Audio mastering isn't "just" the icing on the musical cake or the final polish — since it's what comes last before the world hears your tunes, your release may very well shine or suffer because of mastering. A botched mastering job can destroy the listener's experience, and even if they don't know all the technical terms that an audio engineer does, they feel something is "off", "wrong", "muffled", and soforth.
That's why I'm disappointed by what a poor mastering job Imogen Heap's newest album, Ellipse, has. It shocked me because she's a superb songsmith and surely, wouldn't allow her voice to be relegated to at-points cardboard gurgles. The credits list Simon Heyworth as the mastering engineer, and I've enjoyed other work he's done, so by no means do I jump to conclusions to fault him. It seems a different balance was warranted here — or something big got missed! But at least it's the inverse of the artists telling the hapless engineer to "MAKE IT LOUDER" whilst sacrificing contrast.
I've hyperacusis so my ears aren't the goldenest, but I searched, and found others who've drawn similar, independent conclusions:
- T. Tom's review on Amazon – "… get better engineering/mastering to achieve better dynamics because this album sounds like a demo tape…"
- MuuMuse review – "No matter what adjustments I’ve made to the volume, treble, or bass, Ellipse still sounds like being played just out of reach of the speakers."
- Color Theory observations – "The mastering is driving me crazy though! The dynamics are not only lost, they’re reversed."
Here's Imogen's original version of "Tidal", as officially embedded on SoundCloud (the CD version is essentially the same):
Here's my 5-min. quickie remaster:
… wherein I've not merely boosted the volume a bit, but added punch and clarity to the dynamic range. This avoids the ugly, overcompressed attitude which has claimed many casualties in the loudness wars as you can observe from the waveforms, while at the same time being more competitive with other contemporary recordings. Overall, Imogen's voice is richer and fuller. The difference can be subtle at times, but so compare back and forth. And if you have subwoofers, let them kick in!
Note that:
- These are web-friendly, compressed 128kbps MP3s. There's bound to be some spectra loss and there's also distortion in the original I couldn't fix.
- In my remaster, the volume fades out every now and then because I own no rights to this work and this is for demonstration purposes only. If Imogen or any of her reps want it taken down, they can contact me, and hopefully get a better mastering job in the process.
See and hear why mastering matters?
{ 2 trackbacks }
{ 7 comments… read them below or add one }
Ah, so I'm not crazy!
As a mastering engineer, I'm often made to feel like I'm the only one who cares about such intricacies. Thanks for bringing it up! I'm disappointed that for all the excitement I had built up going into this release, I never ended up buying it. In fact, I barely made it through a single listen, because to my ears, the dynamics issue is a deal-breaker.
Brian Hazard
http://www.resonancemastering.com
Dynamic range and compression have been issues for recording engineers since the 1970s. There's a new campaign in progress recently:
http://www.turnmeup.org/
I signed up on that site at least a year ago, maybe two, and never heard anything. I don't think the site has changed since then.
Holy crap
Imogen get in here
It's funny you should express concern over the dynamic range… If you use Bob Katz dynamic range meter available from http://www.dynamicrange.de/ (which is the now official organisation to raise awareness of "the loudness wars") you will find that your version has a DR of 5 whereas the original is DR 7… 7 is a poor result for the style of music in question but 5 is obviously significantly worse. Just a thought…
@Charlie I express regret at not being a better transient design specialist — please show me what you can do! (And of course, numbers aren't the only thing, some music really does lend itself well to being squashed *waves to Merzbow* but Imogen Heap's voice sounds airier after what I did to her.)
Well I can't do any better without an original unmastered and uncompressed version of the mix – to reverse compression and retain the natural transients and dynamic flow of already over-compressed audio is difficult… (perhaps I could do some tone adjustments without obliterating it's dynamic range further than it has been already – if you want me to have a go send me a 16bit 44.1khz .wav or FLAC or some other lossless format)
The original wasn't actually too badly over-compressed although I think there may be some kind of overall phase issue within the mix (some instruments played in reverse phase or whatever, not to cancel each other out but just not playing at their natural phase) or it's simply not a great mix – I really don't think the mastering engineer can be held fully responsible here.
I totally agree the issue is vastly more involved than just numbers and actually among those who have the correct vocabulary for issues/negative characteristics in music processing/mastering it isn't actually too subjective either. For listeners who can't identify and describe such artifacts the music simply doesn't feel right – the emotional communication becomes weaker, they simply don't enjoy the music to it's full potential.
Perhaps when playing both tracks back with the same amplifier volume, your version may initially sound better – the key thing to do here is match both versions to the same RMS output from your replay device i.e. using the actual amplifier/output device volume knob.
If you do this you will realise that the original is actually much more involving with it's greater dynamic range and transients. This is easy to demonstrate if you play the original then switch to yours but turn down the volume so they actually sound at the same loudness –
I just felt compelled to post because you stated yourself that you added to the dynamic range when actually you decreased it further.
The whole "loudness wars" is about getting people to use their volume control to turn up the music and hear it in all it's glory, transients and all, rather than letting the mastering engineer unaturally force this loudness on us giving no choice but to listen to something that's devoid of any realism.