Something is bugging me again. I encoded to mp3 for a while around V4 / V3 and was usually happy, very convinient. I know there can be slight artifacting, but I tell myself that it will be slight and short - and this is usually correct.

In the last week or so I've experimented again with the new wavpack 4.5b and encoded a few albums @ 270 k and discovered that the quality is very good while staying much smaller than lossless. I can create correction files if needed. This is not as easy as pure mp3 - nothing is. The problem is hard to explain: I felt less preoccupied with possible flaws with Wv than mp3. I know that at transparent bitrates most things are the same perceptually, but in when I start looking mentally into flaws with Wv my mind shoots back saying there are none but a possibility of slight hiss / noise. With mp3 and co. I've heard weird stuff like ringing , distorsions, pre echo, tremollos, whistling etc. Is it a matter of taste - artifacts vs noise ?? is it something else ?

Do you mean the imagined problems are different depending on what you expect?

Or the real ABXable problems are more likely to be heard by you depending on your expectations?

Or something else...?!

Cheers,
David.

I imagine them yes and the same can happen with Wv - imagining hiss. Then there are many real flaws in the actual CD recording that sound like they could be artifacts. The thing is I can rule out Wv on the spot because I know 100 % from abx that it has no artifacts that I 'hear' less hiss. Other times i am sus about tremolos, ringing etc. I know it exists at lower VBR settings and sometimes at V2 or so because I managed to abx it on some of my music. On the other hand I also know Wv issues, But somehow its not the same. the problem is that strange sounds on CD's - artifacts or not trigger a reaction in my mind. With Wv it can only be a hiss or so and not on 'top of the music.

We are never really objective. Often we can't be no matter how hard we try. We should accept this.

I think that's your problem.
You know you can be satisfied in practice with, say, Lame -V3. Exceptions to satisfaction are rare, and if they occur, they usually aren't a big issue. You said it so often and it's true.
However you know that mp3 can behave pretty bad (usually at low or rather modest bitrate), and this brings negative emotion to mp3 in general.

Your wvPack lossy experience tells you that - roughly speaking - the only bad thing you have to expect is a slight hiss. It's a bit oversimplified cause wvPack lossy can sound distorted as well at a low bitrate (in wvPack sense), but with new dynamic noise shaping this is expected to be overcome most of the time. So you have a positive bias towards wvPack lossy dns.

I think it's not objectively justified to prefer wvPack lossy @ 270 kbps over say Lame 3.98 ABR 270.
In both cases you'll be satisfied with extremely few exceptions, and we will never know which one is the better solution.

So the best way to go is to stick with your emotion. It's fine as long as it's not in contradiction with facts.
In the end everybody behaves like that.

You are right I can't say one is better or worse @ 270 k..before --dns Wv also concerned me at rare moments but in a different way. Wv is transcodable, mp3 is hit and miss with transcoding.. Wv has correction file options. Mp3 is more compatible and sounds good at even lower bitrate.

That's a real mess once you have started taking care about encoding problems.
I suffer from it too, I'm afraid of encoding problems when I hear something bad, only to find out it's the CD content which isn't ok.
It would be fine to think like this: with very high bitrate encoding (more or less no matter which way) the top audio quality problem is the recording as it comes from CD - so many distortions, especially with vocals towards which we're so sensitive. Encoding problems are next to nothing compared to this. Unfortunately it's hard to get to this very sane kind of thinking.

Just a practical proposal as you're bothered about these things like me:
Why don't you go higher in bitrate with wavPack lossy? While the situation remains the same in principle, you can look at it in a more relaxed way - wvPack lossy @ 270 kbps is a bit on the cutting edge.

I am experimenting on how far I can push Wv to maximize hdd space. It seems you can't go much further without risking noise. I think a non demanding listener can get away with 240 k on rock / pop music but below that it will fall apart.

Without correction file I would go around 400 k . In fact its now going to be nearly impossible to expose 350 k -hx3. If it takes 5-15 years for someone to find 1 or 2 samples I don't really care. The bitrate is often than half of lossless. I think eventually I will do it this way or lossywav.

Regarding the fundamental question, "Is it irrational?", only if it serves no purpose to you personally. It's obviously giving you peace-of-mind if nothing else, so it's perfectly rational in your specific case IMHO.

Cheers, Slipstreem.

To me this sounds very reasonable (of course it does as I'm doing it the same way, just at an even higher bitrate as I don't use a lossless archive any more and don't have to care about disc space).