Feb 23, 2012
Mastered for iTunes ‘sounds closer to the the CD’ – I’m calling BS
[Edit: Some people have miss-quoted what I'm saying in this video, using it to suggest that the whole idea of "Mastered for iTunes" is bogus. It's not - the full story is far more complicated and interesting than that. For all the details and my latest comments on this topic, click here.]
So, the “Mastered for iTunes” concept has been getting a lot of press in the last couple of days, and there’s a claim that’s being made that bugs me.
Specifically, it’s being said that the “specially Mastered for iTunes” versions, or so-called adaptive mastering, ‘sound closer to the CD’.
Not true.
Test-ably not true.
Don’t believe me ? Watch the video. It proves this claim is untrue, by using a technique know as ‘null testing’ to highlight the sonic differences.
As I say in the video, people may well prefer the specially “Mastered for iTunes” versions, and there’s absolutely no reason not to make a specific master for a particular release format – but to say that it sounds ‘closer to the CD’ is just wrong.
Let me know what you think in the comments !
Update
Just to be clear, it’s the specific mastering processing that’s supposed to be “optimised for iTunes” that I’m calling out here – the sonic choices made by the mastering engineer.
Apple have release an excellent set of guidelines on Mastering for iTunes, and I’m delighted to see a clear discussion of issues like the loudness wars, sample rate conversion, dither and clipping.
BUT the fact that the new Apple encoders can correctly handle high sample rates, and should make a better job on the conversion, STILL doesn’t mean that the files will sound ‘closer to CD’.
In fact, since at the end of the day we’re still getting a lossy encode, it’s my opinion that the advantages of higher bit-depths and sample rates will be completely outweighed by the AAC encoding.
If Apple really care about giving us high-quality audio, they should offer lossless formats… but that’s a whole other blog post !
Update 2
A few people have asked me where the claim that “Mastered for iTunes” masters sound “closer to the CD” comes from.
To be clear, this isn’t something that Apple themselves have said, to my knowledge.
The phrase originally came from Scott Hull’s comment on the first post I wrote about Mastering for iTunes, discussing the RHCP release:
http://productionadvice.co.uk/mastered-for-itunes/
September 2, 2011 at 4:14 am:
“The goal of the unique AAC master of the Chili Peppers album was to make it sound as close to the CD as possible. The ears involved in the process felt it was a success. Pick up the CD and compare to the AAC file with your ears. It’s really damn close. Keep in mind, I’m not addressing whether or not you like the sound of the CD. It was RR’s goal to make the Itunes file as true to the sound of the CD as possible despite the data reduction. A lot of time, effort and careful listening went into this project. This was not just media hype.”
And, here’s Rick Rubin saying that AAC encoding needs to be compensated for:
http://read.mtvhive.com/2011/09/01/rick-rubin-explains-what-mastering-for-itunes-means/
The video in this post was was prompted by the Ars Technica article recently:
- which included quotes like
“The problem? The AAC compression algorithm is “quite quirky.” Without compressing a song, and carefully listening to it, then comparing to the uncompressed master, there’s no way to predict how the sound will change. Vlado Meller, another engineer at Masterdisk, described mastering for iTunes “like polishing your Bentley in total darkness, then turning on the lights to see where you missed.”
“There are no accurate real-time tools to help you hear what the algorithm will do,” VanDette said. “It was not uncommon to revise tracks three, four, even five times until I got something that compared well with the CD.”
Slightly different words, but the same spin. And still not true.
If you found this post interesting, you might like to check out my recent eBook,
“The Essential Production Advice” – it has all the best content from this site site to help you start improving your recording, mixing and production skills today. For more information, click here.
Related posts:
.gif "Dynamic Range Foundation - Active Member")
I think mastered for iTunes means the AIF or wave file has been encoded with apples own droplet.
http://www.apple.com/itunes/mastered-for-itunes/
When iTunes started the available songs were ripped from CD’s. Now it seems many files at the iTunes store are encoded from the original source. Oh and now at a bit rate of 256kbps rather than 128. Also they suggest headroom of -1db rather than perhaps -0.1db to avoid clipping that may still occur.
We us the consumer notice the difference. Maybe from a CD rushed released when CD were first introduced. With many consumers listening on their MP3 players, they probably wouldn’t notice much difference anyway.
I think there is some truth in this claim, but only with the new encoding, which most wont ever notice.
Hi Nathan,
You’re right – at it’s most basic, “mastered for iTunes” is just higher quality SRC, dither and encoding. I have no problem with that (although I doubt that higher sample rates will be audibly beneficial)
It’s the extra mastering work that’s being done in the name of “compensating” for the encoding that bothers me.
Ian
Ian – Where do you see anywhere where “Mastered for iTunes” is suggested to mean “closer to the sound of the original CD”? It’s not in Apple’s “mastering for iTunes” pdf guide, and it’s not on the Ars Technica article that sent me here. It’s not on the iTunes store page either – it just says “experience music as the artist and sound engineer intended”, and “specially tuned for higher fidelity sound on your computer, stereo and all Apple devices” – nothing about CD at all.
If I google for “mastered for itunes closer to cd” this page is the the first hit.
Hi jsd,
This originally came from Scott Hull’s comment on the first post I wrote about Mastering for iTunes, discussing the RHCP release:
http://productionadvice.co.uk/mastered-for-itunes/
September 2, 2011 at 4:14 am:
“The goal of the unique AAC master of the Chili Peppers album was to make it sound as close to the CD as possible. The ears involved in the process felt it was a success. Pick up the CD and compare to the AAC file with your ears. It’s really damn close. Keep in mind, I’m not addressing whether or not you like the sound of the CD. It was RR’s goal to make the Itunes file as true to the sound of the CD as possible despite the data reduction. A lot of time, effort and careful listening went into this project. This was not just media hype.”
And, here’s Rick Rubin saying that AAC encoding needs to be compensated for:
http://read.mtvhive.com/2011/09/01/rick-rubin-explains-what-mastering-for-itunes-means/
But this video was prompted by the Ars Technica article today:
http://arstechnica.com/apple/news/2012/02/mastered-for-itunes-how-audio-engineers-tweak-tunes-for-the-ipod-age.ars
“The problem? The AAC compression algorithm is “quite quirky.” Without compressing a song, and carefully listening to it, then comparing to the uncompressed master, there’s no way to predict how the sound will change. Vlado Meller, another engineer at Masterdisk, described mastering for iTunes “like polishing your Bentley in total darkness, then turning on the lights to see where you missed.”
“There are no accurate real-time tools to help you hear what the algorithm will do,” VanDette said. “It was not uncommon to revise tracks three, four, even five times until I got something that compared well with the CD.”
Different words, same message.
And it’s all nonsense, as the video tries to demonstrate – the vanilla AAC encodes sound OK – reasonable, for lossy encodes – not “night and day” different at all.
What we need from Apple are lossless downloads, not “optimised” lossy encodes…
I just had my laugh of the month. “Mastered for Itunes” (…) “closer to the CD”? REALLY? What kind of idiotic key-gibberish-manager did they hire to come up with this one?
Your factual, sober way of addressing this neo-voodoo bullcrap, without laughing or calling people idiots even once (as far as I can hear), is impressing to say the least.
I won’t be able to discuss this on my own if people disagree. Your video will be my only hope of getting people to understand. So, yeah. Thanks a lot
@Nathan
Quote: “When iTunes started the available songs were ripped from CD’s. Now it seems many files at the iTunes store are encoded from the original source. Oh and now at a bit rate of 256kbps rather than 128.”
This is wrong. The bitrate change from 128 to 256 was made long ago, way before this “Mastered for iTunes” thing came up.
There might be some misunderstanding things.
People read “this article only” through some blogs like http://9to5mac.com/2012/02/28/mastering-engineer-proves-mastered-for-itunes-doesnt-sound-closer-to-the-cd/
Apple haters will think; “Oh, Apple *advertises* ‘Mastered for iTunes sounds closer to the CD’, which isn’t. What a BS.” – apple fans will think; “No, Apple will be right after all.”
The problem is they don’t know, actually the phrase “sounds closer to the CD” came from Masterdisc CEO, which is commented in other article. It is not featured in any of Apple’s advertisements, nor PDFs.
It would be great if you mention this.
Hello Ian,
Do you have a link to the “sound closer to the CD” quote from apple?
I couldn’t find it on apple site: http://www.apple.com/itunes/mastered-for-itunes/.
I could find the marketing slogan “Music as the Artist and Sound Engineer Intended” which means something different.
It’s obvious that compressed files are not going to sound like the CD version. What I really want to know about though, are the differences between “Mastered for iTunes” files and a comparable AAC encode. I have most of the music on my computer in FLAC format, but for carrying around on my iPhone (listening through Westone 4′s) I usually rip my CDs to 256kbps AAC, as I this is the best quality I can get them and still have everything fit on my device.
My question is, will the 256kpbs Mastered for iTunes files sound better than a 256kbps AAC created from the CD? What differences in the sound might be present when comparing the two? I would also like to know your unbiased opinion vs your personal opinion, if you had to choose between the 2 and money was not a factor.
Neither increasing the sample rate nor providing lossless digital audio files will make one iota of difference to how badly Death Magnetic sounds. All the bits in the world cannot make that record sound good.
IMO it’s pointless to provide lossless audio recordings when many contemporary artists allow their records to be compressed into oblivion.
Thanks Don,
I’ve just updated the post with quotes and links to explain that.
Ian
Thanks for the clarification Ian!
Hi Federico,
It’s not a claim made by Apple, but by the engineers who made the masters – I’ve added an update to the post.
Ian
Hi Pablo,
I did a quick test yesterday, encoding some 24/96 sources I have here using the new “Mastered for iTunes” droplet and iTunes built-in AAC encoder from the 16/44 versions I made for CD release.
The differences on the material I tested were minimal.
So, the answer to the question “will the 256kpbs Mastered for iTunes files sound better than a 256kbps AAC created from the CD?” is – it depends.
- If they’re made from a better transfer of the original source – potentially, yes.
- If the new master is superior in some way – eg. more dynamics – then potentially, yes.
- But if the new encode is made from exactly the same source as the CD ? Probably not.
- And if the new master is inferior in some way (or you just happen not to like it) then, no.
Ian
Ian, I have a question, and it may be a fairly ignorant one, but please bear with me…
To me, the biggest problem with the “loudness” wars, the 16bit/24bit, 44.1/96kHz, lossy/lossless all comes down to the -idea- that once the audio is committed to the final distribution format, it is concrete, done, finished. With Vinyl and tape, that was pretty true. Even with CD, because of limited post-processing power, that held to be true. But in the age of AAC encoders that are software-maliable, that isn’t so true anymore.
Looking at the -science- of the Red Book standard: 16bits of dynamic range and a 44.1kHz sampling rate theoretically capable of capturing up to 22.05kHz signals, that standard is pretty much going to exceed the credible hearing range of 99.94% of human beings. Right?
But, again, the problems creep in when the 16bits of dynamic range are frittered away confusing dynamic range with “loudness” and “gain”.
And mastering uses 24 or 48 bit, and 96 or 192kHz, so as to not hinder/encumber the signal path and act to degrade the final result…which ultimately will be the 16bit/44.1kHz of the CD.
Given that most ALL devices that playback iTunes store music are very capable of greater than 16bit/44.1kHz (definitely 24/96 in the case of the iPod) as well as able to dynamically compute compression and gain, wouldn’t it be possible for Apple to take hidef sources and then push them into 16/44.1 AACs along with metadata “hinting” for gain and compression?
I guess what I’m thinking is, take that RHCP song, it has pretty bad DRC. Let’s assume that the difference between 96kHz and 44.1kHz is unlikely to cause the most dramatic difference (which I believe likely, since most people can’t even hear anywhere near 20kHz); let’s just take that off the table if you’ll allow me. So we’re dwelling on the 24bit/16bit. On a CD, part of those 16bits are used up in gain, in pushing “loudness”, because there is no post-processing. But if we COULD introduce a measure of postprocessing, a “normalization” level and a compression ratio, couldn’t we take those 24bits from the source and more effective fill those 16bits? Where 0 doesn’t correspond to noise floor, but some other relative level?
Basically, what I’m saying is couldn’t Apple have said “Look, if you use OUR device, you’re going to end up with a sound that is BETTER than CD because we can offer better dynamic range and better normalization WITHOUT compromise.” “Remaster your music, move away from the loudness war crap caused by the constraints of the CD…this new way will sound ‘good’ on regular kit but AWESOME on and iPod.”? Isn’t there a path forward that fits in with the “container” of the past, while not requiring the weight of 24bit/192kHz file sizes? Use the “computer” in the decoder to improve the output signal?
I can’t see that you “prove” anything usefull here with your null testing.
There is a large problem doing your “null” test, essentially digital difference between files, when any of your sources are based on perceptual coding(AAC/MP3).
Perceptual coding by its nature makes obvious visible changes to the waveform of a file, BUT in places where they are considered to be MASKED by other sounds around them.
If you do a null comparison, you are taking the areas considered to masked and thus deprecated, and highlighting in the difference.
Taken out of context like this, it does not prove these are audbile differences, because it ignores that these are the differences from the masked sections. All it proves is that yes indeed AAC uses perceptual coding.
It seems clear that you don’t believe in perceptual coding, but you should at least understand how it is intended to work, and the problem of making claims proof based on a null test of audio using perceptual coding.
Beyond that, I read the Apple guidlines, they seem excellent, and if I read it correctly the end product is 24 bit, so they could have much more dynamic range potential than CD for sources that need it (not RHCP).
Great screencast, great voice, great explanation. Thanks a lot for this.
In brief, “Mastered for iTunes” is dedicated mastering with lossy encoding in mind. While, obviously, we can’t compare such music experience to a CD one or, for relative extents, a vinyl one, I find this “closer” take a positive one for customers, artists and engineers, and I definitely welcome it anyway.
Another rumored improvement from Apple is that they are developing new audio file format to offer ‘adaptive streaming’ to iCloud users:
http://www.guardian.co.uk/technology/2012/feb/28/apple-audio-file-adaptive-streaming
Of course, Ian, it’s lossless what we all like, which would, eventually, become more and more the standard – assuming, of course, that people still get the value to get the best quality from the work of artists and engineers, and pay a price. Hopefully, you’ll cover the “whole other blog post” too asap
Thank you for your article!
@ Scott: If you’re asking “couldn’t Apple reverse-engineer AAC to sound better than CD”, I don’t know the answer – but I prefer to ask “why not just offer Apple lossless files today” ?
@ Snowdog: Trust me, I understand all about perceptual coding, and null testing
Otherwise, I agree – the Apple guidelines are great. In fact, I’m very optimistic about them:
http://productionadvice.co.uk/what-is-mastered-for-itunes/
And, you’re right that listening to a null test doesn’t tell us all that much – but in this case, it clearly shows that one lossy encode is closer to the CD than the other. (Both files have lossy artefacts, they’re just less obvious in the MFi null because our ears are distracted by the EQ difference)
And, you’re ignoring the fact that my instinct was that the MFi didn’t sound like the CD – the null just confirmed my suspicion.
To be honest, I see the 24/16 bit issue as a red herring when we’re talking about lossy files – and especially with masters of material like this.
@ Martin: Thanks !
@ Franco: I agree, there’s great potential in the “Mastered for iTunes” idea – again, see the link I posted above. I just hope potential is exploited, not squandered.
Couldn’t the more audible difference in the Mastered for iTunes one just be caused by the resampling filter though?.
As a mastering engineer you’ll know that even the teeniest tiniest phase shift can make a dramatic different in null test, but in actual audible listening the difference is barely if at all noticeable. Kinda like the effect of inserting a minimum phase eq and doing a boost or cut of 0.1db. Boom, suddenly a whole section of frequency comes through on the null test listening.
It’s a hard one to quantify because, listening to your aac coded one nulled there seems to be a lot more wide mush going on. Now it could just be thats also on the Mastered for iTunes one as well but we can’t hear it so well cause of the central high freq stuff coming through louder, but I wonder if maybe this is the compromise and thats where the improvement lies.
In short, a null test isn’t really helpful in this situation in definiting audible quality when using lossy algorithms. The best test is just listening normally switching back n forth quickly realtime playing whilst blindfolded.
Hi Laurence,
The answer to your first question is No – there’s no way either of the filters used (for the CD or the MFi version) could cause such a massive difference, unless they were broken. I’ve also done a quick comparison encoding a 24/96 master I have here with the new droplet versus iTunes, and the difference is absolutely tiny.
I also disagree with your final paragraph – there’s no question a null test is a blunt instrument, but my initial instinct just flicking between the two versions was the the MFi version was a little softer in the upper mids (“edge” on vocals and drums) than the CD, and that’s exactly the region that the null test reveals is different.
I think your second paragraph is the right answer – “it could just be thats also on the Mastered for iTunes one as well but we can’t hear it so well cause of the central high freq stuff coming through louder”.
Ian
I think Mr. Shepherd you are missing the point entirely. The whole point is to have Mastered for iTunes albums sound as good as they possibly can. It is not to make them sound as close as possible to the CD version of the album, which is what you seem to allude to. You later correct yourself in the comments section. Your null test while entertaining, proves nothing of significant relevance. Taking one song from one album which was one of the first albums to bear the Mastered for iTunes moniker, released mid last year, before Apple formerly presented the guidelines, and taking that as representative of the whole Mastered for iTunes process is very short sighted.
I am of the opinion that some of the recently released Mastered for iTunes albums sound very good. What sounds good is very subjective of course. Improved sonic quality of music is the goal of Mastered for iTunes not the quantitive aspects of the audio files as compared to the CD version, which is all the null test does. If anything, the Mastered for iTunes version and the CD version should not be exactly identical. If they were then that would defeat the whole purpose of the Mastered for iTunes process. As you correctly stated in the comments, ‘If they’re made from a better transfer of the original source – potentially, yes’. If a mastering engineer were to use the CD masters to create a Mastered for iTunes version, the end result would be no better or worse than the existing lossy AAC version. Thus to get any benefit one should use 24/96 or 24/192 masters which is what the Mastered for iTunes guidelines suggest.
Again the goal of Mastered for iTunes is improved sonic quality of music not better quantitive aspects of the audio files as compared to the CD version. Comparing it to the CD version assumes that the CD version is the best version available.
If your goal was to prove that a particular album had fallen short of the goal the mastering engineers had set, then you should adjust your post, tone and content of your video to reflect that. As they stand now you seem to imply Apple is trying to pull the wool over everybody eyes, which is not the case.
Again what sounds good is very subjective, and ultimately we listen with our ears not eyes.
Hi Ian,
Nice post. I do have a question about your conclusion though:
As you know, additive EQ is used to compensate for perceived conversion discrepancies in the “Mastered for iTunes” process, and additive EQ, as you know, adds phase shift to the original source signal.
While I don’t doubt your thesis that MEs are human and therefore fallible, couldn’t it also be that this phase shift would have the effect if making a similar-sounding file with a closely-matched frequency response null less significantly than a dis-similar sounding file with no additional phase-shift introduced?
That’s to say that even if an ME did a great job restoring the pre-conversion EQ balance, wouldn’t it be possible that the phase shift introduced by using additive EQ could lead to a less complete null?
Food for thought and something to test, for sure.
Still, this was an interesting test, and does successfully raise the question as to whether this particular song was made closer to the original master or not. (Remember that in the MFiT process each song’s AAC is matched to the original version separately).
But for my standards, I couldn’t consider this conclusive evidence that the whole thing’s a crock just yet.
“there’s no way either of the filters used (for the CD or the MFi version) could cause such a massive difference, unless they were broken.”
..or they were using a minimum phase filter rather than linear.
That would give the characterists you mention, softer in the upper mids, whilst also creating the null test difference. This is something I’ve noticed myself when comparing sample rate conversion using linear vs minimum phase.
Hi Laurence,
I’m sorry, but you’re wrong about that. If you do a null test between those two conversions (linear & minimum phase) you won’t hear anything like the difference between the CD and MFi. There WILL be a difference, but not enough to account for what the video shows.
Ian
@ Dexter – Thanks for your comment – mostly I agree with you.
I’m not sure when you read the post, but I think “Update 2″ addresses all the points you make quite clearly. And I’ve added a link right under the video to another post where I explain very clearly the different ways that people are using the term “Mastered for iTunes”, and my hopes that it might one day result in genuinely better-sounding iTunes files. You say you’ve heard great examples, so maybe it already is – if so, fantastic !
Unfortunately – re-doing the video isn’t something I have time to do – and, I say very clearly within the video that mastering for different formats is entirely valid.
Some sites have embedded the video and linked here with descriptions that don’t exactly match the point of the video – I don’t have any control over that, though.
And finally, I stand by the tone of the video.
This album was the first one to be heavily publicised as “Mastered for iTunes” – supposedly the whole idea came from Rick Rubin’s reaction to the initial iTunes encodes. Rick and the engineers say very clearly that they felt the initial encodes didn’t match the CD, and the processing was to address this. That’s still nonsense.
And, this whole thing about “close to the source” versus “closer to the CD” is semantics. Are they saying the CD isn’t as close to the source as they could get it ? If they are, why did they make a substandard CD master ? I’ve been a professional mastering engineer for over 15 years and I’ve never done anything but try to make the CD sound the best it can possibly be.
So it doesn’t matter whether they say “closer to the source” or “closer to the CD” – those things *should be the same*, in all formats.
It’s all smoke and mirrors – or people who should know better fooling themselves.
Ian
@ Justin – As I said to Laurence, above, the phase shift due to EQ types is just a contributory factor to the difference between the two. The key issue is – they sound different, and if the goal is to make all masters sound as close to the source as possible, they shouldn’t – see my reply to Dexter.
I hear you Ian, and I do appreciate your efforts on this.
I still don’t feel this test is conclusive for a few reasons, and have featured your arguments along with some arguments from the other side in a new article on the subject.
( It can be found here, if you’re interested:
http://trustmeimascientist.com/2012/03/03/ace-engineers-share-tips-and-secrets-of-mastering-for-itunes/ )
If you’d like to devise another series of tests that would be more conclusive, I’d be happy to help. As it stands now, I just don’t think this one covers all the bases.
Thanks and have a good one,
Justin Colletti
http://justincolletti.com
http://trustmeimascientist.com
http://sonicscoop.com
Hi Justin,
The test was never meant to be particularly scientific – I used the word “proof” in haste
I saw your post, and if you’d like to open the comments up I’ll reply
But put it another way – I’ve yet to hear a “Mastered for iTunes” album that sounds better then the CD – and they could – see this post:
http://productionadvice.co.uk/what-is-mastered-for-itunes/
So, where are they ?!
Ian
Hi Ian,
I hear where you’re coming from.
I have trouble finding the time to monitor a traditional comments section like the one you have, but I do love letters from readers, and I’m always happy to print a follow-up!
Feel free to email me anytime, and I might very well feature it in next month’s issue.
Thanks,
Justin Colletti
http://justincolletti.com
I think its another way for Apple to say you can make better music on a Mac. They killed off an important recording software on the PC so its in the companies interest to keep up dirty tactics. Notice you cant get the encoder for a PC. Go F yourself Apple and maybe put some nets around another factory. I used to be an Apple fan but now I cant stand this lowlife company.
Hi David,
I agree there’s a degree of marketing spin involved, but I can’t criticise Apple for the initiative. The testing tools may not be available on PC, but there’s nothing to stop you submitting a master made on PC and getting certification for it.
Ian
Do you think this plugin would fill the gap for the windows people. Granted its not for Ipods specifically.
http://www.iis.fraunhofer.de/en/bf/amm/produkte/audiocodec/audiocodecs/sonnox-plugin/
What are your thoughts on this Sonnox product?
Hi David,
I haven’t used it, but if you want to do “adaptive mastering” for AAC or mp3, then yes – the Sonnox product is the kind of thing you need.
Ian
nice work ian.
Hi,
I do not agree completely with your null test.
Have you read the book on iTunes Music by Bob Katz?
He states on page 44 NOT to compare pressed CD to an Itunes encode or a new Mastered for iTunes encode.
Simply beause a CD is not a master; it’s an end-product which has been derived from a (in your case also) a different master. As shown in your video also. (not to mention dither on a cd, clipping levels and resolution differences amongst other things)
So in my view you’ve just showed us that a different master is used for both these end-products (CD and AAC file) And you’ve also proven what Bob Katzis is telling us is totally true (and you’ve proven his point to be correct).
No.
In my tests, there is virtually no difference in sound between a “Mastered for iTunes” encode from a 24/96 master, and from a 16/44 CD made from that same file.
So the format of the file used for the master makes very little difference to the encoded sound. Apple actually want us to supply 24/96 masters so they have a sales feature for a future “upgrade” to 24/96 lossless file streaming – see this post:
http://productionadvice.co.uk/what-is-mastered-for-itunes/
If different masters are used for the CD and Mastered for iTunes encode, then all bets are off – of course they will sound different, as in the example above. But these are mastering decisions and not required by the format – I’ve had confirmation of this from Apple.
Ian