We all know that currently Vorbis is the only codec out there that doesn't add any silence to the decoded file, so when you decode your .ogg file the resulting .wav will be exactly the same length as the original one. This is very useful when encoding gapless songs (like live concerts...). But since Vorbis is a lossy codec, even though there is no silence added, sometimes you can still hear a click between gapless tracks because the end of the first song doesn't match up with the beginning of the second song. This doesn't happen always and as you encode files using higher bitrates these clicks seem to appear less frequently (as expected). One of the solution would be to "teach" the encoder that it has to pay attention that the end of one songs should match up with the beginning of another one. But I'm thinking that there might be a better solution: the encoder shouldn't use psycho-acoustic at the very begging & very end of the songs (for a few milli seconds). This will ensure that the edges will always match... Does this holds water or not? Greetings, Aleksandar --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Aleksandar Dovnikovic wrote:> > We all know that currently Vorbis is the only codec out there that > doesn't add any silence to the decoded file, so when you decode > your .ogg file the resulting .wav will be exactly the same length > as the original one. This is very useful when encoding gapless > songs (like live concerts...). But since Vorbis is a lossy codec, > even though there is no silence added, sometimes you can > still hear a click between gapless tracks because the end of > the first song doesn't match up with the beginning of the second > song. This doesn't happen always and as you encode files using > higher bitrates these clicks seem to appear less frequently > (as expected). > > One of the solution would be to "teach" the encoder that > it has to pay attention that the end of one songs should > match up with the beginning of another one. But I'm > thinking that there might be a better solution: the encoder > shouldn't use psycho-acoustic at the very begging & > very end of the songs (for a few milli seconds). This will > ensure that the edges will always match... > > Does this holds water or not? > > Greetings, > Aleksandar >I would do this on playback if possible. If the levels don't line up at the join, then you'll hear a click or a pop. You can smooth the area around the join (say, +- 10 samples, applying a 10 sample moving average) to try and remove this. This can also happen if you do things like change bit rates in the middle of a piece. -- Regards Marshall Eubanks Multicast Technologies, Inc. 10301 Democracy Lane, Suite 201 Fairfax, Virginia 22030 Phone : 703-293-9624 Fax : 703-293-9609 e-mail : tme@on-the-i.com http://www.on-the-i.com --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
>One of the solution would be to "teach" the encoder that >it has to pay attention that the end of one songs should >match up with the beginning of another one. But I'm >thinking that there might be a better solution: the encoder >shouldn't use psycho-acoustic at the very begging & >very end of the songs (for a few milli seconds). This will >ensure that the edges will always match...A more sensible approach would be to encode the entire thing as a single stream (which is easy enough). You could then, afterwards, split this into multiple seperate files - vorbis allows you to do sample-precise edits, like that. The tools don't exist currently, but wouldn't be too hard to write. Michael --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
> We all know that currently Vorbis is the only codec out there that > doesn't add any silence to the decoded file, so when you decode > your .ogg file the resulting .wav will be exactly the same length > as the original one. This is very useful when encoding gapless > songs (like live concerts...). But since Vorbis is a lossy codec, > even though there is no silence added, sometimes you can > still hear a click between gapless tracks because the end of > the first song doesn't match up with the beginning of the second > song. This doesn't happen always and as you encode files using > higher bitrates these clicks seem to appear less frequently > (as expected). > > One of the solution would be to "teach" the encoder that > it has to pay attention that the end of one songs should > match up with the beginning of another one. But I'm > thinking that there might be a better solution: the encoder > shouldn't use psycho-acoustic at the very begging & > very end of the songs (for a few milli seconds). This will > ensure that the edges will always match... > > Does this holds water or not?Of the various suggestions, I have to say I like this one the most. It's simple and relatively easy to do, gets the desired result, etc. It would have to be an option of course (those first/last frames would be kinda big), but that's not a problem. Monty --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Actually you can get the silence removed on MP3's if you encode using VBR and use the FH VBRI table in the MP3 stream. This might be what the author was implying. The silence is still in the file but the header tells the application not to load it. As far as I know only Sonic Foundry Siren 2.0 encodes and takes advantage of this information. -----Original Message----- From: Chun-Yu [mailto:cshei@indiana.edu] Sent: Thursday, January 18, 2001 12:54 PM To: vorbis-dev@xiph.org Subject: Re: [vorbis-dev] Idea for gapless tracks -------------------> > > > > We all know that currently Vorbis is the only codec outthere that> > > > > doesn't add any silence to the decoded file, so when youdecode> > > > > your .ogg file the resulting .wav will be exactly the samelength> > > > > as the original one. > > > > > > i know im new here but... > > > mpegplus does that and the author claimed that his was the first > > > > No it doesn't. Vorbis and maybe VQF are the only lossy encoders Iknow of> > that don't add silence. > > i dont mean to start an argument but...I don't either, but encode any WAV file in Ogg Vorbis and MP+, decode them, and check the sizes. The decoded .ogg file is the same length as the original, but the decoded .mp+ isn't. I tried it with the lastest version of the MP+ encoder. Chun-Yu --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered. --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
>One of the solution would be to "teach" the encoder that >it has to pay attention that the end of one songs should >match up with the beginning of another one. But I'm >thinking that there might be a better solution: the encoder >shouldn't use psycho-acoustic at the very begging & >very end of the songs (for a few milli seconds). This will >ensure that the edges will always match...>Does this holds water or not?making files that depend on others is a wrong way. suppose, i don't have next file that decoder will do? or what will be if i'll encode cd and then split it on tracks? the only one way to make gapeless output is to make independend frames and buffer them for decoder. as i know, mpg123 makes gapelss output of mp3 (this can be checked with total recorder) even if frames depend on each other just because frames from the next file are buffering and for decoder both files seemed like one. andrei@vest.ru --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.