MusicTrax

diw commented: To my surprise, I found an alarming number of faulty ITunes rips from CD...around 110 of 1500 albums. How did you rip the files in the first place? If you used iTunes, be warned that QuickTime has a notoriously bad ripping engine, even when "use error correction" is checked. You'd be wise to use Chris' suggestion of an AccurateRip method like dBPowerAmp (for Windows) or XLD (for Mac OS). I suspect the rips were always bad, and you may not have noticed them. I ripped about 500 CDs many years ago and subsequently discovered about 5% of them were bad, usually the last couple of tracks. With dBPowerAmp, I had zero-point-zero problems.

Chris commented: <em>"Based on Apple's history I'd say audiophiles shouldn't expect much from iCloud. I'd love to be wrong and announce to the world iCloud supports lossless file formats."</em><br /> <br /> Well, there's always the hope they would support Apple Lossless. Even if it wasn't, I wouldn't consider Cloud storage to be my <em>only</em> option for music listening -- it'd just be another way of listening, after my existing servers and iPods.<br /> <br /> The key to me will be how much storage they make available, and at what price. $1000 a year for 1 Terabyte doesn't appeal to me. If they could do it for a hundred bucks a year, that'd be fine.<br /> <br /> Bear in mind that users are still liable for data charges through their existing ISPs. If you're streaming lossless music 24/7, you'd probably run smack into the ISP's data cap limitations.<br /> <br /> The other question is: how will they prevent users from giving access to friends and relatives? In other words, what if I give you the password to my account, allowing you to stream a terabyte of music from my collection? I can see where the record companies would be nervous about this.

ALAC is variable because it's inherently a variable-bitrate system. It automatically adjusts based on the complexity of the file. As long as what comes out is still bit-accurate with the original WAV file, it doesn't matter. <br /> <br /> By the same token: if you take a print document or photograph, and compress it with lossless Zip or Stuffit, the file sizes will be different. But uncompressed, they come out the exact same way. I can compress a 600,000 word novel as a Word file, then decompress it, and not a single punctuation mark or typeface has changed.<br /> <br /> You can always take the Apple Lossless file, decompress it back to WAV, then pull it into an audio editing program and see if it precisely nulls out with the original file when you invert the phase. If everything cancels out, then for my money, it's identical. <br /> <br /> I don't think there are any mysteries with lossless encoding or decoding. The only real variables are a) how long does it take to compress the file, b) how large all the files, and c) what equipment can play back the file. From a sound quality point of view, I believe all lossless formats are the same.

CSD455 commented: <em>My freshly ripped ALAC files have varying bit rates (typically 800-1100 kbps) listed in iTunes and are slightly different by song. Is that normal?</em><br /> Yes, it's normal. The musical content, EQ, and level of the songs are different, so naturally in a variable-bitrate situation, the bit rate will be different. I've seen old mono WAV files get compressed down to below 700kbps, but they're still exact representations of the original WAVs.<br /> <br /> You can convert the ALACs back to AIFF or FLACs, but they shouldn't sound any better or different than the ALACs. As they often say on the Hydrogen Audio Forums, "lossless <strong> is</strong> lossless." All the various lossless formats sound identical, assuming a perfect bit-accurate rip.

<em>Kirkmc asked: "Is there any reason to rip mono discs (in this case, the Beatles set) in stereo?"</em><br /> <br /> No. All this does is double the storage space. A mono FLAC or mono WAV file will be half the size, but in playback, you'll hear the same channel reproduced equally through two speakers.<br /> <br /> Apple does not implement mono Apple Lossless through iTunes, but you can force true mono encoding through dBPowerAmp. I think their encoding is smart enough to adjust the storage so that the mono encoding file size is almost (but not quite) half the size of a stereo file.<br /> <br /> <br /> <em>Humzet asked: "I assume then that this is partly why you recommend copying to FLAC WAVE and AIFF thereby elimination any conversion problems between FLAC and WAVE?"</em><br /> <br /> There should be zero difference between a lossless FLAC, Apple Lossless, AIFF, or WAV created from the same source. The guys on the Hydrogen Audio Forum have endlessly debated and tested this, and as far as I know, nobody has ever heard or measured a real difference. This is assuming a recent-vintage computer, modern codecs, and good-quality rips. As the saying goes: "Lossless is Lossless." <br /> <br /> The only advantage one lossless format has over another is playability on different operating systems and portable players. To me, Apple Lossless makes the most sense, but I don't pretend it's right for anybody. If Apple went out of business tomorrow (unlikely, since they have about $50B in the bank), I'd just batch-convert all my files to a different lossless format with dBPowerAmp. As long as all the embedded metadata is retained, and the conversion is perfect, there's no issue.

I agree completely with Chris. Sometimes, it's all about the content, and quality is secondary. <br /> <br /> J. Gordon Holt of <i>Stereophile</i> often used to comment that there was a weird inverse rule that applied to "audiophile" releases: the better the sound quality, the worse the performance. And sometimes, just pulling out an old vinyl record gave you the most emotionally-satisfying playback, cracks, pops & all.<br /> <br /> I'll say this, though: I can still hear bad compression artifacts on cheap systems, even boomboxes. Once you know what to listen for, it's grating on everything. At 256K and above, I can't hear it enough to bother me most of the time. But if I know it's lossless, I can psychologically relax more -- whether because of the lack of artifacts or not.

Agreed. I'm more of a pop/rock guy, and there's no question that a large classical collection would be much more of an issue for metadata tagging. In a case like Encore's above, I can see a use for a system like GD3.<br /> <br /> Still, my central point is the same. For an average CD collection, spending $200 on three standalone rippers, dBPowerAmp, and doing a lot of manual intervention is much more cost-effective and can yield perfect results. You just do a lot of the work yourself. I also have a preference for AAC and Apple Lossless, and omitting those from any ripping program is a deal killer for me.<br /> <br /> I suspect, though, that somebody with enough money to buy a full robotic system would just go the next step and just hire a ripping service to do all their CDs. There are many companies out there that can turn around a collection of thousands of CDs for maybe 50-90 cents a disc (plus postage), like GD3's own GetDigitalinc.com service, and it's no muss, no fuss. I suspect this will be the main buyers for systems like this: ripping services, custom installers, and high-end audio stores who want to help their customers adapt to servers. If you're putting in a $250,000 home theater system, spending another $5K to have somebody rip your collection and organize it on a Sonos system (or something equivalent) is chickenfeed.

maplejohn asked: "If I rip my collection to WAV using the GD3 system, am I eventually going to have my all my songs sitting in an "unknown artist" or "unknown album" folder?"<br /> <br /> I think this is a very bad idea, for reasons Chris has explained before. To me, in 2010, there's zero advantage of WAV files over lossless files. I can't hear or measure a difference between a WAV file and an Apple Lossless file, after spending two solid days testing, using everything from test signals to wide-range music. I think the key is: fast computers, fast servers, and great D/A converters, plus ripping the CDs with no compromises (good DVD-ROM hardware and good software). That, plus the willingness to groom the data manually -- which can be done quickly once you get the hang of it.<br /> <br /> The huge advantages of built-in metadata with Apple Lossless (or FLAC), along with embedded image files, make it 100% preferable over WAVs. I can speak with some expertise, because I spent a year or two in the early 2000s ripping a lot of CDs to WAVs. Luckily, I only got a few hundred CDs into the project before realizing the problem. One can make an argument pro or con different lossless formats, but -- as the extremists frequently argue over on the Hydrogen Audio Forum -- "lossless is lossless." The only advantage of one lossless format over another is playability on different systems.<br /> <br /> Apple Lossless worked best for me because it had to work on Macs, PCs, and iPods. That was a much better solution than WAVs. Granted, there's always the risk Apple will go under (though with a stock price of $210+ and a net worth of $40B, I think they'll survive). But I have enough open-source tools that I can always batch convert the collection to FLAC or anything else.<br /> <br /> BTW, there is the Broadcast WAV format, which does support many dozens of headers, but it's not widely used in consumer players. I use it all the time in my work because it can store scene & take info for film production, along with embedded timecode, but it's pretty much a pro-only format.

Doug commented: "Any other tagging software will list everything just as it is on the hard drive, then the reconciliation process is impossible or extremely time consuming."<br /> <br /> Naaa, it's doable. A program like dBPowerAmp can be easily configured to just automatically put every album (CD) in its own folder. Load the newest-ripped folders into an MP3 Tag-reading program (like iTunes or something more sophisticated), using the "created by" dates, and just check them by eye. It takes less than a minute per CD to verify that everything is OK.<br /> <br /> Bear in mind that as two CDs are ripping, I'm checking the tags on a third right before it's ripped. The only thing that slows me down in manual mode is if lots of changes have to be made, which happens on rare occasions. Having access to four different sources of CD metadata (which dBPowerAmp provides) helps quite a bit.<br /> <br /> The reality for me is that three standalone Firewire CD rippers and one copy of dBPowerAmp cost less than a couple of hundred bucks. Any other solution -- particularly one using a robotic ripper -- doesn't make economic sense for me. Especially with a 5 cents charge per disc for metadata. And I'm confident that the quality of my Plextor drives and the AccurateRip software is giving me zero compromises in terms of the data. I can also get more done simply by running multiple drives at one time. One single-drive robotic system can only do -- at best -- a dozen CDs an hour. I can do at least double that using three drives.<br /> <br /> I readily admit that for people who don't have budget concerns, your solution (or any of the commercial ripping companies) work fine. I also know database management fairly well, and I've often worked with keeping track of vast amounts -- petabytes -- of film data files in post-production here in LA. My methods only work if you have the time, the expertise, and the equipment to do it right. But I can do this in a fraction of the cost for any standalone system.<br /> <br /> BTW: every major motion picture ever made, even those shot digitally (like the recent "Avatar," use human beings to input all the data. It's a necessary part of the system. Once you get good at it, you can manipulate, move, rename, and adjust the metadata very quickly. Granted, it helps if you're an inveterate nerd, but that goes with the territory.

After trying various Windows programs on the Mac, I finally decided that neither Boot Camp nor Parallels did what I needed to do. Ultimately, nothing was better than just buying a cheap (under $300) PC and just using that with the same monitor I use for the Mac.<br /> <br /> To use the Windows side, I just switch my monitor over to the PC, dismount the drive from the Mac, mount the drive on the PC, and fire up whatever Windows programs I need to process the files.<br /> <br /> I use dBPowerAmp for ripping, and Tag & Rename (from Softpointer.com) to handle all the metadata. Tag & Rename has 1000 more features than iTunes, and will work with Apple Lossless, FLAC, MP3, AAC, whatever you have. <br /> <br /> I do run MediaFour's MacDrive in order to read Mac drives on the PC. I've had no problem formatting even 2TB drives as FAT32, and they read and right fine. But I take the precaution of making at least two backup drives in the vent of failure.

Sure. dBPowerAmp's batch-ripping solution is described here:<br /> <br /> http://dbpoweramp.com/batch-ripper.htm<br /> <br /> Again, I'm not convinced that unattended robotic CD ripping will work, because of the increased possibility of bad rips, wrong CD identification, bad metadata, and computer hiccups. I think a human being has to keep an eye on things in order for it to work.<br /> <br /> When XM Radio ripped about 25,000 CDs for their library about ten years ago, they used humans on shifts, working 24/7. I think that's extreme for most people, but I think it's reasonable for one person to use three rippers at once (and three simultaneous copies of the program on one computer) and check everything manually. This will be much faster than a robotic ripper, but the trade-off is the manual labor. Still, I've been able to do at least 25 CDs an hour this way, more if I fire up multiple computers.

Note that the dBPowerAmp people now have low-cost robot CD changers and an automated version of their ripping software available. This will rip to WAVs, FLACs, MP3s, AACs, Apple Lossless, you name it -- and can be configured to batch converts two sets of files, like one lossless and the other as MP3s. I think this is a far more economical solution.<br /> <br /> My own experience is that you can't rely on any of these programs ripping CDs accurately because of the flaws in the metadata. You still need to do a lot of manual intervention in order to "groom" the data (as some of the ripping services call it). I would rather do this than pay 5 cents a CD for the information.

Beautiful job, Chris. This is very close to the article I was planning to eventually write myself! I agree with a lot of your concepts and the approach. My experience is also that dBPowerAmp is the best, most flexible ripping engine out there, and I prefer using it to anything else -- despite the fact that my main servers are all Mac-based. <br /> <br /> My preference is for Apple Lossless, which to me, makes more sense than FLAC, simply because it works with iPods and any computer that can run QuickTime. I might change my opinion if iTunes supported FLAC, but Apple is very much a "not invented here" company. I love the Mac, but I often disagree with how Apple runs their company.

The drawback to both the servers Chris has outlined ("CA01" and "CA02") is that they put the computer in the listening room. To me, the amount of noise from fans and hard drive motors would interfere with the music playing from the speakers. <br /> <br /> I think a better choice would be to just have a touchscreen monitor alone in the audio system, then put the computer and drives in an adjoining room (or a well-ventilated closet). I think this is a far better solution, and heavy-duty shielded cables for the monitor and digital audio connections are not that expensive, nor are they complicated.<br /> <br /> I do like the idea of a touchscreen interface for a Mac-based audio server, and ultimately I think that's what I'm going to go with.