Sunday, 2 February 2014

Muse 2nd Law Dynamic Range Research (Avoid the CD version)

Dynamic Range Measurement and Visualization of the Previous Muse 2nd Law Listening tests

So I wanted to do the Muse listening test without having researched the available versions beforehand and prejudicing the listen.  However researching afterward, shows that there was quite a lot of material saying that the CD version was definitely not the one to buy, having a very compressed dynamic range.  So at least I wasn't imagining things, and pretty happy to have heard things backed up by the technology, but it wasn't actually the best back to back to have selected to work out whether 24/96 was worth it.

It also turned out that the internet is full of angry people, with quite strong opinions on dynamic range, and it gets pretty complicated quickly as there are a whole bunch of different ways of measuring it, a fair amount of suspicion around the pro-vinyl brigade's measuring practice, and also on the practices of recording companies reflogging remasters or format copies into HDs on the internet. Also a piece of music without a big dynamic range, might be meant to be all loud, so it's not necessarily a measure of quality.

Simply though, my summary would be that the quiet bits should be quiet, and the loud bits be loud.  To make things sound better in cars, compilations and CD shops a lot of CDs are mixed so the difference between the small bits and loud bits is much smaller, compressing the range of loudness so it's all loud.  Things can also get too loud, and have to be chopped off, with a variety of limiting or clipping at the top.

So could you see the difference in the tracks if you looked at them, and how bad was the compression on this CD?  It turns out, yes, it's blinking obvious, and yes it's really pretty bad.

You can get an overall number for your Dynamic Range of a track with Foobar (whether it's measure right or not is a bigger question) with this plug in.   There is also a database of these numbers you can look up, so comparing different versions, remasters, and formats can be done before purchase.  This is something I think I will be doing in the future.  It's not that low is bad, but if the CD is much lower than alternative formats, I think I'll be thinking fairly hard before purchasing

CD 16/44.1 scores 5.  This is bad.

The HD 24/96 scores 9.  Decent

The Vinyl apparently scores 10, but I'm just focusing on the Digital as I don't want to get into digitizing the vinyl to measure at the moment.

Sicking it into Audacity it became immediately obvious how compressed the CD master is.  I've colour coded to accentuate, but the latter section really shocked me.  The CD is at the top and the HD is at the bottom - this is the Madness track I was mostly listening to.  I really wasn't expecting it to be so immediately visually obvious.

Amusingly I was mostly stopping my listening before the compression got really bad!  No wonder if I let it play on it got tiring.  The CD track bounces off the top, all the time, and the latter section has almost all the dynamic range wiped out with compression.  You can see on the HD track how it should look more like with the the loud bits going very nearly to the top, but with lower bits in-between.

I think it's amazing how much the CD has been compromised by something that has been deliberately done and not anything to do about the technology.  This seems from reading to be worryingly common.  So buying Vinyl or HD is much more likely to get you it looks like a better recording, but it's not necessarily for anything to do with the format, but rather the market in which CDs are being sold in.

Annoyingly it took me about 2 minutes to re-sample to 44.1k and save to 16bit WAV the 24/96, and you get the right uncompressed range, in the 16/44.1 format.  So basically it looks like a new back to back listening is required: HD against the re-sampled down to CD quality HD

No comments:

Post a Comment