Bravestar - The Metallurgist's "Tragedy" Mix.

[The_Metallurgist]

This part is a sincere question: I was under the impression that the 32 bit to 24 bit crush isn't a huge deal since modern PCs are capable of handling the long decimals... and the noise floor for 24 bit audio is too quiet to be an issue. What am I missing there? I'm still learning how to hear dither, but with that much headroom, what is there to hear? There seem to be to many opinions on the web to sort out the facts...

thanks for your notes. especially glad you liked the version.

it's debatable whether dithering from 32bit to 24bit is necessary, especially given that i'm finishing the process at 24bit for this particular file, and given the nature of the material in this project! one might actually suggest that some harmonic distortion from quantisation would be a benefit (lol), or conversely, we might prefer to dither and have the hiss rather than add further sonic degradation from the additional harmonic distortions which would inevitably occur. but would they be detectable? most likely not.....so let's be sure and dither, i say.

a crucial factor getting in the way of judgement, is the presence of background noise. this will tend to mask a decent low noise floor anyway, and even a REALLY BAD ONE of which there's plenty of example mixes in the forum.....and issues throughout the MDL are also noted.

and if the SPL is cranked right up, we are far more likely to hear the cumulative noise of all the semiconductors and goodness knows what else in the signal chain, even cables can be a culprit and especially the amps in active monitors! cheap monitors will have lower-grade components which means a greater preponderance to emit more self-noise than more expensive stuff (we have to pay for quality because it costs more to manufacture low-noise components). it is possible that quantization noise could be detected, present in the form of buzz, harmonic and inharmonic distortions resulting in brittle or harsh audio, for example. but i also think genre is a factor...a smashed metal song isn't going to show truncation issues that are detectable even at 16bit...or less But it's not a cut and dried case...and music material and recording quality will influence the decision whether to dither, and even the kind of dither selected.

not many people will be using high resolution monitors or a decent hi-fi setup; such enthusiasts are in the minority, but the fact that some will be shouldn't be over looked in a lossless audio presentation. one regular forum participant actually condemns quality and has derided and even ridiculed some of my views in that regard. don't ask me why

perhaps understanding quantization issues and truncation isn't necessary any more than it's necessary to understand how your car's gearbox works when you push the stick. just so long as you know what the stick does when it's pushed and when it's best to push it. but it is fascinating stuff...

if you dither, make sure it's the ultimate process in your signal chain, and post fader. but unless one's aspirations are towards a commercial grade presentation, i would only bother if wishing to be pedantic

i firmly recommend Bob Katz's Mastering Audio and get the technical stuff directly from the horse's mouth. music based forums do one thing extremely well in my opinion - they help perpetuate myths. myths get in the way of knowledge, but how to tell if one is hearing a myth or not? dither