[damians]

Good morning everyone, on windows I use the mediaval cue splitter program to split a large flac of an entire audio cd into separate flacs. This program does not carry out any re-encoding of the data stream, at least that's what it says on the official website:
http://www.medieval.it/cuesplitter-pc/menu-id-71.html
I recently switched to Linux and mediaval cue splitter is not available, for Linux there is the flacon program which performs the same function.
But when flacon divides the flac file, it recodes the data, it doesn't just cut the tracks.
According to the developer this is not a problem:
https://github.com/flacon/flacon/issues/103#issuecomment-543838432
Even if it weren't a problem, I would prefer a program that simply cuts without remodifying, what do you think?

 

[gregorio]

Even if it weren't a problem, I would prefer a program that simply cuts without remodifying, what do you think?

I’m not a Linux user and therefore can’t recommend any Linux programs. However, what the developer has stated is true. FLAC is a lossless compression scheme, so decoding and re-encoding will make absolutely no difference whatsoever, even if you were to perform that cycle millions of times (unless the program has some sort of serious bug of course). This is not true with a lossy compression scheme, such as MP3 or AAC, where you don’t want to keep doing decoding/re-encoding cycles (or transcodes) as there will be a loss of data/audio quality but with FLAC it’s not a concern.

G

 

[Davesrose]

It seems Flacon is the main FLAC editor for Linux. If you have a cue file, it looks like another popular option is using Terminal and apt install cuetools shntool flac. Looks like Flacon is a graphical interface of shntool. https://askubuntu.com/questions/521388/native-flac-cue-splitter

 

[71 dB]

...decoding and re-encoding will make absolutely no difference whatsoever,...

G

The only difference can be how much data compression is achieved (how much redundancy is being removed).

 

[gregorio]

The only difference can be how much data compression is achieved (how much redundancy is being removed).

Sure, but I was referring to the difference when the FLAC is decoded (EG. For playback).

G

 

[71 dB]

Sure, but I was referring to the difference when the FLAC is decoded (EG. For playback).

G

Obviously there can't be differences in sound quality as FLAC is lossless, but file sizes and bitrates can vary depending on how thoroughly redundancy is removed from the music. Some/many people may only care about sound quality, but I pointed out there are a potential difference with lossless formats when encoding them again.

 

[damians]

The only difference can be how much data compression is achieved (how much redundancy is being removed).

Yes, in fact, on the flacon I can choose a compression level ranging from 0 to 8, I imagine that at 0 there is no compression or that it is minimal. However, on the mediaval cue splitter I cannot choose the compression level as it does not reencode during the split.

 

[71 dB]

Yes, in fact, on the flacon I can choose a compression level ranging from 0 to 8, I imagine that at 0 there is no compression or that it is minimal.

The range 0-8 has more to do with how much computational effort is done in the encoding in order to remove redundancy. The difference in the bitrate/filesize isn't that huge. I just tested (in Audacity) with a song (49,457,900 bytes at 44.1 kHz/16 bit) and the file sizes are:

Level 0: 27,445,151 bytes (55.5 % --> data compression 44.5 %)
Level 8: 24,531,439 bytes (49.6 % --> data compression 50.4 %)

How much lossless compression is achieved depends quite a lot on the recording level and music genre. This test song is (Hip-House) club music from 1989 (2 IN RHYTHM - WE WANT FUNK - CLUB MIX) ripped from vinyl (XL-Recordings XXL 001).

However, on the mediaval cue splitter I cannot choose the compression level as it does not reencode during the split.

I believe not re-encoding the music means the sample points where splitting can happen are "sparse." For example splitting may be possible only every 128 sample points. However, I am not an expert of lossless data compression so I might be wrong about this.