After a certain point, lossy compression doesn't create any perceptible loss in audio quality as long as the rip is done well (with a good encoder, etc.), and you don't know every part of the piece in question.
e.g. in classical music, you can tell subtle differences if you've listened the piece live or performed it inside an orchestra. However, that's a pretty edge case. There are always differences if you know where to look for, otherwise it's pretty insignificant.
e.g. in classical music, you can tell subtle differences if you've listened the piece live or performed it inside an orchestra
Can you explain this a bit more? How is having heard a piece live, which by definition means a unique performance, going to affect whether or not you can pick out whether the recording of (likely) a different performance has been put through lossy compression? Or do you mean a recording of that same live performance?
I mean I'm used to subtle differences coming and going depending on room/speakers/crappy compression, but only because I use the same source as reference (say, the same CD). Using a live performance as reference sounds strange, because there I would be able to here one of the musicians doing something different, but difference isn't there on another recording so not usable as cue for hearing differences in sound reproduction.
In classical music, the performance is of course unique, but the piece is not. What I mean is, even if the arrangement has changed for a particular piece, the underlying score, the foundation is same.
As you know, classical music is layered. It can be scaled for different sized orchestras, which can be akin to tessellation in graphics. You can add more nuanced scores or details if your orchestra has enough members. Of course this has a limit, which is the full score written by the original composer. Similarly, you can remove some layers or simplify the piece if you're smaller orchestra without compromising the piece.
What I tried to say is, if you've listened the piece from or performed with a relatively big orchestra, you'll know that which instrument shall be there, where the small optional triplets are, how the piece should sound or where's that little oboe shall come in, where the little cymbal adds that little crash, or how the harmonics affect each other and create that atmosphere.
So, you'll notice something is missing or off or not as it should be especially in the high end. Classical music has a lot of perceptual tricks under its sleeves to create a specific ambiance and sense of space and most of this lays in the higher end of the spectrum, and they get shaved off first with lossy compression.
Hope this helps, because it's something more felt than can be said with words, how you can't really hear the double bass but feel how it's there. It's that kind of perception.
Edit: Just wanted to add that one musician's or orchestra's specific style of course will be different, but a good orchestra is very faithful to the original score of the piece. Even if an orchestra is playing a little fast or more aggressive, or a simplified version, base instrumentation and atmosphere is the same (as long as the orchestra is not doing Metallica S&M style play the right thing with wrong instruments kind of deliberate arrangement).
Another extreme example would be the band Pink Martini. They have an on-stage audio magic which allows them live with the exact sound of their studio recordings, albeit live. It's surreal to experience.
I sort of get what you're hinting at, but I still think it might be inaccurate; to me your reasoning come over like 'played live there's detail X and Y, when listening those details might be vague or don't come out properly, so that might be lossy compression at work' (please correct me if I'm wrong). Thing is: just poor microphone placement or poor recording equipent or poor mastering can have those effects as well, no?
In my comments, I assume that the recording and mastering is done indeed properly. If you can't carry the orchestra's sound to the playback medium, everything is already moot to begin with.
The thing I'm looking is musical dynamics rather than details itself, but it's equally lost with poor recording and mastering as you say, since they're also captured by the microphones. The thing I'm trying to explain was they are not "finer details" like "oh! I hear the bow of that player", but a bigger feeling that the orchestra creates by playing together, and that effect is independent from individual instruments, most of the time.
It's a somewhat difficult concept to put into words and explain. It's more about feeling the music and decoding the brain, and I think it needs some experience. Being unable to translate this into words makes me sad, because it carries music to another dimension IMHO.
It's a somewhat difficult concept to put into words and explain
Don't worry I understand what you mean wrt dynamics etc, it's just that I'd never thought of linking it to lossy compression, because there are so many other things which make it hard to reproduce that live sound.
But that still had nothing to do with comparing lossy vs lossless, or am I misunderstanding you?
How does a live performance that you hear with your ears at a specific place in a room help you pick out missing parts in a different recording, played by different people in a different place, recorded with multiple mics and then mixed and mastered?
It seems plausible to me. I assume that when you are doing a comparison, you are comparing a single source to a memory (does anyone do comparisons by playing two synchronized sources together, possibly into different ears?) In that case, I can well imagine that listening to multiple live performances primes one's mind to remember clearly how a given presentation sounded, and to pick out small differences, precisely because live performances are all slightly different. I would further imagine that performing a piece, and particularly practicing with the rest of the orchestra or conducting a practice, further enhances one's ability to notice and characterize small differences.
Of course, this might be utter nonsense, and I will bow to bayindirh's judgement on that!
Of course you can train your ears, to be able to hear more detail and learn to differentiate and identify frequencies.
That'll definitely help with hearing differences between two tracks. But i don't think you can compare live music to recorded tracks. Especially acoustic instruments, the room is such a big factor with those.
I am not actually suggesting that one should compare recorded music to live performances for the purpose of comparing audio encoding technologies. Oddly, your reply to bayindirh is in complete agreement with what I wrote here.
I think he explained it a little better. I totally understand that you'll learn to identify which frequencies and sounds belong to which instrument and in turn learn to identify when those are missing.
I guess that also teaches your ears to identify differences in other situations more clearly.
That's what mixing and mastering engineers practice their whole career and get really good at.
It has, but in a different w.r.t comparing different sound systems with the same recording. Let me try to explain. You might know some of the following, sorry if it's a re-explanation.
In a proper concert hall, sound is expected to be homogenous, so you should be able listen to the orchestra equally well, with the same sound balance (or mix) regardless of the place you sit. Similarly, recordings are done from suspended or positioned (and ideally tuned) mics, so you can capture the orchestra as someone sitting in the audience. At least this is how our performances were recorded.
The mastering is then done to match the recorded sound to the hall's sound, and balance any imperfections or clean the orchestra's inner talk between pieces (yes, we communicate a lot :D ).
When you listen an orchestra live, you will have a lossless blueprint of the piece in your mind (track by track if you can separate the instruments). If you can get a recording of the same performance, you can compare it with the live performance. That's absolutely correct.
But if you listen to a recording of a different orchestra playing the same piece, the arrangement and instrumentation will be same (you may have 8 violins instead of 12 but, violins won't be changed by violas most of the time). So, the atmosphere of the piece will be the same. Assuming the recording is done by competent folks, the spectrum would be the same (~20Hz -> ~20Khz roughly).
After some point, even if you're listening to a different orchestra, you can start to point to the things that should be there. It's very hard to describe, but every instrument has a base sound and details on top of it (you can tell they're all trumpets, but different brands or models. Similarly you can tell they're double basses but they're different in some ways). That base sound starts to erode too when you have a lossy compression, and in turn it affects the sound of the piece, regardless of the finer details (which are mostly affected by resins, bows, styles, etc.).
It's a "these two instruments shouldn't interact like this in this piece. Something is missing!" kind of feeling. This missing part is either something at the high or low end, almost an harmonic. It's not noticeable unless you're looking for it, but it's there.
That difference can be clearly heard by re-encoding a FLAC as a high bitrate MP3 and taking their differences. It's a hiss-like sound by contains a lot of the said harmonics and you can almost listen to the piece just by listening to it. Someone did that and published the differences, but it was some years ago. I'm not sure I can replicate or find the article. That article took differences of the exact same recording but, it can be applied by your brain to different recordings after some time.
Hope I've succeeded to clarify it somewhat. It's something very hard to describe by words. Please ask more questions if you want to. :) I'd be happy to try more.
Comparing two orchestras can be similar to comparing a recording in MP3 to FLAC.
I think i get the point in that learning to listen to those details and recognize frequencies can enhance your ability to spot differences in encoded audio.
I don't disagree with the fact that you perceive more in a live performance -- after all there's a wealth of spatial information that you don't get in stereo.
But that has absolutely nothing to do with compression. All that would matter is whether you're missing the "nuance" or "layers" that are there on an uncompressed CD, but that you would perceive to be gone in MP3.
I've performed and listened to a ton of classical music in my life, and I've never heard a difference in what you're talking about between CD's and MP3's. It doesn't really make any sense in terms of how MP3 compression works, either -- the compression artifacts it introduces are pretty orthogonal to nuance in classical music. at 128+ kbps
It sounds to me like you're describing the difference between a live concert and an uncompressed stereo recording, no matter how well it was mastered.
e.g. in classical music, you can tell subtle differences if you've listened the piece live or performed it inside an orchestra. However, that's a pretty edge case. There are always differences if you know where to look for, otherwise it's pretty insignificant.