AIFF vs ALAC with iTunes?

To clarify, I think the effect is actually slightly softer, but more natural. Less of a digital style sheen that can come when a recording is nearer to the edge than it ought to be.  Not quite so upfront, and further back with detail more integrated ...

Is this right? Possible? Or am I too far gone, and describing a nonesense.

I am asking if it is possible really.

ATB from George

Hi George -

Anything is possible, even hearing a difference between ALAC and AIFF (or FLAC and WAV).

I believe you are using a PC and an external DAC, correct?  Well, the reason why Naim chose not to support a USB connection on their DAC was because of how difficult it is to prevent a PC's electronic noise from being transmitted over that type of connection.  You may very well be hearing an increase in noise caused by an increase in processor load due to the need to uncompress ALAC, and when you listen to AIFF, the processor load and noise level is lessened. That's the theory anyway! 

Also, there are a number of tools out there than can batch convert your ALAC files to AIFF, so it would not be much work for you at all.  May take a little time, overnight perhaps, but it is easy enough to do.

Good luck!

Hook

Dear Hook,

Yes I am using a small PC [external PS like a laptop] and USB DAC. It runs totally cool even with the slowest fan speed. I am going to try to rationalise this, but now I have found the difference - after undifference on my own part as I did not believe it! - I have identified what it is that is different, and it is ultra-subtle. Like stepping back two rows, but sitting in a nicer clearer hall.

Less demonstrative, and yet better at a slightly quieter setting than before. I converted a couple of 78 transfers, and the effect is even more marked. I guess that 78s are more of a challenge in any case. 

I am frightened to use a batch converter, as the whole lot could be ruined in a night of mayhem. First I have to buy another external drive, so as to back up today's position. I have one external drive here that offers the perspective from five months ago, but I have fixed a huge number of tagging issues - how roughly these are done! Wrong publication catalogue numbers, and even wrong musical keys given. In three year I have fixed literally thousands of errors that come fromt the record companies! 

That work is also worth saving!

It makes the search engine a very sharp and effective machine.

But this SQ advance is worth a good deal. Like the difference in the old days between the CDX and the CDS2, which I chose over the CDX!

ATB from George

Backing up to, and then regularly maintaining, a separate external hard drive is a really good idea George.

Whichever tool you pick, it would also be a good idea to run some small test cases, just to make sure all of your edited meta data comes across ok.  In theory, unless you tell the tool to replace the ALAC copy with an AIFF copy, it should simply make a 2nd file in a separate directory.  But once you grow comfortable with whatever tool you pick, it should be straight forward process to select all files (or your top most directory), and hit go.

Personally, I can not hear a difference between FLAC and WAV, but I have absolutely no doubts that other people do. With your keen ear for classical music, you are perhaps more able than others to recognize these subtle differences in presentation.  But the theory of why they exist is solid, and it exactly what Naim has been telling us for a few years now.  Also, it is why some guys prefer to place a USB-to-S/PDIF digital-to-digital converter between a PC and a DAC, even when the DAC has a USB connection -- lower jitter, and better noise isolation resulting a lower noise floor.

Will be interested to read if you hear these subtle differences across your entire collection, or if they are more obvious with some pieces of music versus others.

ATB.

Hook

Dear George,

This is fascinating. Although I am certainly no authority on digital replay it seems to me that it would perhaps be more surprising rather than less if there were not audible differences between file types. After all, just about every other variable seems to have an effect on the sound!

Something I wonder about here. You have listened to files converted from ALAC to AIFF. Have you tried actually ripping to AIFF, assuming that is possible, and comparing with a rip to ALAC? I just wonder if the conversion process is introducing artifacts, albeit pleasant ones? Or indeed if an actual rip to AIFF rather than a conversion may be even better.

Best,

Peter

PS. Just thinking out loud here - with my (lack of) computer audio knowledge I may be talking gibberish.

Hi George

A few months ago, I went through the whole" which codec shall I use" debate between WAV, FLAC, ALAC & AIFF.  After taking into account the need to replay music via both itunes (for most of the family) and Naim (especially for me!), plus some back to back listening tests, I ended up with AIFF.

At the Signals DR power supply demonstration, the difference between FLAC and WAV encoded versions of the same track was not subtle when replayed via a top flight NDS/500 system, with WAV versions having demonstrably more life, air & vitality.  The FLAC versions sounded as though the life had been squashed out of them.

Perhaps the ALAC codec crushes more of the life out of music than the AIFF one.  Whatever the theoretical differences between the 2 (which I certainly don't understand), AIFF just sounds a bit better to me.  However, once the file has been converted via ALAC (and some of the life lost), I would not expect that life to be somehow re-injected with a conversion to AIFF.

So, looks like you will inevitably be destined to re-rip 7000 songs into AIFF then....

Best regards, FT

Why not FT ? ALAC stands for Apple Lossless Audio Codec. All the information from the CD is in the ALAC file (provided it was a bit-perfect rip). If the conversion tool is not buggy, the very same information that was in the ALAC file is in the AIFF file, just uncompressed.

It is not the information that is at stakes here, it's the processing needed to play the file. The musical information remains absolutely the same whenever you convert any lossless file to WAV, FLAC, ALAC or AIFF (or any other lossless format for that matter).

If you find that it's worth the trouble, just do it au fil de l'eau as we say George (or hire a bunch of friends and their PCs to help you over a beer or a six pack) !

ATB

Maurice

Well Peter, it shouldn't if the conversion tool is not buggy. In doubt, ask the usual suspects to teach you how to check the data chunk of both AIFF files.

Thus said, bugs happen.

ATB

Maurice

Dear George

I can only comment from a technical perspective and subjective viewpoint on my own system.

Subjectively I can hear no difference between ALAC and AIFF and it makes no differences which format I rip to or store my files in. However, I avoid USB preferring S/PDIF optical and my music server is a Mac and DAC is Naim ... so it is not really like for like. 

ALAC and AIFF contain exactly the same PCM data - just like a zip file and a document contain the same text. The difference is in processing, it takes longer and more effort to open the compressed (ALAC, Zip) than the uncompressed (AIFF, original document). However, the processing difference is small and possibly offset by the disk access which is less with the smaller file. 

You are using iTunes and send the music data (PCM) to your DAC. iTunes will send exactly the same output to your DAC irrespective of whether it is ALAC or AIFF. Your DAC cannot (doesn't need to) know the original file format. 

You are using USB which has some electrical noise. The async mode will obviate the jitter, but not the RFI/EMC. An optical connection is better in this respect, but not an option for you. So it is possible you may get a little more noise from ALAC, but .... 

You have a PC ... if you bring up the Task Manager (you right click on the strip on the bottom and it is an option) then you can see exactly what you are running - a lot of it will be garbage (stuff you don't need to play music and it will also change depending when you do it). All of this will inject noise and that noise will vary over time. You can see Apps, Processes, Services and Performance. 

Can you hear the noise .... don't know, but try doing something aggressive like running a PC application to see if affects the sound while it is loading ... if it doesn't (and I suspect it won't) then the difference in residual noise between ALAC and AIFF is also very unlikely to do so. 

So I never doubt people can play something twice and hear differences, but my take is it has nothing to do with whether the file is in ALAC or AIFF (this may be different on an integrated player like the Naim NDX or NDS though it should not be in my view). 

In the analogue domain the difference are subjective ... but in the digital domain it is more black and white (1s and 0s), However, interference from the digital domain can affect the analogue (a mobile phone can interfere, yet it is often used as a remote). 

I would stay with ALAC if I were you, as it takes up less space and is quicker to load in to memory than AIFF. You may need a bigger disk if you expand everything to AIFF, which you could do by letting iTunes process all of your files overnight. 

All the best, Guy

Well I have opened a can of worms with this!

It was a discussiion with a leading Forum member a couple of weeks ago, and he said the FLAC was not as good WAV. Then we had a few thoughts on what "might" be the case between WAV and AIFF. AIFF keeps the splendid tagging of ALAC.

So I tried converting a few tracks, and perhaps I was not receptive - too tired, not drunk enough, no sufficiently recently broken legs and loads of morphine - but honestly it made no difference when done as an A/B between the two versions of the track in question.

But in the process I adjusted the default rip in iTunes to AIFF, and later ripped a CD as per the opening post.

After a few passes just adoring the music making, it suddenly seemed to me how bloody good it was actually sounding, So I made one of the AIFF tracks into ALAC, and yes it was back to my normal happy listening, but by now I could find the difference, which I described above. 

I wish it were not so, but it is. The difference between good and poor recordings is much more significant of course, but it must be down to the "uncompressing" of the ALAC that is somehow less good on the fly [i.e. it is affecting the analogue output in some way] than simply allowing the ITunes to convert from ALAC to AIFF, whilst not playing the track, but playing it afterwards without the need to uncompress the new file on the fly. 

There is absolutely no way that I would re-"rip" the whole lot! But conversion to AIFF works for a neat little free gain in quality in my latest experience.

As my hard disc is less than one third full - a dedicated hard disc for the iTunes library alone - then about doubling the actual file sizes will not matter.

It will probably be a gradual process, ... of doing them as I listen!

If anything interesting crops up, I'll let you know!

ATB from George

George, i think you have  now heard what some  of us have also heard, and once you aware of it, it's very hard to ignore. As you say the difference between good and marvellous.

Although I don't think I have ever streamed ALAC to my network player I would expect the differences npbetween AIFF and ALAC to be broadly similar to WAV and FLAC.

Simon

As Simon suggests, once the improvement is found it becomes hard to ignore!

I have been converting a lot of chamber and solo instrumental music from ALAC to AIFF, as it seems to me that this quiet music is where the change is most clear. In music that is most nearly correct in recordings in any case. Segovia playing the guitar is more obviously right or wrong in replay when things improve! An orchestra of 120 playing hammer and tongs is less open to this sort of improvement, because the recordings are far more compromised in the first place.

What is a surprise is that files sizes are actually increasing in the order of three times larger on quiet music. 

It maybe that I have to replace my half TB drive with a one TB unit before the end!

It is a discrete drive dedicated to the iTunes liabrary alone so that is not a problem with the current hardware. I would rather keep at least 30% free space on any hard drive.

As Simon said, this is going from good to great. I cannot disagree with his comment.

Thanks James for your observations when considering much better equipment than I have.

ATB from George

Aiff every time.

Tog

I can't hear any difference between the PCM extracted by my Mac from an ALAC (Apple format) or AIFF (Amiga format) and sent through an optical interface to my Naim DAC. As I understand it the data in the DAC's buffer are identical in both cases and it plays these from that buffer. So it is playing the same thing. if instead of sending it the Naim DAC I send it to my UQ and out again and capture it ... the PCM is the same. 

Using Audacity I see no difference. 

Even if you load the Mac with a program just to divert processing power from rendering music for the sake of it, I can't hear a difference. OK, if you nice the rogue program you can introduce a stutter, but the effect is the same irrespective of file format. 

Also if check the processes running on a Mac then they differ over time as the Mac does its housekeeping, but despite this behaviour I can't hear any difference. 

Of course I prefer AIFF because it was designed for the Amiga and the sentimental attachment remains, but I really cannot hear any difference between WAV, ALAC, FLAC and AIFF ... and when configured for bit perfect playback I cannot hear the difference between iTunes, Decibel or other players. 

I can clearly hear the difference between a Naim DAC and the Chord DAC though. 

However, I still advocate there are other reasons that folk hear differences and that the file format when using a computer to feed a DAC is not the reason.

However even if the differences heard are psychological (and I'm not saying they definitely are) then it still a good idea to choose the format you like best.

I have found that I like AIFF..

I wish I did not, because I am only 10 per cent through making AIFFs out of ALACs, but the consideration is my main and only music system, and no money is at stake - just the patience to make the conversion, 6974 tracks, one at a time ...

It would have been easier if I had not found this.

Even on a modest system, Quad ESL 57s would show up a gnat having a poo on a microphone stand, so that I am well enough placed - quite apart from my hearing - to find this most unfortunate development ...

Getting a better DAC will only magnify the improvement, but will involve less work if more cost ...

ATB from George

George, if you have your file in ALAC, you don't need to re rip to get into AIFF, you can simply use a format converter such as DBPoweramp to do a bulk conversion. There must be OSX equivalents.

As ALAC is lossless, if the software is working correctly no info is lost in this conversion.

Simon

ITunes will do this. No need for any external converters.

James

Thanks James... Yesure opulent remember  whether iTunes could and didn't want to give a bum steer.

Simon

If you set iTunes to rip to AIFF and then select your entire library - just select one track and hit Command-A - then do a Ctrl-Click - a context menu will appear with the option Create an AIFF. Be careful though because you end up with all your files in both formats and then you need to delete the ALAC files. So if you leave the conversion process running over night you need to make sure your disc has sufficient capacity. 

You could create an AppleScript that converted an album did the deletion and then did the next album and so on ... 

Or you could buy a big NAS and store every album in lots of formats .... I'm still mystified how the same data loaded in to the Naim DACs buffer can sound different depending on what format it was stored on disc though.

This is not the same as the NDS/NDX where you are sending it ALAC or WAV and relying on Naim to do the rendering ... assuming Naim rendering is optimised for WAV, differences are understandable.

If you give a DAC just the bits then how does it know what format you stored them in originally. Are we saying there is more jitter produced from ALAC than AIFF .... if so I still can't see how? I can't understand how RFI comes in to play if the Mac, that is isolated from the DAC by some glass, is doing the rendering. 

The CPU usage on a Mac is not constant ... iTunes takes between 6.2 and 6.7% of the CPU to play Arcade Fire's Woodland National Anthem (ALAC) and between 5.8% and 6.3% for AIFF. So yes there is a measurable difference, but I can't figure out how my DAC knows this (or how it can possibly make a difference to SQ). 

As I say I can't hear any difference, but that could just be my age. Is there any logical explanation why others could hear a difference? 

George

Not a criticism but for info: 

Are you running i-tunes on a PC or a MAC? According to your profile looks like you are using a PC.

If it a PC it is also of interest to know what Windows version you run. Windows XP and Windows 7 process audio through different software manipulations neither of which are good for the best rendition of audio files. In the case of XP for example I believe everything is adjusted to 48 KHz even if it starts out as 44.1. There are ways to circumvent this situation but they are op system dependent. 

Just a suggestion but it maybe the audio differences between ALAC and AIFF are made apparent by this.

Regards

Geoff

Dear Guy,

I cannot risk batch converting with iTunes, because it might hit the limit of my dedicated internal hard drive. Only iTunes library on it. It is clear that uncompressing to files to AIFF is in some cases even mutlipying the space taken by four times. None is less than twice the size. While I might get the whole lot on the drive in AIFF, I would definitely run into trouble if the ALACs were not deleted as it went along!

In any case, I'll have to get a large internal HD so that I may eventually put onto it the complete Bach Canatas sometime soon! 200 of them or slightly more! This is the biggest gap in my library of recorded music, and of course I may want two cycles in time, as rarely is one cycle lovable in every single piece!!

I have started manually converting, and it is a good, but before I do too much more, I'll get a one TB USB HD to back up the current part and part position, and also then get a one TB drive inside the little PC. Then I can carry on. Possibly even batch convert the lot once I have adequate larger back-up storage. My current back up is a half TB USB HD, last fully adjusted five months ago. Not many CDs added since then ....

As for the sonic thing, I was convinced that there was nothing in it at first as reported above. The improvement is marginal, but if the only cost is a pair of new one TB hard drives plus my time, well I'll go for it. 

I can understand how some would not find the same results as I did, and at that more or less by accident. Certainly I had not released that I had ripped the Schubert Great C Major Symphony in AIFF, and after three serious listens to a wonderful performance, I suddenly thought that I had never heard such a vintage [.c~1960] recording sound so splendid. Then I tried to find out why ...

So not a question of expecting the result, but really an accident of a harmless sort! 

But it does involve quite a lot of work to actually do it all!

I am absolutely not going to use a separate NAS. With the absolute requirement to set up a network, it is too complicated with too many additional wires and boxes to accommodate  All I have now is an amp, a small DAC, and a PC apart from the old ESL 57s. I am certainly not going to add more to this. I can put all this into some small IKEA style cupboard with an open back for ventilation, but with a door, so I don't have sight of anything but the speakers, and monitor [used for DVDs also], plus mouse and keyboard.

ATB from George

Dear Geoff,

It is only a little PC with XP. Two internal drives - one for the operating platform and the other [half TB] exclusively for the itTunes library folder.

I was possibly wrongly under the impression that it is possible to make it work in 44.1 sampling without conversion with a certain amount of setting up. I did not do this. It was done for me, when I got the little computer about a year ago. It goes very nicely.

Of course a MAC Mini would be very nice, but I have not budgetted for it.

ATB from George

PS: Maybe next year should see a Mini come? I don't know, but it might be the best next move. Even then it is not such a big expense compared to the proprietary solutions that really are far beyond me in price, although guaranteed to work ... 

Dear George

One option with AIFF is mono. I'm not sure what the effect would be, but it is certainly an interesting option where I have some recordings that were enhanced(?) to stereo by some well meaning marketing studio person, as it may send the back to something akin to how they were meant to be heard. iTunes dors not give options with ALAC.

I am surprised that AIFF is more than twice the size of ALAC on my Mac ALAC files are about 60% of the size of AIFF ... however I do embed 300 x 300 (low resolution) artwork in my files and I don't think you bother with artwork if I recall correctly. With a CD the AIFF cannot be much more than 700 MB as AIFF is more or less what is on the CD. 

I do hope at some stage we can forget about formats and just play our music, but as long as we can hear improvements then I think we'll be trying to optimise. Funny thing is with vinyl I just play the music and forget tweaking - wasn't it supposed to be the other way around. 

The Mac Mini may not suit you as well as your current set-up without significant investment as I think it works best with OS X and iTunes are on Solid State Drives and an add-in like Bit Perfect push the music from and external drive in to memory (this makes it work silently from a mechanical viewpoint). You can use a DAS rather than NAS to avoid networking using its Thunderbolt interface, but again the price is not good. 

I obviously don't know much about PCs (yes I have to use one for work, but if it goes wrong I just reboot and if that doesn't work phone support and they remote control it) so I wonder does iTunes get the sample rate right. On my Mac it is not very clever at ensuring the incoming sample rate matches the one it puts out: Apple recommends up-sampling to 24/96, but that doesn't seem right to me. I use bit perfect because it sorts out that problem/feature. I guess you can fix it to 16/44.1, but I do have a few high resolution recordings. There is a paper (please click here) that discuss it. 

Did you switch to a Rega DAC? I know you were thinking about it. 

All the best, Guy

Dear Guy,

My observation of four times the size is always with mono recordings. Perhaps the compression see mono as being dual mono saved once and recreaeted a dual mono in AIFF?

Stereo is about twice as a rule. But about four in ten of my recordings are mono in the first place. I am growing to appreciate stereo as it comes. Because my ESLs are so close together than stereo is never exaggerated with them. It is as narrow as in real life, and I am not disturbed by vague directionality, but only when this is exaggerated to a Technicolour Panavision version, which normally spaced apart stereo speakers can well achieve.

More follows.

ATB from George