basically convert the mp3 to a wav or aiff file - there will be no loss at this point.

edit the wav or aiff file

then either play the wav or aiff, or convert it to apple lossless if you want to save space

converting it back to mp3 will introduce a quality loss

there is no real way to edit an mp3 directly

As long as u save it as an mp3 with the bitrate equal to or greater than the original bitrate, u shouldn't lose any quality.

that is most untrue

I know of two Mac applications that allow "lossless" editing of MP3 files...in other words, editing without decoding/re-encoding.

deepniner.net

www.rogueamoeba.com (can also do lossless editing of AAC files).

Neither of these will enable anything more than the most basic level of editing and adjustments, but for simple edits and gain changes these may be exactly what you need to avoid quality losses.

Absolutely. *ANY* lossy encoding step will *ALWAYS* degrade the sound quality, regardless of whether you're going higher, lower or the same bitrate as the source file.

The advice given earlier in the thread, that any lossy files decoded to linear-PCM then edited/mixed/whatever should be rendered to a "lossless" format for use, was correct.

The only type of transcoding that does not degrade audio is that done between lossless formats (FLAC to Apple Lossless, for example...or Apple lossless to AIFF for editing then back to Apple Lossless or FLAC).

these will only cut on the frame boundaries of the mp3, meaning there is no sample-accurate editing, so the editing ability is very rough and limited

Use a non-destructive editor, move around and slice in edit/multitrack view, bounce the arrangement to a new file, save as.

So if you're starting out with a lossy file & you're editing in the digital domain & you save it at the same or higher bitrate, how do you lose quality?? (Not trying to argue or be a smart ass, just trying to learn something!)

I use ableton drag and drop the mp3 into the session view. Then edited what I need to then export as wav file then convert back to mp3. It maintains the original quality that way

false...read what DJMark posted

Every time you do a lossy encode, you're dumping 70 percent or more of the incoming audio data.

The codecs are treating all incoming audio as linear-PCM. If you do a re-encode of a previously-lossy-encoded file, the algorithm has no way of "knowing" that a bunch of data was already removed from the file. It'll go ahead and remove more.

In very simplistic terms, "transcoding" with lossy codecs is like making a copy of a tape and should be avoided for the same reasons. The type of sonic damage is very different, but conceptually/practically it's the same principle.

I believe AAC is somewhat more forgiving of repeated encodes than MP3. I participated in some of the listen-testing that was done in the development of AAC back in the late 90's, and I recall that files repeatedly encoded/re-encoded was a (small) part of the testing. The issue of transcoding was already well-known back then, as a lot of radio and TV stations had adopted various record/playback technologies starting in the late 80's that involved lossy-encoded media. Such as computer-based automation playing audio from hard disk, "digital cart machines" using removable media, and digital studio-transmitter links.

I personally first heard the "transcoding" issue in 1993, when the radio station group I was doing engineering work at decided to install a fancy new digital studio-transmitter link. I think that STL was using "Musicam" compression at 128kbps per stereo pair, in an early form of "joint stereo mode"...something that initially caused us a lot of mystery when trying to do routine audio calibrations since we could no longer do an "L-R null"...but I digress. Then a few months later one of the two stations sharing that link put in a new computer-based playback system, which used some other form of data-compression for the audio. One of these systems processing audio at a time was acceptable, but when put together the sound was absolutely terrible. I think the end solution was keeping the computer, and changing back to the old analog STL.

Mentioning all this detail to illustrate that this is actually a fairly "ancient" issue that anyone working with audio in a "professional" capacity ought to be well aware of.

This does makes sense. I've never thought about it this way!

So I guess if you HAVE to go back to mp3 format, you're kinda screwed! If I did an edit on a 192kbps file, I better save it as a 320 to at least try to preserve as much of the lossy info that I had to begin with, lol! Wow, I guess gotta hop on the Serato FLAC support bandwagon!

No, Mark is right, but I think you guys are misunderstanding. If you make any changes to an mp3 in a destructive editing application (cut, normalize etc) no matter what way you try to save your changes, it will degrade. You may not be able to notice it after one save, but if you keep adjusting it and coming back it will start to get more and more artifacts until you get that digital wash sound. You definitely don't want to save to a higher bitrate, once it is what it is, it is what it is and can only go down in quality.

The reason Tunecrew said to save to a PCM, is because you are no longer making changes to the original mp3, but the copy of it. What he's suggesting isn't going to make the quality any better than it is, it just makes sure that it won't degrade anymore because it's a lossless copy of the original. You can close your program, come back at a later date, cut it, gain it, eq it, whatever, it won't matter. But once you've made those changes and saved, you are stuck with those changes since undo capabilities go out the window once you close the program and reopen.

Now the method I mention is if you want to make quick cuts, rearrangements etc. for whatever custom play you have in mind. With this you don't have to save anything but a bounced version of your final and nothing else is touched. Think Pro Tools, you take a long vocal track in which the engineer just let it run and there are long breaks in between where the vocalist maybe breathing, his headphones are too loud (turn up m'a headphones!) or whispering the other vocalists part whatever, you trim out the bullshit and mix it in with the rest of your sound and bounce it out, but you're never actually making changes to the file. This is non-destructive editing.

For just making simple custom edits of track, PT is not necessary. When in a simpler program like Adobe Audition, you can insert a track into multitrack view, separate, merge, slide around, lop off end pieces, envelope, whatever you want to do but you're not compromising the mp3 because this view is non-destructive, any sort of EQ or reverb/delay etc is all in real time so it's not applied to the actual mp3 either. When you bounce down your work, you're able to save it to whatever format you want, and if you think you may want to make different edits to the track, you can always save the session, and have different versions of the session too (without taking up any HD space) if you want come back anytime you want without ever compromising that original mp3.

I hope this makes sense.

The reason I personally no longer do what Tunecrew mentioned (in terms of editing club songs) is I'm tired of having a large master file for a shitty track that I may have to keep around, just in case I want to re-edit it again. It's much more efficient to just do everything non-destructively and be done with it.

However where I do still use Tunecrews method is slicing things up like you would in a sampler. It comes in real handy there if you're a Fruity Loops or Reason user working with samples. If you play back a sample and realize it may need a trim, you can apply it directly, reopen your sampler and the newly overwritten and better suited file is where the old one used to be.

Granted I'm on Maschine now so I no longer use these programs, but I would still need to bounce to a PCM if I want Maschine to see the file. (I just find it easier to read the file directly from disk, rather than sampling in a whole bunch of long unecessary pieces where the sound could also be affected)

You could...or you could use Apple Lossless (ALAC) which Scratch Live has supported for quite a while now.

Using "Max" sbooth.org you can transcode from FLAC to ALAC with no quality loss, and all the tag info from the original file will be retained.

To my knowledge, there are only two programs that allow you to edit and save MP3's without compresssion, that being NHC Wavelab and X-Wave MP3 editor. Unfortunately, both lack in precision editing like you would have with Sony Sound Forge. NHC Wavelab is the best out there for now, it's a bit fussy to work with, but will allow cutting, splitting and volume leveling before saving as lossless.

NHC Suite offers a host of audio and video editing programs. I purchased the whole lot of them back in the early 2000's and have been using Video Pad Editor, DVD Burner, Tone Generators, and a few others for a while. Some are good, some are fussy, but overall - a good buy.

mp3DirectCut can do this, but it's only for Windows -

mpesch3.de1.cc

It can even do batch editing - I use it all the time to trim silence from the ends of songs so they work better with AutoPlay.

if your source file is MP3, re-exporting it into an MP3 is going to be transcoding regardless of the program you use.

To avoid lossy transcoding, you pretty much have to save in a lossless format and then you'll basically have a wav-sized file that sounds like an MP3.

The only basic editing suite that won't affect audio quality right now would be using Serato Flip and that's because all it does is automate cue points in the track instead of actually editing and re-exporting in audio.