I am attempting to repair some music files that are incorrectly mastered. What I mean by this is that there are incorrect track boundaries on certain CDs I own and therefore my rips need to be edited. In the screenshot below, you can see how the beginning of a track is attached to the track that comes before it on a CD.
What I want to do is take the sliver of audio at the end of the first track and attach it to the second track so that the second track starts correctly. I know the sliver looks small but I realize now that if I zoomed in more, you could really get a feel for how big it is and it is 100% audible when I shuffle my music.
I have read in other threads that I need to disable the dither feature to ensure that when I export the files, they stay the same as their originals. Is it possible to make the edits I want to make and keep the file quality in tact?
If you still have the CD, you could make a new rip, but as one continuous file rather than separate tracks. Several CD ripping applications have this option. This isn’t an essential step, but it is a bit easier than aligning the separate tracks so that they play seamlessly.
Without that step, you need to align the tracks as you have done.
Check that playback is smooth across the transition from one track to another.
When you are happy that it plays correctly:
“Ctrl + A” (Select All)
“Tracks menu > Mix > Mix and Render”
You will now have one long track with all of the songs.
You can now add a “Label” at the start of each song (including a label at the very beginning for the first song). Then you can “Export Multiple - based on labels” to create new files that start and end at the correct places.
Ensure that you use lossless formats for all of your files (I’d suggest WAV format).
Turning off dither is hardly worth the effort because it makes so little difference, but with the above steps it does provide a small benefit. Dither should be turned off before you do the “Mix and Render” and turned back on after you have done the “Export Multiple”. Dither can be disabled / enabled in Preferences: https://manual.audacityteam.org/man/quality_preferences.html
Strictly speaking, the “Mix and Render” step would be considered “processing”, but assuming there is no overlap between the Audacity tracks (there’s never more than one audio track playing at a time) then the “mix” track contains exact copies of the tracks that were concatenated. In other words, for consideration of whether dither is required, you can consider the audio to be “not processed” and it is safe to turn off dither.
Note that if you use “Amplify”, “Normalize”, “Fade”, or any other processing, then ideally “dither” should be enabled.
Thinking about it a bit more, if both tracks come from 16-bit files, and neither has been “processed” in Audacity, then it doesn’t actually matter whether they overlap or not, so long as the mix does not exceed 0 dB. The reason being that adding two 16-bit values will produce an exact 16-bit value, so long as the sum of the values does not overflow (does not go above 32,767 or below -32,768).
The reason that dither is normally necessary is that 16-bit divides the amplitude range of +/- 1 into 65,535 values. At any point in 16-bit audio (any sample value) will be exactly one of those values. Now let’s say, for example, that you apply a fade to the audio. The sample values will be scaled according to their position along the fade. Because Audacity works in much higher precision, the new values can be thought of as being the “exact” right values, but the new values may not be “exact 16-bit values”. The new values may lie between 16-bit values. In this case, to convert back to 16-bit requires that the in-between values are rounded in some way to 16-bit values. “Dither” is a special form of rounding that avoids “quantization noise” by replacing it with a less unpleasant “dither noise”.
That is very interesting. I do not plan to overlap the tracks. I’m just going to do as you said and introduce labels to signify where I want the track beginnings to be. I am also not adding any effects because I’m trying to get files as close as possible to what the CD is supposed to be. All the files are regular 16-bit AIFFs.
