Yeah, -0.1 to -0.2 for CD as it is a wav format, I don't know if you missed I meant that -0.5 before converting to Mp3 may be enough too as some people state. I don't know why as I've never heard clipping, converted full-power 0dB Wav to mp3.
1dB attenuation seems like sacrificing too much loudness, but you gotta be careful with some things like streaming that's right
I hope there are alternatives to such codec preview in offline mode
Hmm.. on my stuff it always seems to need -1 to -1.3dB. I'm curious to know what makes the difference. I'm not technical enough to know why or how it happens, it's just something I observed when messing with Ozone.
Loudness isn't a big priority for me. I think most of the industry will go towards auto-levelling like on iTunes and Spotify.. and those standards are lot less loud than what everybody is used to. There will always be a place for ultra loud squashed stuff, but I really like the extra dynamic range that it gives and want to use that to my advantage.
What I normally do is make a 24bit wav that peaks at -0.3dB. That's the one I play for myself and keep as a master file.
To make an mp3, I lower the volume by -2db, convert it to mp3 and on that file, I limit at -2.0dB so I'm just shaving the newly added peaks and amplify it back up to -0.3db.
I could set the limiter to -2.3dB and catch them all, but oh well.. call it the 80/20 rule.
If you're just sending your WAVs to a distributor or aggregator who then pipes it to DSP's you have no control over how it gets converted... it might get re-encoded ten times automatically. iTunes will make AAC's, Soundcloud MP3, others will have proprietary formats. The only choice you have then is: do I want to live with the possibility of clipping (which is minimal, hardly anyone notices) or do I sacrifice peak level for pristine conversion? I guess that's why the new loudness standards have lower maximum peak levels, like at -1dB... to account for this effect.