Describe differences in LP playback vs CD playback

[magiccarpetride]

Could it recreate McCartney's vocal on Hey Jude, and how did it cope with the inordinately long fade out?

Sure. The way parrots cope with inordinately long fade outs is they fly south, and as they're flying away from the listeners, the sound is getting weaker and weaker.

[magiccarpetride]

I have had spent a lot of time listening very carefully to 16bit( ripped CD), 24bit and DSD (if available) and CD versions of the same album. Two albums in particular come to mind Dire Straits (Love Over Gold) and The Wall (Floyd).
Anyone who can at least get a 16bit rip and 24bit version of these and listen carefully via a revealing DAC will come to some conclusions and they may not be what you think!

Lets put it this way 24 bit Hi Rez does not necessary mean you are getting anything better at all!

Then compare them against vinyl.

Ok I will tell you what happens.

The Wall. 16 bit sounds way better than the 24 bit version. The vinyl version slaughters them both completely.

Dire straits. 16 bit good. 24 bit version - more air and very slightly better. Vinyl version on a par with the 24 bit release.

Make of that what you will.

I'm amazed that you could actually hear the differences between 16 bit and 24 bit hi rez tracks. I have personally been through the 'cargo cult' phase in my audio journey where I firmly believed that hi rez format was hugely superior to the red book format. I was always on a lookout for hi rez reissues of famous albums. I remember being head over heels when I downloaded Weather Report "Heavy Weather" in hi rez (and also SACD rip of Billy Cobham's "Spectrum"). I was marvelling at how smoother, less fatiguing, more musical those FLACs sounded.

Until one day a buddy of mine dropped by and listened to me raving about hi rez formats, and forcing him to listen to some tracks with me. In the end he said "why don't we do a little bit of experiment? I'll choose one track in red book format, and then the same track in hi rez. I'll choose them randomly from the remote, and you tell me which is which. We'll repeat the exercise as many times as you'd like."

Sure! I was excited to try that. And then -- flop! Much to my chagrin, I could never guess, with better than random probability, which format is which.

That was a wakeup call for me. It demonstrated to me that I'm mostly listening 'with my eyes'. In other words, if I know it's a hi rez format playing, then I'll read all kinds of virtues into the listening experience. And if I know it's a red book format, then I'll read all kinds of harshness and issues into my listening experience. But if I don't know which format is playing, then all these virtues and issues suddenly become moot.

Try the randomized test, see what you make of that. For 100% honesty, I wholeheartedly recommend you do that in the presence of other audio enthusiasts.

[Jimbo]

I'm amazed that you could actually hear the differences between 16 bit and 24 bit hi rez tracks. I have personally been through the 'cargo cult' phase in my audio journey where I firmly believed that hi rez format was hugely superior to the red book format. I was always on a lookout for hi rez reissues of famous albums. I remember being head over heels when I downloaded Weather Report "Heavy Weather" in hi rez (and also SACD rip of Billy Cobham's "Spectrum"). I was marvelling at how smoother, less fatiguing, more musical those FLACs sounded.

Until one day a buddy of mine dropped by and listened to me raving about hi rez formats, and forcing them to listen to some tracks with me. In the end he said "why don't we do a little bit of experiment? I'll choose one track in red book format, and then the same track in hi rez. I'll choose them randomly on the remote, and you tell me which is which."

Sure! I was excited to try that. And then -- flop! Much to my chagrin, I could never guess, with better than random probability, which format is which.

That was a wakeup call for me. It demonstrated to me that I'm mostly listening 'with my eyes'. In other words, if I know it's a hi rez format playing, then I'll read all kinds of virtues into the listening experience. And if I know it's a red book format, then I'll read all kinds of harshness and issues into my listening experience.

Try the randomized test, see what you make of that.

That would be an absolute way of assessing digital tracks and I agree blind listening is the best way. But I can assure you where I listened and the means of delivering the audio was so clear and unambiguous and at such a high level of resolution it would still be obvious.

As Marco alluded to earlier mastering in any format is everything but i would also add not all Hi Rez material is actually what it purports to be. Quite often it may be at lower bit rates and unless you have the means for checking you don't actually know unless you obtain Hi Rez material from a reputable source that checks and assures the validity of the provenance.

I have actually found ripped 16 bit actually more preferable on many occasions.

[swampy]

If 16 bit was good enough then no DAW would use 32 bits for all audio processing and manipulation internally. Similarly if 16 bits was good enough then no mastering engineer would demand 24 bits as minimum for processing. Everything would be 16 bits since its good enough. When I render a DAW track the 24 bit version sounds better. Not a million miles better but better. When you render 16 bits you need to worry about quantisation noise and what dithering method to use to avoid this, triangular or PoW-R2/R3 since they also make the track sound slightly different.

[magiccarpetride]

That would be an absolute way of assessing digital tracks and I agree blind listening is the best way. But I can assure you where I listened and the means of delivering the audio was so clear and unambiguous and at such a high level of resolution it would still be obvious.

As Marco alluded to earlier mastering in any format is everything but i would also add not all Hi Rez material is actually what it purports to be. Quite often it may be at lower bit rates and unless you have the means for checking you don't actually know unless you obtain Hi Rez material from a reputable source that checks and assures the validity of the provenance.

I have actually found ripped 16 bit actually more preferable on many occasions.

So what might be the case is that when we purchase a hi rez download of, say, Weather Report "Heavy Weather", the vendor could be offering this hi rez format as a mere up sampling of the good old ripped CD? Hmm, I can do that on my Macbook -- load an AIFF ripped from the CD, then upsample it to 24 bit/96 kHz. I don't think those acrobatics will result in any perceivable difference in the sound quality of the original AIFF file and the 'up sampled' FLAC file.

For most of the experiments we did during the blind testing between formats with different resolution, we were using Paul McCartney and Wings "Band On The Run" reissue. This reissue was offered on a CD as well as a hi rez download. I am assuming that the manufacturers didn't cheat and that they have honestly offered a true 24 bit/96 kHz digitized master tape. Can anyone confirm that?

I recall on an occasion preferring mp3 to ripped 16 bit CD. This can be explained (I think) with the fact that lossy formats tend to introduce certain type of distortion that may appear to listeners as 'enhanced details' (especially when the distortion results in boosted high frequencies). I think these exaggerated highs that may occur during the mp3 playback contribute to the initial fascination with the mp3 playback.

Of course, any prolonged exposition to this mp3 playback inevitably results in massive headache. Which is why we tend to prefer CDs/LPs.

[Rothchild]

If 16 bit was good enough then no DAW would use 32 bits for all audio processing and manipulation internally. Similarly if 16 bits was good enough then no mastering engineer would demand 24 bits as minimum for processing. Everything would be 16 bits since its good enough. When I render a DAW track the 24 bit version sounds better. Not a million miles better but better. When you render 16 bits you need to worry about quantisation noise and what dithering method to use to avoid this, triangular or PoW-R2/R3 since they also make the track sound slightly different.

Meh, recording in 24bit and mixing with floating point (most engines are now 64bit rather than 32 but the point remains the same) doesn't mean that 16 is not sufficient as a distribution format. We need headroom in our processing chains and floating point delivers that, but as long as the level at the master is kept below 0dBFS (which it has to be when you go back to fixed point regardless of whether that's 16 or 24) it's not an issue.

Mastering engineers may have preferences around deliverables but very few will 'insist'. If the choice is 16 or no work they'll take 16bit. Indeed I suspect that if the option was 'well mixed 16bit with a decent slice of headroom' vs 'crappy, too hot 24bit' any worth their salt would plump for the former.

Yes dither has a sound, but it's marginal stuff and mostly affects the quality of the noise floor and things like reverb tails (where they hang in to silence) it's not going to change the tone of a guitar (for instance)

[magiccarpetride]

Meh, recording in 24bit and mixing with floating point (most engines are now 64bit rather than 32 but the point remains the same) doesn't mean that 16 is not sufficient as a distribution format. We need headroom in our processing chains and floating point delivers that, but as long as the level at the master is kept below 0dBFS (which it has to be when you go back to fixed point regardless of whether that's 16 or 24) it's not an issue.

Mastering engineers may have preferences around deliverables but very few will 'insist'. If the choice is 16 or no work they'll take 16bit. Indeed I suspect that if the option was 'well mixed 16bit with a decent slice of headroom' vs 'crappy, too hot 24bit' any worth their salt would plump for the former.

Yes dither has a sound, but it's marginal stuff and mostly affects the quality of the noise floor and things like reverb tails (where they hang in to silence) it's not going to change the tone of a guitar (for instance)

This is similar to my conclusions. I used to religiously track all the instruments while recording in 24 bit/96 kHz resolution, but later on I've noticed that for the simplistic production I am doing, 16 bit/44.1 kHz cuts it just fine. In other words, no audible, discernible differences in the final master, regardless of the upstream resolution.

[swampy]

Meh, recording in 24bit and mixing with floating point (most engines are now 64bit rather than 32 but the point remains the same)

True, I switched from 32 bit to 64 bit daw recently. I guess nearly all pro audio interfaces are 24 bit these days so it does seem to be a standard. And yes a good 16 bit will better a poor 24 bit master but is a good 24 bit master better than the equally good 16 bit one ?

Not that it matters much as I said in an earlier post. Most listen via streaming services now in mp3 or on a smartphone with cheap ear buds and its usually only collectors or audiophiles want the CD or LP and DVD-A SACD are dead so the CD vs LP or 16 bit vs 24 bit matters little in the real world.

BTW, I like your last 2 EP's on bandcamp.

[Rothchild]

This is similar to my conclusions. I used to religiously track all the instruments while recording in 24 bit/96 kHz resolution, but later on I've noticed that for the simplistic production I am doing, 16 bit/44.1 kHz cuts it just fine. In other words, no audible, discernible differences in the final master, regardless of the upstream resolution.

Yeah, don't get me wrong though, I'm saying there are good and valid reasons for recording at 24 bit (and even better reasons for using high bit depth, floating point maths for mixing) just that when it comes down to it, for the great majority of music once it's (properly) dithered down to 16bits for distribution you're not actually causing any significant harm.

David, thanks for listening I'm glad you liked them - I hope they're a good demonstration of what love and attention can bring to a recording, in the absence of expensive recording studios and piles of equipment! Someone (with significant mastering experience) recently told me they sounded 'very analogue' but the truth is they only analogue stage in their production was the recording (ie they are real instruments recorded with microphones) everything else was done in a computer using free or bundled software plugins - it can be done.

Whether they'd sound better pressed to vinyl is a question I'd love to have answered, if anyone wants to fund us to find out please get in touch! ;-)

[swampy]

The section on bit depth is worth a read. http://tweakheadz.com/16-bit-vs-24-bit-audio/