[erl]

This is a follow-up to a phenomenon mentioned in passing in the thread "Comparison between BH and TBH" in the Headroom forum, which I think deserved it's own thread.

Eagle_Driver wrote:
Quote:

Now, I've heard modern, mostly acoustic recordings - and they all sound hollow and artificial to my ears, even on a relatively high-end home system. And in my experience, digital is NOT better than analog. The trouble with most of today's digital recordings is that the engineers have actually misused the technology: Unless I go for the classical music recordings that very few people around here buy these days, all pop/rock/rap and most jazz CDs have suffered from severe compression, caused in part by the fact that most such CDs were mastered too loud (or too hot). In fact, this race for maximum loudness has GOT TO STOP! Otherwise, we'll all be stuck with nothing but a big, steaming pile of dogmeat for available recordings.

There was also a reference to a list of recordings graded by loudness

I don't understand the reasoning as to why "CDs have suffered from severe compression, caused in part by the fact that most such CDs were mastered too loud (or too hot)."

Being a software engineer, I would think that the highest quality (lowest quantization noise) would be achieved by calibrating the recording such that the highest points on the waveform use the extreme available sample values (ca +-32767). That way quieter portions of the recording would use higher sample values and have a lower quantization noise.

Does Eagle_Driver at. al. mean that the recordings are amplified so much that the samples don't fit in 16 bits, and that the recording engineers compensate for that by compressing the samples?

 

[lini]

Quote:

Originally posted by erl Does Eagle_Driver at. al. mean that the recordings are amplified so much that the samples don't fit in 16 bits, and that the recording engineers compensate for that by compressing the samples?

Digital clipping is indeed a not so uncommon problem. There are quite a few cds where one can find that. For dynamic compression, my opinion is a bit split, as it's usually done on purpose (like with fm radio for example...) in order to give better results on lower quality equipment (boomboxes, microsystems...) at lower volumes or in listening situtions with higher background noise (car, portables...)... If it's done carefully, I'm not that much against some dynamic compression - but it's overdone rather frequently...

Greetings from Munich!

Manfred / lini

 

[kwhead]

I also read that post and article with some interest. I too think normalizing the recording for its peak amplitude to be near 32768 would seem to lower quantization errors as well. Strange...

Well if it does in fact clip that'd be bad...

If they are in fact using dynamic compression to be able to turn up the gain that'd be bad too...

What seems odd to me is that most "pop" recordings have lower dynamic range than other genres so I don't know why you'd compress more (to be fair, any acoustic recording can have pretty huge dynamics)...

Anyway, sorry, that wasn't much of an answer, just more questions and thoughts, but I too wonder what the root cause is...

 

[davei]

Look at the compression thread in the Members Lounge.