[ag8908]

I'm using a Burson Conductor DAC/amp.
 
I'm pretty sure that if I'm just listening to CDs, there is no benefit to setting the sampling rate to anything other than 16 bit 44.1khz. So what about Spotify? I listen to "extreme" quality on Spotify, which should be 320kbps.
 
Digression, fyi http://www.2l.no/hires/ can anyone tell the difference b/w the various flacs?

 

[Duncan]

It all depends on if you believe up, or over sampling pushes the distortions that may occur at 16/44.1 outside of hearing range, or introduces potential for more distortion (dithering and what not)...

No right, and no wrong answer, but plenty of room for debate...

I suppose for this purpose you also have to factor in the mp3 encoder / decoder (assuming spotify extreme is still 320kbps) so for redbook or lossless playback you would have 1411kbps being upscaled, vs 320kbps spotify, think if the en/decoding of that mp3 file isn't up to scratch how little data is being 'expanded' to 24/192...

Sorry, probably makes no sense, just woken up, hopefully if it doesn't someone will fill in the blanks

 

[ag8908]

well a CD which is 16 bit 44.1khz turns into 1,411 kbps (according to foobar), so Spotify must be like 320/1411 = like 23% of the quality of 16 bit 44.1khz? so what's the point in setting the sampling rate any higher than that?
 
I obviously don't understand the actual mechanics of the sound, just using random made up rules of thumb.

 

[Duncan]

Going back to the original theory of how MP3 (and all other lossy) encoders work, it is mostly to do with the high frequencies where they 'save' their data - stripping out everything above 16khz, and anything below that, if there are two samples (for example a cymbal, and a vocal) at a certain frequency, they 'assume' you can only hear one, and take away the data for the other - this is done over and over until desired bitrate is achieved, the lossier the encoding, the more brutal the stripping down of that data (and why, the lower the bitrate, the more distorted it sounds) - this data cannot be retrieved because it is removed at time of ripping, whereas for lossless encoding, this is just (with the exception of WAV) compressing the file, in the same way that you can compress a word document with winzip - the file, once uncompressed is identical to if it hadn't been zipped...

 

[ag8908]

I'm going to leave it at 32 bit 48khz for a while and see if it creates any quality issues, not that I can hear any difference between that and 32 bit 192 kHz.