Past EED rants


Live leaderboard

Poker leaderboard

Voice of EED

Wednesday 10 October 2007

The Radiohead In Rainbows Experiment [Lurks]

In one bit of majorly good news we didn't anticipate, Radiohead came out and announced their latest album In Rainbows (which we didn't even know was near done) would be released soon. Today in fact. Better yet, following on from countless comments of ours on this web site over the years, they decided to embrace a pretty damn brave distribution strategy. It would be available for download with DRM and you can decide how much to pay.

That's brilliant. I only want the download and I think albums ought to be worth a fiver. I'm sure record company execs don't but fuck them. Enough people buy a digital download for a fiver and that's plenty of money in anyone's book.

There are some oddities. First of all I preordered it. Then they sent an email this morning to download it. So I did and I expected it to be slow as fuck, fully expecting to end up downloading it from a well known music torrent warez site instead (but feeling obviously morally justified). However it came down nice and fast and in no time I was listening to it.

This isn't a review of the album. It's brilliant though, a proper Radiohead album. The only real question that remains here is whether this album ends up getting to us the same way as had a record company been involved, a CD bought and then ripped. So has it?

No. There are two issues. Firstly the MP3 is a strange old version of LAME 160kbps CBR. That's a fair bit short of the mark of what anyone has been using to rip to MP3 lately. Secondly, and you know I suspect this is pretty subjective, but I suspect that with no record company to front the dosh for a proper sound studio and the valuable time of a recording engineer, the album itself sounds not quite like you'd expect.

It is, in short, very lo-fi. There's a lot of mixed in guitars with an inherent distortion to them. This done in such a way that to my ears it makes the vocals sound distorted, even though they're not. The stereo imagery is also not exactly subtle. It's like something said "this should be on the left and this should be on the right". Maybe they wanted this effect but I think it's pretty out of character with previous Radiohead albums.

So, in all, it's good news. Lots of artists are watching this and commenting on it. However the delivery could be a little tweaked. We'll get better quality MP3s at some point when the discs ship and people with a clue do the MP3 rip. However the audio mix is probably always going to sound like this. So hopefully it'll just grow on me to match the level of my enthusiasm for the superb content itself.

I do need to be clear about this though. A certain amount of the lo-fi approach I could put down to an artistic direction. So many artists have used it to good effect. I don't think it's particularly well used here, to me the album sounds like it's been produced in an distinctly amateurish way. Some of that could be put down to a pretty bad MP3 encode. Certainly it doesn't help.Portions of it sound genuinely bad which is practically a disaster and pretty sure to have record company execs chuckling into their sleeves somewhere.

Other commentators have been right, in my view, to also point fingers at the whole endeavor as being something that can only ever work when a band has made it big. And they're right too, you know. Clearly only a percentage of Radiohead's fan base is going to fork over money for something they can just as easily grab from a mate or download from a site. Yet when your fanbase is in the multi millions, and given you've just chopped out a significant middleman and, it would seem, heavy production costs too; they're probably not worried about the upcoming price rise price of a can of baked beans.

The thing is, it was basically record companies that helped them get this popular. New artists don't have the luxury of this approach. And you know, what we'd ultimately like to see is that pretty much all music is available from a single store for about a fiver (ten dollars) an album and in a variety of formats to suit your needs. However record companies don't show any sign of 'getting it' any time soon but with this, who knows. Maybe some lightbulbs will switch on in the boardrooms?

I'm guessing that for that to happen this would need to be a resounding commercial success too. And you have to wonder if it will be, in terms that a record company exec would define as a success anyway. After all, he's not going to trade in his Maseratti because the new age of music has less profit margins there for him and his ilk, are they?


  1. Alright my guess about the amateur recording scenario of the album appears to have been on the money. Here's some pictures of the recording environment: - So cut down there's no walls - Nice wood panelling - Nice homely couch


  2. I really only see intentional or ripping as the cause of this which can you leave with a result of either'interesting'and / or 'damnfrustrating' respectively. I'm sure its not the recording environment for a few reasons. In other pictures you can see isolation screens in the recording environment and since everything is close-mic'd in the studio from guitar amps, to the singer, to the drums at a very close proximity of centimetres or a few inches, you don't get any environment in the sound unless you deliberately want it by adding in room mics to pick up room sound.What can cause nastymixes issound from the guitar amps"spill"ing into the drum mics during recordingbut this is prevented by the isolation screens and no different a problem in a pro studio or the room in a house.

    The gear they are using will be their choice by this stage of the career and you can hire what you like, so its not that they recorded on a cheapy which ends up sounding bad because it's ina house. Lots of bands do this these days, from U2's Unforgettable Fire album in Slane Castle to Blur recording parts of 13 in Marakesh to a lot of Gorillaz being done in a shack with wood wallsin Jamaica.

    If they'd been applying a ton of EQ to sounds at the recording stage but were being badly affected by room dynamics in the monitoring room where the mixing desk is, that could have been a problem but no-one adds EQ at the recording stage pretty much anymore so it won't be that.

    If they'd mixed in a bad room, this could throw them badly, but frankly I find this really unlikely. An experienced band instantly knows what a shit room is like. You try out some tracks you know in the room to see what it's like when you start there and to acclimatise your ears for mixing. And even if theysomehow hadn't spotted it, an album like this is subject tomix-listening inlots of different environments like on shitty little boom-boxes, in people's car stereos, on home stereos to make sure the mix is going to work all round.If you work out the mixes sound bad in any of these environments you re-do them.

    Lastly I spose they might have botched it if they'd done it themselves and lacked the skills. You CAN hear this to good effect on the original Hard-Fi album which was recorded in a room above a shop before they had a contract and some tracks on that album are definitely *&!%ed. But Radiohead are massively experienced and moreover they've been using Nigel Goodrich, their longtime producer throughout all the recording and the man's a recording god and thealbum has also been professionally mastered.

    So either its supposed to sound like this or there's something bad in the rip. I suspect mostly the former (one reviewmentions 'amplifiers that sound like they are gasping their last and have beentortured to death by Jonny Greenwood)and a little bit of the latter if you are listening to it on a hi-fi as opposed toearbuds on an ipod whereyou just aren't going to notice.


  3. Beej for teh love of Jebus please fix this wordsandwichshit we're looking really dumb here!!


  4. A few points. I don't accept that the recording environment is satisfactory for just about anything. However you're clearly right in saying that the environment itself is not responsible for the sound. I mention it really, and I should have been more clear about this, as I think it shows their attitude to recording this album. It has a very DIY look to me.

    The poor quality of album is not the rip. I know MP3 well enough after all these years to know if you fuck it up very badly it doesn't sound like that. Actually the audio quality is simply too poor to really tell how much the substandard MP3 encode is hampering the sound so I choose not to cover that at all.

    There's a good deal of stuff which is supposed to sound like that which I accept. Fucked up distorted amps, sounds like a valve on it's last legs. Eg. very bad crossover distortion. That's alright, I don't have an issue with that even.

    The issue I have is the distortion which is clearly clipping, no ifs or buts, it's fucking clipping. And this is prevalent on many sections in virtually all tracks. It kicks in on bass, multipart vocals and multipart keyboards. I've listened to the album about a dozen times today and really I have to come to the conclusion that what they've been doing is simple recording the tracks too loud so when they overlay, they're overdriving. It seems to be overdriving in the analogue domain (thank Christ) so the mixer is running out of headroom to sum the channels being added.

    This is proper proper satanic stuff where audio engineering is concerned. Of course you can throw out the rule book if the result is good but I'd put it to you that the result isn't good. This album *would* benefit from being recorded correctly and that's leaving aside the more subjective issue of the school-boy stereo downmix effort.

    I aint being snobby about this. I don't hold much with that but I do have some audio mastering experience and I think this album is a pig's ear. It's a fucking great album anyway but I just don't buy the fact that it had to sound like this just because they ditched a record label. Radiohead deserve a better sound than this. I can still enjoy it. It's not that bad but it makes me want to keep the volume low and avoid using headphones. And that's really not great. There must be some audiophiles going nuts about this?


  5. Having returned home, downloaded the version (the 7th time I have bought a Radiohead album on day of release including Pablo Honey fuckers ;) ), I can now see precisely what Lurks is getting at. Album opener Videotape has all sorts of distortion and clipping going on across practically every instrument at sporadic points. It sounds like when you push cheap headphones too high and they go 'zzt' on loud notes. It is then totally prevelant throughout.

    Well the clan is too damn fuckin leet to do anything other than fire up Wavelab from Steinberg and have a look at wtf is going on at a track level.

    So to illustrate some differences;

    Track 2 Bodysnatchers starts with Jonny Greenwood bashing out the central riff to the song. Here there's distortion like we're used to with rock albums but it is not the smooth creamy distortion of valves (think AC/DC or Faith No More's ringing electric guitar chords) but there's the buzz and clank and frankly nasty shit of a speaker being driven too far. Zzzt / bzzzt / zzzt. Any kid who ever playedtheir parent's hi-fi too loud knows this sounds similar. At 00.40 a single guitar line comes in that's compressed and in a tight band sonically but it isn't distorted. The rest of the track is mish mashing in and out of this 'blown speaker' sort of sound at points.

    So this could be the original guitar amp speaker, something nasty happening on the recording desk or something nasty happening in the mastering or rip where there's digital distortion introduced. To categorise; natural creamy valve distortion is fantastic for the ear. Human beings lap that sound up. Secondly, blown speaker or overdriven speaker is much harder to take (the original overdrive was The Who and others physically slashing their Marshall amplifier speakers with knives to get this effect - that's where the rock sound started kids). Thirdly and lastly, digital distortion is fucking hideous - it sounds like a nice sound getting its head rammed through a thick glass wall - it's the clipping and blipping people know when overloading their soundcard on a PC....sort of.....

    Firing up wavelab we can see the intro to Bodysnatchers You can see here the guitar and the overall levels of the track are well within the maximum 100% volume level for the track overall. If it was over 100% it would start horribly distorting digitally and that sound is hideous. What this means is that the blown-speaker sounding distortion is at the source level - it is NOT the rip or the mastering. So for right or wrong this is exactly how Radiohead recorded it (my bet) without fucking anything up later. But even if something weird happened thereafter what you definitely know is that it sounded exactly like this to them at the mixing and mastering stage (if it didn't it would be clipping over 100%). This is the sound they heard, mixed down, listen to elsewhere and approved. Later on in the track we get a lot of things getting the same sort of sound including drums and voice. It sounds weird but it's not 'offside' sonically.

    Track 2 Weird Fishes _ Arpeggi starts with a drum beat exhibiting no distortion then pretty soon we get some familiar clipping coming in. Here the waveform might look busy but in the modern mastering culture where loud is better you aren't getting digital clipping until you see significant square waves (big horizontal lines at the 100% level for a significant left-to-right length). There's a few minor clipped peaks but it's in the realm of normal albums these days. It is not accounting for what you are hearing.

    Probably the strongest sign that it is how they meant to start it is on opening track 15 Step. At 10 secs in you already have distorted drums and fucked up bitrate processing from a program called Absynth that I know Jonny Greenwood uses that has a plugin that downsamples stuff to a lower bitrate to get this fucked up sound. Sound is unmistakeable. It's happening at the track level


  6. Jonny Greenwood has commented on the bitrate on RollingStone

    How would you respond to complaints about the sound quality – that 160 isn’t a high enough bitrate?I don’t know, we talked about it and we just wanted to make it a bit better than iTunes, which it is, so that’s kind of good enough, really. It’s never going to be CD quality, because that’s what CD does.

    Which as anyone (millions of us) who is / are vaguely technologically sussed which absolutely definitely includes Jonny Greenwood who is an electronic junky knows is total bullshit. A 320 rip or a FLAC is precisely that quality and these guys have been recording with digital and analogue gear for years and years and years and using pc / mac based programs where all this stuff is completely in front of your nose. For the first time I smell bullshit about Radiohead. Sad and bad.


  7. "What this means is that the blown-speaker sounding distortion is at the source level - it is NOT the rip or the mastering."
    It's not the rip but you can't say it's not the mastering. A mixing desk has many channels and in the process of recording there's a downmix made resulting in a single master level out which really has no baring on whatever earlier per-channel levels were. The distortion happens on certain types of things as I indicated. It sounds to me like these have been recorded per-track and mixed together for laying down as master track. What you'd use overlaying for in the old days.

    But then later on you say they must have known. So we're splitting hairs between whether it's deliberate or inept mastering. It's impossible to tell I guess.

    Regarding the MP3, Greenwood is talking rot and that is disappointing. Why shouldn't it be pretty much CD quality for all intents and purposes? Do they believe that basically the download should sound notably worse? Another thing that's impossible to tell is their master level prior to MP3 encoding but certainly from your level plots the MP3 is recorded far far far too loud. On 15-Step you've got peaks at 85-90% of clip. There's no other way of saying it, that is fucking grossly inept.

    For me that makes mastering ineptitude more likely than design. Although it could be once they started getting this horrendous sound they collectively started rationalising it as being by design.

    P.S. Those jpegs could really have done with being resized.


  8. Unfortunately its all part of the 'loudness war' which has seen mastering levels pushed through the roof in the last twenty years because louder is perceived as better. This has squashed all the dynamics out music in order to make it as consistently loud as possible. It has got so bad there are online petitions for the remastering of Californication by the RHCP which is pretty much unlistenable even to people who are less sensitive to it and even old muso stalwarts like Rush have been roundly criticised for it.

    This article explains it simply and has an extremely stark picture of a waveform from 20 years ago vs a modern one. One look and you'll get what this is all about.

    Over compressed / loudness boosted music is tiring to the ear, it robs the music of loud / soft dynamics which come through marvellously on older recordings and it appears only classical music is really protected because the audience there are too experienced to hear the music without the volume dynamics of loud and soft in it.

    For all the advances in technology, music recording quality has been assaulted in the past few years and now we get shit rips as well. I just wish more people were aware of how bad things have got.

    p.s. on the size of the jpegs I was trying to make them big enough so the scale could be seen but yet small enough for people of work puters to see them. Oh well.


  9. Indeed, I know the issue well. The thing is the loudness war lead to an industry stablisation, if you can call it that, of an average level of around -10dB. Mate, in those recordings there, it's less than -2. It's proper proper wrong.


  10. I have just loaded parts of this album into an audio editor after hearing horrible digital clipping distortion throughout. The levels don't actually enter clipping territory whilst playing, but the waveform shows highly compressed levels. If I had made a recording at too high a level, causing digital distortion, and then cut back or normalised to just within 0db, the actual volume level would have been curtailed, but the distortion would remain. This is how the waveform of this album appears to me.