last night i couldn't sleep and i listened to some mp3 music on my computer and because i heard some much good about mpc i was plannning to rip also my first mpc, i didn't knew something about mpc, but when i played it on winamp i tought that mpc was kind like an mp3 but not has no vast bitrates as 32-40-..-192-256-320 vbr mp3, mpc just picks the right variable bitrate at a given time ex 198, is this true or not true??

next question

i saw this Mid/Side Stereo: enabled in the ape v2.0 tag when i pressed the i (nformation) button in winamp, i saw somewhere ( in the mp3 forum) that mid/side stereo isn't very good, correct or not? and how can i make it that it is joint stereo or full stereo?

i'm new @ mpc and mp3, don't know very much about it, a friend of mine showed me a month or tree ago that i have to use eac and lame to rip mp3's instead all the other crapp, i used to rip my mp3's to 256kb/s cbr

if i use --quality 8 --xlevel as command line for mpc, is the quality better than mp3 256 kbit's cbr??

last question

i very new @ mpc and mp3 and my native language is dutch , what does gapless mean?? somebody can explain it please?

thanks, greetz koen

I can answer a couple of your questions right now



Generally, using --quality 8 is overkill. Most files will be transparent (as in you won't be able to tell the difference between that and the cd) at --quality 5. Try using that and see how you find it. But anyway, yes, it will be better quality.



This basically means that if you rip a whole cd, there wont be any gaps inbetween the tracks. The tracks will flow perfectly as if you were playing one big file. MP3's aren't gapless so sometimes you will hear a slight "skip" inbetween songs.

Hope this helps

True. MPC is "fully VBR" with 1 kbit steps.



mid/side is a different name for joint stereo.



Yes. But even --quality 5 is already better.



Gapless means that i.e. live albums have no short blips of silence between the songs.

Yes, but within the scope of a quality setting. If you specify --quality 5, then you'll get a moderate VBR depending on the complexity of the music...perhaps averaging 180-220 kbps. --quality 8 for the same music might net somewhere between 260-320 for an average, and --quality 10 can send it much higher. You're right to assume that MPC "figures out" what bitrate to use, but it's baswed on the --quality setting you specify. The higher the --quality, the higher the average VBR.


As far as I know, there is nothing wrong with MPC mid-side stereo encoding. I use mid-side stereo for my FLAC encoding which is lossless. Unless I'm mistaken, using mid-side vs. full stereo should make no difference in sound quality, it only specifies how the encoder will use bits to define stereo seperation. Someone could correct if I'm wrong on this, but this is my understanding of it.


"Better" is subjective. You should ABX MP3 cbr 256 against MPC --quality 8 --xlevel and see if you can hear any difference. If your hearing is good enough, then you could tell a difference between them. 256 kbps MP3 is barely transparent for most people I think, again depending on the music being encoded. Hard techno sometimes breaks at 256, light pop would probably be fine at that rate. Lots of variables to consider. The solution is to ABX them yourself, though. Edit: I got off-track a little on this one...I think most people would say that MPC --q 8 is a far better quality setting than MP3 cbr 256, but my point is how much of a difference between them could *you* hear? That's what determines what bitrate or --q setting you should use..if all your other needs are met (compatibility, gapless playback, etc.) then the lowest average bitrate or the lowest --q setting that's transparent to your ears is probably your best option.


"Gapless" playback means that music will not have gaps between tracks that don't exist on the original source CD audio. MP3 writes frames in such a way that truly gapless playback is impossible. There will almost always be a gap of a few hundred to a few thousand samples, a fraction of a second, but almost always detectable. As a matter of fact, a desire for gapless playback (and a few other things) is why I am re-encoding my collection in FLAC right now. Lossless codecs are gapless. MPC is gapless. I think Ogg is gapless, but verify this before believing me. But MP3 is *not* gapless.

Importance of gapless playback for you will depend greatly on your musical taste. For me, it's mostly Pink Floyd, techno and some classical music that causes me the most trouble when it comes to gaps between the tracks.

Hope this helps...

Edit: Three people answering in three minutes...that's pretty funny.

thanks for the replies, guys, i know now something more about mpc and i'm considering to use mpc instead of mp3, many thanks, koen

You surely have not experienced foobar's newest betas.
Peter has introduced gapless MP3 playback.
Even though he claims it to be not perfect, audibly it is.

And I have tried with Dark Side and Wish you Were Here.

I had the same mistaken impression as you before, thinking that mpc only encodes songs in a small bitrate range. In fact, mpc can encode (if I am not wrong, otherwise a mod please correct) frames with bitrates from 32 to 700+ kbps. This gives mpc very flexible encoding options, and tries to ensure that you get the best quality/filesize ratio (I don't want to leave vorbis and aac out, they try to do this too, although mpc is more advanced so far). Imagine, mp3 can only encode from 32 to 320 kbps, and only at the certain fixed frames you mentioned.


Indeed, mid/side stereo is joint stereo, and for most mp3 encoders besides LAME, they have a badly tweaked implementation of it. Joint stereo in LAME is audibly flawless, and is well-designed to save file sizes, whenever you use it. Anyway, it is on by default, so use it . Joint stereo also works audibly flawlessly in mpc.



It will definitely be better, and it is also guaranteed overkill UNLESS you have double-blind tested that you have excellent hearing, in which case you are one of the very few for whom mpc has audible artifacts, and even then, they are hard to ABX already.
Otherwise, use --quality 5 --xlevel. It is THE recommended encoding setting, and already surpasses mp3 256 kbps in quality.

Thank you for the correction. I have heard about the new foobar beta supporting gapless MP3 playback, and I even have it installed on my PC (though I haven't had much time to spend with it yet).

What I should have stated is that MP3 doesn't have natively gapless playback on all compatible players as some other codecs do, though some (or is it just one?) specific player(s) *can* play MP3s gaplessly. foobar2000 can do this, but most other software players, PhatBoxes, and many portable players cannot. Whereas, for example, FLAC is gapless on all compatible players (AFAIK).

Side Question: Out of curiosity, how does foobar2K do gapless MP3 playback? It was my understanding that MP3s wrote only full frames at the time of encoding, and that is why true gapless playback wasn't supported (prior to the latest fb2K). Does it parse the frames of encoded files as it reads them, and skip any silence at the end of the last encoded frame?

All recent LAME versions write gapless playback info (amount of samples that need to be skipped at the beginning/end of file) into file header (unfortunately those values are incorrect in certain conditions). Currently, foobar2000 (only recent 0.7 betas to be exact) is the only player capable of using that info (even LAME --decode ignores padding). This process does not involve any kind of file analysis or guessing how many samples to ignore; all info needed to play the file gaplessly is stored in the header.
Of course, this header isn't a part of MP3 standard (just like Xing header storing VBR seeking table and length) and existing software/hardware players require modifications to make use of it.
Also, latest foobar2000 beta is capable of editing gapless playback info, so you can basically make any MP3 file (also not LAME-encoded one) play gaplessly as long as you know encoder delay (576 samples for all LAME versions as far as I know), and exact length of track (can be eg. extracted from cue sheet).

@zZzZzZz: Thanks for the explanation. I had no idea LAME put gapless p/b info into the MP3 header. Maybe other software/firmware players will follow foobar2K's lead in making use of this data to give people even more incentive to stick with MP3. I'd love to have gapless MP3 p/b in my car, for one.

@schytluisje: Sorry for hijacking your thread and turning it partially into a discussion of the intricacies of MP3 gapless playback.

No harm done, I think.
To me, is kinda good that the almost automated response that comes whenever posts the words "MP3" and "gapless" together ("that is not possible"), has to change.
The MP3Splice guy as well as Peter Pawloski have provided clever solutions, and it has to be known everywhere.

I am sure that the same companies that make MP3 hardware players, even "costumer conscious" ones like iRiver are full of those "automated response" people.

To me, the solution provided by Peter is the best, because it both encourages people to use Foobar and to encode with LAME. Two birds with one stone. MP3 has many years left in it. They do not have to be unpleasant.

heum, am i right to say to an mpc file with quality 10 is near a lossless file (ex ape ..) true or not true? if i want the best quality but no lossless i should use quality 10 or should quality 8 be enough?

Saying that would probably be a bit far fetched. In general mpc -5 is considered to be transparent to 99.9% of all people in 99.9% of all cases. Choosing a higher quality then that should only make sence with certain problem samples and even then it might be that choosing a highter bitrate wont help that much as the problem might be more related to the encoder then the bitrate. Anyway these problem samples seem to be extremely rare.

If you want true lossless you wont get around using one of the lossles codecs and that will increase storage space drastically.

Meaning that lossy codecs designed to be lossy and are meant to (and mostly do) achieve transparancy at a certain quality setting (ie. mpc -5) and lossles codecs are the ones that are truly lossles.

Despite all this smart talk I totally have to admit that I encode all my mpc's at quality -8. But hey it is just for that warm fuzzy feeling and a bit of extra sence of security and absolutely not based on that much rational reasoning. And in that case those few extra bits of storage space really don't matter to me.

Well, I think you could say that meaning that yes -q 10 keeps more of the original information that -q 8 and so is "closer" to lossless. Although I would feel somewhat unconfortable using that term on lossy codecs, except for non-perceptual ones (i.e. Wavpack or Optimfrog)
Again, allow us to reiterate that MPC has been found transparent (that means, impossible to distinguish from the original = best quality!) 99.9% of the times at -q 5 and even the most exigent are usually satisfied with -q 7 (insane), since what still fails at this setting (a rare occurence), will most likely fail at higher 'q's.
If what best quality means to you is that it will sound exactly like the original, then it is a relative term which should be determined personally. (Again, it is recommended to try MPC -q 5 and ABX it (if you are not familiar with ABX, by all means, ask)). I daresay that you will be quite conviced that MPC -q 5 does it for everything you try.
If, on the other hand, by best quality means "highest settings", then yes, the best quality you can get in MPC is -q 10. But I guarantee you that there will be a lot of wasted bytes, perhaps not so many as in lossless, but still quite a few.
My personal recommendation would be to use ABX on the most complex and difficult sample you think you've got encoded by MPC -q 5. If you succeed in detecting a difference between the original and MPC, and it annoys you, then I'd go up to -q 7 'insane' straight away. Its 220 kps average bitrate are still usually affordable, at least more than Lame cbr 256 kbps.
Hope this helps. But seriously, try ABX.

Good luck,
~Dologan

just to say: mpc is really gappless, i used to rip my mp3 to an image with an cue image and cut it in songs with musiccutter, this isn't gapless just discovered, just when it has to change from song with winamp ( buffer ahead on change song = 2500 ms, buffer 5000 ms ) if you hear closly to the speaker you can hear a slight short dip, just like only 1/2 frame is missing ( 1 frame = 30ms???), with mpc no dip, no cut just great , this test is correct did blind listening, i think it's strange, with an mp3 image that you cut into pieces there arn't any frame missing or not??

mpc lover for live

edit : typos

Many people would agree that MPC -q 5 is better then any MP3 that you can make. That makes it somewhat ridiculous to use any other quality setting.

I just started a new job a week ago and I've been downloading songs of my computer onto my work computer. My upload is capped at 256. Boy, am I glad I encode everything at -q5. Just wait. You may think it's cool to waste bits now, but small files can really come in handy sometimes. Especially since your essentially trading nothing.