find an exact description of what it does.

well, after I learned a nice lesson ( http://www.hydrogenaudio.org/show.php/act/.../ST/f/16/t/8046 ) with the presets, I will go for 3.90.2 "--alt-preset standard -Y -F" for the next few albums, if I can't notice any high frequency distortion -Y may be my new favorite switch.

Hi,

Although I don't use mp3 anymore, I have a few comments about your command line:

- The --nogap switch, despite its name, unfortunately won't suppress gap problems between your tracks. If you need gapless playback, you need either to encode the whole CD from a single WAV file, or to use a gapless codec.

- for the average listener, due to mp3's sfb21 limitations your --lowpass switch will waste more bits, thus slightly decreasing quality on more important parts of the spectrum (0..18.5 kHz).

- the -m s switch, in your command line, will cripple the quality of your encodings. Many bits will be wasted in the stereo separation, while they would be much more useful where artifacts are to be avoided.

Bottom line: if you trust --alt-preset standard, use it. Otherwise, use your custom command line, but expect lower quality (unless you spend months of tuning LAME's code and doing bling listening tests).

For your information, every switch that you included, is better implemented in the core of --alt-preset standard. For example, this preset dynamically switches between -m s and -m j. When you override this intelligent behavior with a basic setting like -m s, you're doing yourself a huge disservice.

Cheers

why --lowpass 19.5? aps has a lowpass of 19.0, can you hear the difference?
why m s?

AAAaaaaargh

NumLOCK: I was not aware that the --lowpass 19.5 and -m s would likely degrade the quality of my encodes, as for --nogap , i need to use it mainly for my portable mp3 player, it manages to strings tracks seamlessly if they are encoded with --nogap , as opposed to normal (Some kinda compatibility thing heh)

Yep, 'nogap' works here, too.

If you get satisfying results when associating --nogap with your portable player, I'm glad to hear it

Be aware, though, that your mileage may vary (depending on both player and encoder).. so, don't be too disappointed if glitches pop up with your next mp3 player - and those couldn't be prevented by --nogap. Well, all in all, this setting won't harm anything at all.

About the --lowpass, it won't harm much either. If you have truly exceptional treble hearing, it might help a bit. In other cases, it's overkill and plain --alt-preset standard would be a slightly better balance.. in the end, all depends on you.

About -m s, yes, you should definitely avoid it.. it's really a step back in mp3 encoding.. the --alt-preset standard is clever, and does use this mode when necessary (thus saving bits for more important stuff in all other cases).

I think -m s is the most dangerous setting.. because you'll instinctively use it (many people do), even if it's harmful (compared to presets)

The most harmful scenario, is when your bitrate tops out at 320kbps (difficult part of music) and the mp3 player has to kill precious audible bits because it can't do channel decorrelation.

When you realize that this setting (-m s, used just instinctively) is harmful, you have made a next step in mp3 encoding. BTW I'm pretty damn sure you're about to dump this setting

Welcome to hydrogenaudio.

[Edit]
In other words - there are two painful limitations in the mp3 format:

- the "scalefactorband 21" issue, which raises in importance when your raise the lowpass frequency (--lowpass switch)
- the "pre-echo" issues, which raise in importance when you consume ~30% more bits in the stereo image (-m s switch).

If you use these two switches, you'll get the (misleading) impression that quality has improved. In fact, you'll lose a lot in time resolution (softer sound), which is mp3's big weakness (as shown in listening tests).
[/Edit]

NumLOCK: You are absolutely right actually, my command string is now:
--alt-preset standard --nogap
using lame 3.90.2 of course.

thx for the input,
I actually used to use lame with the --r3mix setting till I came across this site and became better informed.

1) You're a little late for April Fools' Day.
2) --nogap disables the Xing/VBR header, which means that the bitrate and time displays, and even possibly the playback, will be really screwy on some players.
3) This is more in response to the others... I think APS actually lowpasses at 18.5kHz. In any case that's certainly high enough.

If I were you and I'd absolutely insist on "gapless" playback, I would use a CBR line like --alt-preset cbr 192.

not fussed about bitrate / timedisplays... and have not yet suffered a problem with the playback... can it really mess it up?

p.s. I think he's a genuine noob

You're welcome !

I'm glad you switched to a nicer command line

As many people on this board, I too went through r3mix not so long ago. That preset sounded good, but obviously lacked in time resolution (produced smearing etc).

So, the plain --alt-preset standard is much sharper (and truer to the source).

If you have a few hours to spend, I'd recommend you to try mpc someday.. I switched a few months ago, and was astonished. Drawback is lack of portable player compatibility of course - but files are quickly converted for portable use.

Have fun.

PS, about NeoRenegade's post: it looks like VBR Xing headers are a bit screwed up anyway (tested in Winamp 2.9 and WMP8). You can use another tool to generate it after encoding (I too encountered time display problems - and "--nogap" wasn't involved at all).

just --alt-preset extreme here.

I thought I'd have a hard time finding out what the best parameters are for my music since I am so new to encoding MP3s but I am all the way for the --alt-presets now, after reading some threads here and articles I found on internet. Most of experienced users recommend (the newbies specially - not your case I beleive) not to mess with the preset settings as they have been carefully tuned already.
The extreme --alt-preset sounds good enough for archiving to me.

Is there really any reasons not to use presets, other than if one needs real small files or constant bitrate? What would they be?

Xing headers are okay... only software is broken and expects only FhG headers,
for example Winamp 2.81/3/Foobar2000 is okay, but WMP can't read Xing tags.
Although I haven't tried it, I suppose Winamp 2.9 does it too.

Umm.. I know WMP supports as little of mp3 as possible - but I mentioned that because I had issues with winamp 2.81 as well...

I just use plain aps.

I use it because it was systematically tested and tweaked by people who really had nothing to gain except the satisfaction of providing maximal quality and some kind of unified preset system.