Audio codecs: 16/24 bits FLAC. Vinyl. etc ..

[Aaron.Walkhouse]

…except, of course, the effect that sample rate has on the highest frequencies.

If the original has detail in the upper band which is clipped off by a later encode
at a lower sampling rate, good ears will be able to easily tell the difference if the
speakers or headphones are good with high frequencies.

"Warmth" would be lost with higher rates, as high frequencies subtly scrub it off.

[Moe]

The highest frequencies above 22K?  No one is going to hear that, on the odd chance the frequencies were present at time of recording, microphones were used that can capture them, they weren't filtered during later mixing and mastering, and you happen to own speakers or headphones capable of reproducing them.  If you manage to make it past all those obstacles, then yes, your dog might notice.

Warmth scrubbed by high frequencies?  You have gone off the deep end.

[blu_people]

Thank you for your posts! I somehow understand what you guys are saying, although I am kind of lost in all these information. I am still noob at 'digital music stuff' and all I know is "The higher, the better", which is obviously not true as this depends by a lot of things.

For the future, not to mention the frequency, should I choose 16bits or 24bits FLAC? -- I assume it's useless to take a 96kHz one as The highest frequencies above 22K? No one is going to hear that. I'll just stick to 44.1kHz :-)

[Moe]

At the consumer end of things, beyond 16 bits and 44.1kHz sampling you are just wasting disk space. If the recording is really really well made, the content has a wide dynamic range, your system is top end, and your listening room is really quiet, you might notice a little less noise in the softer parts on a 24 bit file. For pop music and noisy movies, those extra bits are wasted.


There is plenty of value for those higher bit depths and sampling rates during the recording, mixing, and mastering steps, but not in the final product. And frequently, those reasons are technical rather than sonic reasons.

[blu_people]

At the consumer end of things, beyond 16 bits and 44.1kHz sampling you are just wasting disk space.

I am only a consumer. :-)

If I would re-encode the 24bits to 16bits 44.1kHz, would this process loose quality? If it does, how much?

---

I would be using FFMPEG, by the way..

[Moe]

If I would re-encode the 24bits to 16bits 44.1kHz, would this process loose quality? If it does, how much?

The number of bits per sample determines volume resolution.  More bits means a larger number of steps between very loud and very quiet.  Reducing the number of bits doesn't mean there will be any noticeable loss in quality as the actual dynamic range of the majority of recordings is relatively small and can be accurately captured in 16 bits (or less).

That said, noise is the biggest factor in keeping more bits around.  Room noise, equipment noise, and noise introduced during processing is ever present.  If the noise level is very very low, then a 24 bit recording can bury it further than a 16 bit recording can.

There is a cost of converting as you can't just whack the last 8 bits off of each sample and call it done.  That would introduce some very unpleasant distortions.  The process is called dithering and any reasonably competent piece of software should handle that conversion without issue.

It all depends on the quality of the 24 bit recording.  If the noise floor is low and the audio content was mastered to use the full 24 bit dynamic range, then there should be no noticeable difference.  But if the noise floor is high, or the recording level is low, then it would probably be better to not convert them.  Try it and see what you hear.

[joew771]

What I would suggest is download 2 versions, the flac and the mp3 or whatever version, no matter the source, vinyl or digital or whatever. If you think one sounds better than the other, then keep that one.

All of this is dependent on what the user thinks. Some people think a vinyl flac sounds best, some think a cd flac sounds best, some think an mp3 sounds best, etc.

I would reccomend that you get what you like, and if they sound the same, then keep the smaller file. That's it. It's up to you.

[blu_people]

Thank you all for all your posts! :-)

[chrisfound]

Bit depth has nothing to do with frequency response.  It is quite impossible for a 24 bit recording to sound warmer than a 16 bit recording based solely on that.

Rather, bit depth relates directly to dynamic range, and more importantly, noise floor.  That doesn't mean volume or loudness, however.  A 24 bit file isn't louder than a 16 bit file.  It just means there are more steps between absolute silence and full scale loudness.

In short, any differences you perceive between 16/44.1 and 24/48 are very likely to be the source or creation of the file itself rather than the bit depth or the sampling rate.

As m engineer who has recorded a lot of artists, I feel angry that the SACD does not sound better than 44K1 and Still you go and mess up the file by providing a flac of the original, STOP IT, we spend years downloading to find in the end its a degraded copy of the original, WE DON'T NEED this crap.  
The time I have lost in finding the end result I refuse to listen to it, artists and engineers spend their whole life to search for the goal to find it is been buggered up with
Keep the original and its format and please even the deaf

Chris Found