Digital Playback Comparison

Which sounds better: Original CD, CD burned from harddrive, playback from computer HD through USB, or optic cable, or Airport Express? I cleaned my contacts, installed NOS Amperex tubes in my preamp, had new tubes installed in my Trivista SACD/CD player, and had a go at a comparison test for these playback methods. The HD used was from my MacBook Pro. The piece of music used for a comparison was the Hallelujah chorus from Handel's Messiah, Solti and the Chicago Symphony Orchestra and Chorus (Decca 1985). See my system for equipment used. The USB playback was via a HagUSB via S/PDIF into the Trivista coax input. Here are the conclusions, going from best to least good sound quality:

1. (Tie): HagUSB and Burned CD (CD was burned from HD at 4x, from Apple Lossless file). I couldn't tell a difference between these two after extended listening including other music besides the test music. Great dynamic range, wide and deep soundstage, a "rightness" to the sound (at least within the limitations of digital) that made you want to quit analyzing and just listen.
2. Original CD. I read somewhere recently (some of you might know the source) that there is an explaination for why burned CD's sound better than original, but until I made the comparison, I was skeptical. The original had a slight loss of "presence", and seemed slightly compressed compared to the #1 playback methods.
3. HD through toslink. 1 & 2 are closer in sound than 2 & 3. There is a noticable shrinking of the soundstage and air with the optic cable. The cable is a good one--Van den Hull glass cable with mini-plug out from Mac. I was disappointed in the sound from this method.
4. Airport Express and optic. A significant drop in sound quality here. If you have been using this method I suggest doing what I will now do which is run USB out of computer to HagUSB ($119) via long run of coax (Canare for me) to your DAC. The Airport Express is OK for "working around the house on the weekend" listening, but is a serious compromise from what you can get with only a minimal incremental investment.

Notably absent from this comparison are music servers, or a good quality USB DAC, or good reclocker/converters that could be used with a computer HD and conventional DAC. Hopefully someone else can do comparisons of the burned CD with some of these methods to see what is the best sound for, let's say, a $2000 or less investment (obviously not including the cost of a computer). That is a level that many of us might be willing to make if there is a significant improvement over the #1 methods above.
i don't see how a burned CD could sound better than the original, unless there were errors that showed up during normal playback which the burned copy was able to correct through multiple reads. but as for a CDR sounding better than the orignal, that seems unlikely. i have done A/Bs between the two and have never noticed a difference.
I was skeptical too, but apparently it is now "conventional wisdom." See the first paragraph of Robert Hartley's article on p. 60 of the December 2007 issue of The Absolute Sound. Elsewhere I read that Peter McGrath was astounded to hear a CD-R of one of his CD's sound better than the original (someone may be able to say where this was in TAS or Stereophile recently). There was a detailed technical explanation of exactly how this is possible. Interesting.
Wouldn't this depend on the quality of the CD burner? I read the article by Robert Harley mentioned above but didn't see the type of burner used. Did I miss that?

11-30-07: Vcoheda
i don't see how a burned CD could sound better than the original...
Back when Genesis made the Digital Time Lens, they used to make CDR copies through the time lens, and use the CDRs to demo their speakers. They explained that the recorded disks sounded better than the source CDs because the Digital Time Lens removed a lot of jitter from the source disk before burning the CDR.

These days we often burn CDRs on a computer, which--like the Digital Time Lens--buffers the data stream and burns the CDR with a rather short signal path. Also, it's my understanding that burning a CDR over a USB link could perform the same function because it is bi-directional and solidly locks in the word clock.
CD's can't have jitter....there is no timing information on a CD...just data order to get jitter you need some kind of clock or a clock signal.

The is No clock on a CD...

....these audio urban legends go against all logic. The jitter must come downstream of the CD.

So what explains the source of urban legends like thiis? Read on if you care...

Imagine a wobbly surface CD or one where the tracks wobble a lot - this is clealry more difficult to read for a lazer/mechanical system, as the CD rotates the mechanism performs more cyclical corrections per revolution of the disc (in fact the focus corrections will be periodic as each woble repeats on each revolution).

Now the lazer will require more focus adjustments per block of data than a well constructed CD (perfect center and perfect flat surface) if you postulate that the rapid oscillations of the focus system for the lazer and the transport mechanism require periodic power and this causes oscillations in the power rails that in turn affect the power to the clock in the D to A......then this may induce jittter in the clock signal.....worse if the jitter is periodic (due to periodicity of the power rail fluctuations) then it will have a much higher chance to be audible than if it where random.

However the problem is the CLOCK timing accuracy! The result may be that one CD perceptively has more jitter than another....but the CD isn't actually the cause. It is the BAD system design that does not properly isolate the lazer tracking mechanism from the clock used to time the D to A.

I hope this helps to explaiin how a CD might "appear" to have more jitter.....what it means is that it may or may not have more jitter depending on the robustness of the CDP player!

Which CD will sound better in a badly designed CDP player? CDR-R? Commercial stamped CD's? The reality is that it will vary....a kind of crap shoot. Genesis found something that worked for them but it may not be the same answer with a different batch of CD-Rs....even from the same manufacturer.

What can you do about it other than playing wiith marker pens and various CD-R manufacturers in an unscientific fashion? ( A hit and miss approach)

A more robust transport, copying either CD to a hard drive and using a computer to play the file back, or a system with an external clock, such as EMM labs use, will generally guarantee that they sound the same. YMMV
Apparently, the CD burners burn the CDR disc more effectively (i.e., more easily read by the CDP) than the CD original was burned, and that allegedly accounts for the better sound. I'm not convinced. My feeling is that, like with everything Hi-End, each burner will have its own sound and sonic character, a frequency response that works very well in some systems, not so well in others, and that these differences will be the bases for the claims of superiority. I agree with those who say that a well made player with a well designed and built transport, laser, power supply, etc., are what we need to focus on for better sound. But if a burned disc sounds better to your ears in your system, and you've got the cash to invest in a full-blown audiophile-approved CDR burning system, then I say: go for it!
again, we are faced with the eternal question: what does better mean ? if it means i prefer a to b, then a is better, there is no definitive answer as to whether an original sounds better than a copy.

if better has a specific sonic denotation, then i would say the answer is that it is cd dependent.
Bruce_1, thanks for your thorough write-up of a comparison that is of great interest to me. I have a HagUSB unit on its way to me and I also have a MacBook Pro and an Airport Express. I will attempt to replicate your experiment with two additions: going through a Squeezebox3 and burning a CDR from an Alesis Masterlink 9600, which is a hard-drive CD burner, among other things.

For what it's worth, I have always been disappointed with CDs burned off my MacBook (ripping via Apple Lossless), which is why I bought the Alesis.
Drubin, I'll be very interested in the results of your test, particularly in adding the SB3 and Alexis into the mix. I've always burned CD's from my iMac and not my MacBook Pro, but just because I have my complete iTunes library on the LaCie HD hooked up to it and not the laptop. I might try comparing CD's burned on the two computers to see if I can tell a difference. According to some of the posts above there might be a difference in sound quality based on burner. We'll all be interested in your conclusions about the Alexis.
I ended up returning the Alesis, deciding it was not a product I needed in my rack. But I will report a huge difference between (1) Mac wirelessly -> SB3 -> DAC and (2) Mac via USB -> HagUSB -> DAC. The first scenario is like FM radio and the second like CD (although I have yet to compare to actual CD playback). In other words, the USB approach is a contender for serious listening, while the SB one is not. I was shocked at how much better the USB is.
Just a note - some early Decca CD's made in the UK have CD rot/CD bronzing issues. This might explain something you observed.

I have a Rossini Decca that bronzed so badly it no longer plays at all. I have several others that are still playable but may be degraded (nearly impossible to tell) I also have several PDO manufactured discs that have an increased noise floor and these are definitely degraded and sound bad.
Shadorne...The "Clock" is sometimes embedded in the data, so that no separate clock signal is needed. Check out "Manchester 2" code, one of the first.
BTW - that you had issues with airport express is not surprising if you think about it ... a tiny little box with SMPS power supply sandwiched close to the rest of the circuitry. However it is the analog that stereophile took exception to and not the optical out.

Perhaps the PLL design in your Tivista did not like what Airport express was doing jitter wise?

How close was the Airport express to the other I mentioned the packaging of this thing means it is more likely to raidate EM/RF noise....

Thanks for sharing - interesting
The CD itself of the 1985 performance is new, but I've gotten the same results from other CD's tested this way, so I'm thinking the difference is in the playback medium, not the particular CD. But, I'm glad to learn about the bronzing problem--something I did not know about.
The AE definitely has limitations, but they are much less when you use the digital out vs analog out. When I sent the TriVista back for retubing and to have the power supply rebuilt I had to go from the AE directly to the preamp to play iTunes playlists. There was a HUGE drop in quality going this route (similar to Drubin's "FM radio" results reported above). I thought that if this was the best my system were capable of reproducing I'd end up never listening to music. I think the lesson from my experiments and that of others such as Drubin, is that we need to do some direct comparisons of digital playback metods and equipment to learn how to get the best quality sound when using a computer as a source. The recent report in TAS is a start, but much more needs to be done to see if there are ways of getting great sound without having to make a five figure incremental investment.
My own system consists of an Apple G5 with CDs burned using lossless into the Slimdevices Transporter via an ATT Gateway Modem to my Audio Aero Capitole using the AES/EBU digital connection.

This sounds really good and much better than streaming audio. However most of my original CDs sound better on the Capitole which means that I have degraded the signal along the way or, I suspect, the Audio Aero was designed to do it's best with an original CD. I still prefer using the Audio Aero DAC over the really audiophile quality DAC in the Transporter as the Audio Aero sounds 'warmer' while the Transporter is much more analytic.

Either way I like the system and I use the Computer generated music set an random song mix.
I have been listening to the following digital playback chain for a while now and can report on the quality compared to the other playback methods described in my original post: iMac-->USB cable-->Hagman USB-->9m of Canare coaxial cable-->Trivista as DAC. Cost: $119 for Hagman, $102 for cable. So, for a little over $200 I can use my computer in another room as the digital source for playing iTunes playlists. The sound quality? Excellent, the equivalent of a CD-R burned from the hard disk, and therefore tied for the best digital sound of all the options described in the original post. This is better in every way than using the Airport Express and Toslink output, and it isn't a subtle difference, either. You can obviously do much better with a high end transport and DAC, or one of the high end servers tested by TAS, but you'll spend a lot more than $200!! I guess the point here is that it wouldn't cost you much to try something like this setup, if you already have a good DAC, before you drop five figures on a transport/DAC or music server. You might end up putting that money somewhere else in the system where you can't get such good results from a $200 expenditure. YMMV.
here is an obvious question that nobody answers:

what does better and best mean ?

it seems that these are subjective terms, such that one man's trash is another man's treasure.

if one is trying to make a point that "good" sound can be achieved without spending $10,000, i suppose that one can agree with such a statement.

what is the point ? component comparisons have an uncertain outcome. opinions abound and this thread confirms that fact.
Vcoheda - simple, the difference in the burned CD is that the pits are more accurately place, which results in more accurate timing, which reduces jitter over the orginal CD. Digital playback comprises the data, which is generally error-free and the timing.

This is really old news. Read the forums over the past 3-4 years. There are products such as Reality Check that rewrite CD just for this reason.

Jitter is what has always been wrong with digital playback.

Steve N.
Empirical Audio
Johnnyb53 - your understanding of USB is wrong. Except for asynchronous interfaces, most USB intefaces are synchronous, and depend on the rate of the datastream in real-time. The thing that makes some USB converters better than others is the chip that is used for the conversion, which can lower jitter significantly, the clock in the converter, and the implementation, or design.

Steve N.
Empirical Audio

admittedly, i haven't read much on the topic, and i generally only listen to original CDs, so the issue has not really come up for me. however, as mentioned in my own experience, the few times i have compared a wav copy with the original, i did not detect any differences and i was looking for them, as that was the purpose of the comparison. so i would say thatg if there is reduced jitter on a CDR version, it appears that the improvement may not be audible (at least some of the time).


I’m not as technically oriented as a good amount of the posters here… I only know what I’ve done, and what I can audibly feret out over a fairly lengthy time of ripping and burning comps and dupes.

I have no decent path just yet to compare HD tracks via USB just yet… I will have an outboard dAc soon however with which to compare non CD/CDR music, shortly.

That being said, I've personally found that there are audible differences in copied & burned CDR vs. Store Bought (origrecorded) CDs. Many times the orig, store bought is better…. But not always.

Sometimes the noticeable diffs one expects to hear in a duplicated disc aren’t there at all. Occasionally, the diffs are so subtle as to passs without detection for a good while… so remote a quirk or slight is the change in the dupe.

I’ve found some conditions play a fair enough part in the reproduction of CDRs too, such as one’s choice of blanks, ripping speed of the orig disc onto the HD, burning speed, and the burner as well. In fact after adding my PC to an inexpensive PS Audio power filter another level of improvement occurred. Using some isolation devices under the tower helped too. Cooling the CPU also is a real gainer.

BTW Physically heavier orig CDs do better overall during the whole of the process. IMO.

All of these changes played either marginal, or just couldn’t help but hear it changes. A nagging notion for me is I’ve always used Sony CD rom drives, or DVD rom drives to rip and burn with. One of my PCs has a different type burner and I’ve done some comparisons between the performance of both in various fashions, rip with one, burn with the other… same same… different CDRs, etc.

It didn’t take much to discern which was the better unit. All else being the same during the process, the Sony DRU 800 won hands down over the Optiarc 3750A DVD burner. The Sony DVD 510A was I believe better still…. But I did that one in finally some time back.

Using the Apple Lossless codec, using error correction, ripping & burning at 4X, and simply sticking to the blanks acceptable to the burner in question have served me very well.

Much of the audible diffs or losses, occur in the lower regions delineation or articulation. When noticeable losses occur it usually is in that area. It does seem to depend upon the bass info being burned. Very busy or synthesized tracks which contain subtle shifts can be blurred now and then during copying.

Trust me here, I can’t recall how many times I’ve played a disc and wondered when or where I had bought it, only to find out I had made it myself!

Also, dedicating the PC more so to burning, via eliminating background services, processes, and closing out apps normally running in the sys tray helps, as does ensuring the burned tracks are all the same file types. The bit rate of them doesn’t seem as important. Chalk that up to the software most likely. Separating drives onto differing IDE channels was a positive move too, but only just.

When I rip or burn, it’s a dedicated event. I’m doing nothing else with the PC.

Making exact copies using NERO has proven repeatedly the best path… either to HD and then CDR, or marginally better, CD to CDR immediately, using two drives at once.

Mixing up varying file types such as mp3, WAVE, etc, during a compilation, does seem to mess up things now and then at the initial opening of the track. When two differing file types follow one another, for example things have a better chance of going south. I’ve completely quit doing that sort of arrangement altogether now.

Does Oppo make a CD burner? They should if they don’t.

FYI…. Much of these tests are really blind tests…. Though not by choice.