Mojo doesn't know if the source file was .wav or .flac.
It just sees a bitstream that the transport device has already decoded and sent across the digital-out, as PCM
But MP3 Vs Flac there will be a difference? Ogg Vs Wav ?
Ahh, that's different.
FLAC is a so-called 'Lossless' codec, that does not discard any information that was present in the original PCM / .WAV file. It just 'packs' the data more efficiently. The same is true for .ALAC and .APE, which are equivalent lossless codecs, but less-widely-supported by playback devices, which is why audiophiles have a tendency to focus upon .flac as the open-standard lossless codec of choice.
So-called '
Lossy' codecs like .mp3, .AAC, and .Ogg
do discard some information from the original file, in order to achieve a much-smaller file. This discarding of information tends to be done according to an algorithm derived from a psychoacoustic model that tries to discard only elements of data that are considered to be inaudible (or
almost-inaudible). They're very clever, but they're certainly not perfect.
So, you can see that a .FLAC, .ALAC, or .APE file, when decompressed, becomes (
broadly-speaking) 'identical' to the .WAV / PCM source file from which the .FLAC / .ALAC / .APE file was originally derived.
In contrast, however, an .mp3, .AAC, or .Ogg file, when decompressed, does create a PCM data stream,
but it
lacks some of the original detailed information that was contained within the original .WAV / PCM, so the sound quality may not be subjectively perceived as being as good as that achieved with a file that has never had any information discarded.
That, then, is the reason why audiophiles like to use .FLAC, as it is a compromise solution that retains the original quality, but allows about 30% or so reduction in filesize, by more efficiently
packing the data, on the storage medium,
without irretrievably discarding any audio information.
The issue can be complicated a little further, though, because some transport devices (and DAPs) may generate more circuitry or RF noise (due to increased CPU activity) when decompressing files on-the-fly, and/or may generate more circuitry or RF noise when transferring large data files from memory cards. Even some full-sized desktop transports can have similar issues, especially if their storage media is a conventional electro-mechanical computer hard-drive, because the electric motor in these hard-drives can be a potential source of physical
and electrical
and RF noise.
In most of these scenarios, well-implemented buffering can help reduce such issues, but there are some devices where the noise is quite obvious, and there can be some variation depending on how tightly-compressed the file data is (which influences not only the stored size of the file, but also the potential CPU-load during on-the-fly decompression. Even .flac files have 9 different compression levels available, with 5 / middle generally being the most-recommended - some devices will even refuse, or fail, to play .flac files encoded at higher compression ratios).
So,
sometimes induced noise can be worse when retrieving larger, losslessly-compressed (.FLAC, .ALAC, .APE), or non-compressed (.WAV, .DSD, etc.), files, from the memory card or hard-drive, which is an irritating (and, strictly-speaking, unnecessary) paradox.
Aside from minor hardware issues like those described above, broadly-speaking, a .WAV, .FLAC, or .DSD file does have a better chance of sounding better than an .Ogg, .AAC, or .mp3 file, since no original data has been irretrievably-discarded, although you can get surprisingly good results with .Ogg, .AAC, or .mp3 if it is at a
high-bitrate, since less data is being discarded by the encoder algorithm.
Lastly, it does depend, to a degree, on the complexity of the actual music you like to listen to. If you listen to real-life acoustic or orchestral recordings, that contain masses of timbral detail, numerous seperate instruments (and seperate, simultaneously-plucked, strings), complex spatial cues, etc., etc., then differences between codecs might be a little more apparent, but if you listen to synthesized, or computer-generated music (EDM, etc.), then differences between codecs might be a little less apparent. To put it another way, some types of music pose more of a challenge to a psychoacoustic-based algorithm than others do.
.
