Has anybody ever heard about the --nogap option in LAME ?
It's there since 3.89alpha and I wonder what it does.

Btw, after reading the FAQ about the gap-problem, I thought the best way to get rid of the gaps could be an encoder front-end, that sends the WAV-files through the encoder as ONE stream, and then splits the single resulting MP3-file into the individual tracks again....this should create perfect and gapless MP3s.
Has anybody seen a software doing this ?

Dr. Pupvogel
mk2 #080000169

Btw, after reading the FAQ about the gap-problem, I thought the best way to get rid of the gaps could be an encoder front-end, that sends the WAV-files through the encoder as ONE stream, and then splits the single resulting MP3-file into the individual tracks again....this should create perfect and gapless MP3s.
Has anybody seen a software doing this ?

I thought that as well, but then more knowlegable people here corrected me: MP3 frames are not totally independent, so even with that method some kind of transition artifact would be heard, unless decoder was also hacked so that it 'carries over' info from last frames of one track into beginning of the next (if they were marked as continuous in some way).


Dragi "Bonzi" Raos
Zagreb, Croatia
Q#5196, MkII#80000376, 18GB green

One can use the --no-res option in Lame to disable the bit reservoir, making frames independent. Of course, this applies all through the piece, not only at the ends, so you lose the benefit it gives (even if you encode with VBR as I do).

--
Toby Speight - 040103385

I thought the best way to get rid of the gaps could be an encoder front-end, that sends the WAV-files through the encoder as ONE stream, and then splits the single resulting MP3-file into the individual tracks again....this should create perfect and gapless MP3s.
Has anybody seen a software doing this ?

Yes, I do. Well, sort of.

As was mentioned elsewhere, the bit reservoir gets in the way on playback. I'm interested in messing with the new version of LAME when they get the nogap stuff working well.

Anyway, what you're talking about is creating a single huge rip of the entire CD, then splitting the resulting huge MP3 into smaller sections after it's been encoded.

Right now, I don't know of any software which does that "automatically", but you can do it yourself in two separate steps.

First, do the "big rip". Most ripping software lets you adjust the start/end points of the tracks before ripping. So de-select all but one track, and adjust its start/end points to match the beginning and end of the CD. For instance, in Audiocatalyst, you do this by unchecking all of the tracks except the last one, getting the properties of the last track, and adjusting its start point to 00:00. Then rip it.

So now you've got one huge track that represents the entire CD.

Now load up the latest version of my GapKiller software. Right now my regular web site address is down (I HATE MY ISP!) but you can temporarily get it from http://pweb.jps.net/~tfabris.

There is a feature in this new version that lets you split up a large MP3 into many small MP3s. Enter the track times from the original album. Preferably the times you got from AudioCatalyst (or whatever ripper you were using), as the album sleeve might not be correct.

That should do it. Now you have MP3s made in the way you suggested.

However, this is still not perfect, as the playback software might or might not Do The Right Thing with timing and with the bit reservior at the track transition points. You might still get a minor glitch at track switches. Remember that it's a two-sided coin: To get truly gapless playback, you need both the encoder and the playback software cooperating.

So until LAME gets that stuff working well, and playback software is designed to play back those kinds of files, it's still a hit-and-miss business.

___________
Tony Fabris

So until LAME gets that stuff working well, and playback software is designed to play back those kinds of files, it's still a hit-and-miss business.

If 'lame --no-res' does what it says on the tin, the playback code shouldn't need to know. Has anyone tried using this and seeing whether a car player plays back gaplessly? (As tms13 says, it will degrade quality -- if the format allowed for it, it would be nice if lame had a --no-res-at-beginning-or-end option.)

Peter

lame --no-res won't magically make separate inputs seamless - it just means that the result can be split with mpgtx or whatever. There will still be a click at the end unless the input makes up an exact number of frames.

There's also the issue of overlap - seamless encoding is intended to be concatenated together, meaning a ½-frame of overlap of the result:

 input:                                                     

                                                            

 -----------------------------+-----------------------------

                      file 1  |  file 2                     

 -----------------------------+-----------------------------

                                                            

                                                            

 output 1:                                                  

                                                            

 -----+-------+-------+-------+                             

      |       |       |       |                             

 -+---+---+---+---+---+---+---+---+                         

  |       |       |       |       |                         

 -+-------+-------+-------+-------+                         

                                                            

                                                            

 output 2:                                                  

                                                            

                              +-------+-------+-------+-----

                              |       |       |       |     

                              +---+---+---+---+---+---+---+-

                                  |       |       |       | 

                                  +-------+-------+-------+-

When encoding single files, the start and end are padded with silence.

I can't find documentation for --nogap, but from what I understand of the code, it should deal with both of these issues.

--
Toby Speight - 040103385

A while ago, I used Tony's program and a program by, I think, a friend of his that trimed off frames on the end of the track. It allowed you to listen to the end of the track to see when it sounded best. There was only about 10% of the splits that didn't come out sounding good.

Tony, do you know what software I might be talking about?

Also, I have been told that Ogg Vorbis is also capable of no gap playback.

Alex Lear

Tony, do you know what software I might be talking about?

Yes, it sounds like you're talking about my Gapkiller software, mentioned earlier in the thread.


___________
Tony Fabris

According to Mark Taylor, --nogap should work okay if you dont use tags (VBR or id3), and during encoding the the mp3 files have to be in the same directory as the
.wav file.
There is also a small bug with the interaction between --nogap
and the Xing header.

I thought the best way to get rid of the gaps could be an encoder front-end, that sends the WAV-files through the encoder as ONE stream, and then splits the single resulting MP3-file into the individual tracks again....this should create perfect and gapless MP3s.
Has anybody seen a software doing this ?

Yes, I do. Well, sort of.

musiCutter
(c) Macik ,
http://macik.homepage.com

Could be what you are looking for. However, i was unable to reach that website lately.

cu,
sven

proud MkII owner (12GB blue/green/smoked, was #080000113 is #090001010)

Excuse my ignorance, but are there benefits to encoding a full CD in one file and then splitting this file up into individual tracks?

___________________________

Steve - 18GB Mk II Sssmokey

Excuse my ignorance, but are there benefits to encoding a full CD in one file and then splitting this file up into individual tracks?

It is only an advantage for continuous-track material, such as live concert albums or Pink Floyd albums. That's what the whole "gapless playback" discussion is about.

I'll rewind and give more history for you:

MP3 does not, by default, properly support gapless playback of continuous-track material. If you encode a set of separate .WAV files, there will be tiny bits of silence inserted into the actual MP3s themselves, at both the beginning and the end of each and every song.

This isn't because there's anything wrong with the .WAV files. The wave files will play back perfectly gapless. It's the process of ENCODING the files that inserts the bits of silence.

By creating a single huge MP3 file and splitting it, you are working around the problem of this inserted silence. Each split point is made on an exact frame boundary, and there is no artificial silence inserted at the beginning and end of each track.

There is still a drawback to this method, though, because the frames are often interdependent, relying upon a feature known as a "bit reservior" that stays active between frames. Decoders empty this bit reservior between files, which is the correct thing to do if the files weren't meant to be played in sequence.

Finally, many players artifically pause between tracks because they need CPU time to prepare the next file for playback.

So fixing gapless playback is a multi-faceted problem. The authors of the LAME encoder are currently working on a method to tackle it at the encoder level.

___________
Tony Fabris

WOW - I have so much to learn. And I thought it was just a case of slapping in a CD and letting it go. Someone posted a link to info about different encoding methods and bit rates that I looked at last week ( sorry, I've been away since and forget the specifics ) but I took a look at that and there is some serious rethinking I have to do. I've started looking at different players - been happily using MusicMatch up till now. It's no bother redoing my CDs as I can do this again and again, the problem is where I've transferred stuff from tape, done the filtering and improvements and then converted to MP3 at 128 kbps and then in my naivety deleting my source .wav thinking if I want/need to improve things then I can just redo the MP3 at a higher rate.

The way I've been doing things, it sounds fine on my current system but I'm starting to worry that I might start to notice imperfections when I get a decent set of amp and speakers in the car.

___________________________

Steve - 18GB Mk II Sssmokey

And I thought it was just a case of slapping in a CD and letting it go.

Usually, it is. The thing about gapless playback you have to realize is:

- There aren't that many albums which are recorded as one continuous track. You might not even own one of these albums.

- Even if you owned one of these albums, you might not ever listen to it in the correct playing order. With so much music at your disposal, it's often more fun to shuffle.

- Even if you do listen to the album sequentially, it's not a very big problem to have the MP3-induced gaps. It's just a tiny little "bump" in between the tracks. It almost sounds like a skip on a vinyl record.

So the only time you'd want to go to all this trouble is if you were a perfectionist and absolutely HAD to hear Dark Side of the Moon the way it was meant to be played.

___________
Tony Fabris

Quote
So the only time you'd want to go to all this trouble is if you were a perfectionist and absolutely HAD to hear Dark Side of the Moon the way it was meant to be played.
/Quote

Or The Wall, Momentary lapse of reason, Wish You Were Here, Pro's and con's or hitchhiking plus many songs that have an intro as a sperate track from the main song. also recordings or concerts

I have wish you were here on the empeg as one track but that sucks because if I pause it for some reason I almost always hit the wrong button (I know that's my fault but it still sucks) and end up back at the beginning also the titles are not displayed for each song

32Gig MK2 In 2001 VW Golf TDI

By creating a single huge MP3 file and splitting it, you are working around the problem of this inserted silence. Each split point is made on an exact frame boundary, and there is no artificial silence inserted at the beginning and end of each track.

There is still a drawback to this method, though, because the frames are often interdependent, relying upon a feature known as a "bit reservior" that stays active between frames. Decoders empty this bit reservior between files, which is the correct thing to do if the files weren't meant to be played in sequence.

Hmm, so what if you store the bit reservoir at the split point into an id3v2 tag of the next track? Then the decoder could prime it from there if it needs to play the tracks out of order. Just a thought - don't know much about the nitty-gritty details myself.

Borislav

it sounds fine on my current system but I'm starting to worry that I might start to notice imperfections when I get a decent set of amp and speakers in the car.

Y'know, chances are it would sound better if it were re-encoded perfectly using the best ripping/encoding hardware and software at a higher bitrate... and chances are, unless you actually did a careful A/B comparison under the very best of listening conditions (car parked inside your closed-up garage, engine off, windows closed, high volume, listening very carefully) you would never notice the difference. There are many variables that affect how good your music sounds that are more significant than the bit rate you used for encoding or the presence or lack of a few minor imperfections and artifacts. Things like quality and power of your amplifier(s). Things like the quality, frequency response, size, number, and location of your speakers. Things like how good a job you did equalizing and tuning your system.

A marginally encoded song (for example, a 96 kbps download from the internet) will probably sound better played on a really good stereo system than a perfectly encoded (Plextor CD player, EAC and Lame at 256 kbps) song on most original factory stereos.

Unless you are really seriously into stereo quality, and think it is more important to impress your friends than it is to just listen to and enjoy your music, I'd recommend leaving your existing music alone, and just worry about improving the quality on new music you add to your empeg.

tanstaafl.

"There Ain't No Such Thing As A Free Lunch"

Hi Doug.

I have to disagree (well, sort of) with you on some points, while agreeing with the tenor of your post:
I would not worry about the 128kbps encoded files that were ripped from tape. They probably wouldn't sound any better than they are now when re-ripped and encoded with even 360kbps. It sure wouldn't be worth the hassle.
Regarding the music that was ripped from CD: I would vote for a re-rip if Steve is not satisfied with the sound of a song.
You said:
A marginally encoded song (for example, a 96 kbps download from the internet) will probably sound better played on a really good stereo system than a perfectly encoded (Plextor CD player, EAC and Lame at 256 kbps) song on most original factory stereos.
Problem with this is: He isn't listening to the song on a factory stereo, but on an empeg. I am mostly unable to hear a difference between a 128kbps (LAME) encoded file and a 256kbps encoded one on my computers (but can tell quite a difference between different encoders). I can easily tell the difference when listening to those files on the empeg, no matter which speakers I use (ranging from el-cheapo PC-speakers for approx. 10US$ to my factory speakers in my car).

There are many factors that are influencing the quality of your sound system, among them the choice of amp, speakers, cables, equalizer settings etc. However big those influences are, the encoding quality still is a big factor. Not one to overly worry about, but still one to consider.

cu,
sven

proud MkII owner (12GB blue/green/smoked, was #080000113 is #090001010)

Many thanks - basically, I worry about it once I've finally gotten around to upgrading the sound system which could be a while yet. Oh well, I'm happy with it at the moment and I don't travel in anyone's car who has a better system so no worries.

In reply to:

---

- There aren't that many albums which are recorded as one continuous track. You might not even own one of these albums.

---

Actually, I do own one of these - Lovesexy by Prince. It took me a while to realise that I didn't even have it on my empeg after I thought I'd transferred it. Guess the file size was too big. I went through an exercise of ripping it again, decoding it to wav, loading up cool edit pro, splitting up the file and then re-encoding back to MP3. I've since found software that mean't I didn't have to go back to the wav, but we live and learn. I made a slight mistake and have a section that is an ending on one track and the beginning on the next but as I mostly put my whole collection into shuffle that shouldn't present any problems.

___________________________

Steve - 18GB Mk II Sssmokey

Actually, I do own one of these - Lovesexy by Prince.

As I understand it, that album takes things a step further and indexes the entire CD as a single track ("Track 01"). That's different than what we're discussing, although my Gapkiller program will allow you to split that large track into separate ones if you want.

The gapless playback issues we're discussing are in regards to albums with a collection of discrete songs which happen to flow together (Dark Side of the Moon, Wish You Were Here, etc.). In those cases, the rippers automatically encode separate tracks, with each one getting the encoder-induced silent bits at the beginning and the end.

___________
Tony Fabris

I have to disagree (well, sort of) with you on some points,

I'm shocked! To think you might actually have the nerve not to accept my half-baked opinions as absolute truths! I tell ya, you youngsters just have no respect for your elders!

tanstaafl.

"There Ain't No Such Thing As A Free Lunch"

Tony,

I haven't worried much about gapless playback yet (for the time being I just make 1 big mp3 of these live CD/DVD). But have you looked at this 'encoder offset' option in Exact Audio Copy ? _Maybe_ it has some effect ?

[email protected]
--------------------------------------
Reg:1934/Mk1:158-6Blue(sold)/Mk2:380-12Amber

That link doesn't work for me, I get an error. Although I don't see how a setting in a ripper could affect anything, since the problem exists in the MP3 spec and at the encoder level. As I said before, the problem is not with the ripper. The ripper will create perfect gapless WAV files, it's only once they're encoded that the problem crops up.

___________
Tony Fabris

Hi.

No, that encoder/decoder offset setting in EAC is only used when creating an audio CD from MP3s. It just skips that many samples of the output Wave (of the decoder that was fed with the MP3) before writing to CD.

cu,
sven

proud MkII owner (12GB blue/green/smoked, was #080000113 is #090001010)