Columns Retired Columns & Blogs |
Loudspeakers Amplification | Digital Sources Analog Sources Featured | Accessories Music |
Columns Retired Columns & Blogs |
Loudspeakers Amplification Digital Sources | Analog Sources Accessories Featured | Music Columns Retired Columns | Show Reports | Features Latest News Community | Resources Subscriptions |
The OPPO BDP-95 has the finest DAC of any player on the market now IMO.
I bought the Ayre C5xe/MP player in 2005 for $6000, and until the BDP-95 I had never heard any player at any price surpass it in sound quality.
The BDP-95, for a mere $1000, is clearly better-sounding; absolutely amazing.
I think it has pretty much made every other CD/SACD player from $800 to $8000 obsolete and irrelevant.
This brings up some things I've been thinking about ever since I started getting into this audio stuff.
Now, this is my first time posting anything here, but I've been reading about and listening to various things for a while now. What I was windering was this: given that you can rip a CD with 100% accuracy, as can be confirmed with dbPowerAmp, or EAC, or whatever other piece of software, what is the point into having a high quality CD player, assuming you have a high quality DAC?
I realize high quality CD players take efforts into making the motor moving the CD as accurate and vibration free as possible (among other things related to reading the CD), and I realize that this can be really expensive. However, isn't the point of all of this to simply read off the digital data as accurately as possible? In this respect, isn't the data going to the DAC inside the CD player the same as the data that was ripped from a CD using a laptop's built in disc reader? Is the purpose then of CD players (now, as in, when computers are in abundance) a matter of convenience of not having to rip all our CDs and/or not having to somehow incorporate digital storage into your system? Because it seems to me that if you can have 100% accurate data off a CD with a realtively cheap CD reader, money could be better spent going into a high quality DAC.
Anyways, I did not mean to derail this conversation, but the topic was about the importance of DACs, and I felt my thoughts were inline with the conversation topic. It is because the DAC seems to me to be the most important part of a CD player that I brought this up.
Whether in a receiver, CD player, or a stand alone DAC, I think there is a tendency sometimes to focus on the digital side of DAC performance. Don't forget the analog portion of digital to analog conversion.
The differences I hear when comparing equipment with DAC circuits may have as much, or more to do with the analog output path that follows after the conversion has taken place. That's one of the reasons I would have a hard time deciding whether a CD player is inherently better than a computer, plus DAC. I think the total implementation has more bearing on what I hear than any individual aspect, like disk reading or choice of DAC chip, etc...
jherrera, I also wonder about the pure reading side of the equation. It does seem to me that that DAC and what comes after it should be essentially the only determinants of the sound from the CD, given that any cd player should be reading the same data with equal accuracy.
I read that some people get better sound with diffferent transports feeding their DAC. That may be true but I'm don't know why that would be the case.
Though I think people tend to associate DACs with digital data, their entire purpose is to reconstruct analog signal. Demondog, could you explain what you mean by the digital side of DAC performance?
How good a DAC sounds is basically a function of how accurately it can reconstruct a signal, and it being an IC that does this, the chip's performance is dependent on analog things, such as its power supply. Also, how the signal is reconstructed (i.e. how well the chip's output can swing from one voltage to another to represent, in analog, the digital data) is determined by the chip's design, and again things like power supply. So yeah, Demondog, I agree that the analog performance is the most important. What I was thinking, however, is that this analog section comes in what we think of as the DAC part (with the DAC part being: Digital Input --> DAC --> Analog Output --> Filters). So again, I don't see why spending money on the making-the-digital-data part of things is a good use of money if we can get the same exact digital data with something like a laptop's built in optical reader.
And yeah, BillB, I also don't see why using a excellent performing digital cable would help things out. I can imagine a cable being subject to things like EM noise, impurities in the cable, etc.. However, a cable would have to be extremely terrible to degrade signals enough to confuse the bits going through it. The only thing I could see affecting sound would be jitter, but since most DACs re-clock the signal, I don't see why having a high quality digital cable would help.
Anyways, if anyone could point out anything I might be missing in my thinking, it be cool to learn some new ways of thinking about this stuff.
Although the data are reclocked in the DAC, any jitter introduced by the cable and/or receiver circuitry can only be low-pass filtered, not eliminated.
John Atkinson
Editor, Stereophile
Thanks for this comment. It made me look more into how DACs take in the digital signal. It also made me start looking into the data sheet for the chip in my Musical Fidelity M1, as well as some papers ragarding jitter. It does take some thinking to understand how jitter can go from cable to receiver circuitry to other parts of the DAC, and take into account the possible sources of timing innaccuracies along the way. So again, thanks for the thought inspiring comment.
You're welcome. Every D/A processor and CD player has a memory buffer that is supposed to isolate the DAC conversion timing from timing fluctuations in the incoming datastream. However, unless that buffer is large enough to hold the entire track (which means a large time delay between pressing Pplay and hearing music), there has to be some kind of link between the DAC clock and the incoming clock so that the buffer neither over- or underflows. And that link provides a transmission pathway for jitter.
John Atkinson
Editor, Stereophile
There is clear, simply understandable mathematics and technologies based from work done in TDM in the very early days of multiplexing, from good old Ma Bell, explaining how to eliminate problems with clock jitter in situations such as you describe.
I won't go as far as saying people in general have learned these lessons, sadly, but it's not a hard problem to handle in, of all things, an analog phase locked loop setup with proper loop design and switchable gain.
I've been thinking about John Atkinson's comment for a while now, and I understand how a buffer can allow jitter to go through it. However, I think this is only possible if the rate at which data is released from the buffer is roughly the same as the rate at which data is arriving at the buffer. The situation I am thinking is the following: data arrives at the buffer and gets released in such a way that the buffer doesn't hold the data for very long. If data begins to fill the buffer, then any data being relased will be relased at a rate governed by the new clock which governs the buffer, and any errors in timing prior to this are pretty much erased, since the data is just waiting to be released at the buffer. However, if the buffer releases data quickly enough that it is always almost empty (as in, releases data as soon as it gets it, and there is a small period in which it is empty), then I can imagine if there are timing errors can allow a buffer to spend some time waiting for new data to release. These errors, if they occur at all, are bound to be small, but this is the only way I can think of jitter passing through this buffer. I don't know enough about rates at which the buffers or the digital signal is transmitted, but since usually these DACs can operate asynchronously, I imagine these problems don't tend to happen. Again, if anyone can point out where in the above I'm missing something, let me know.
Another thing, I don't know if there can be any thing in the audio realm like wander? Usually, this means jitter below 10Hz, but I think 10Hz in terms of audio is pretty huge. In the above, I was thinking if there was something like wander, then this would also cause problems for this buffer, but again, I'm not sure this can even happen.
j_j, I found your comment interesting, especially the part about people having learned lessons. Since all of this is from an engineering standpoint, I think that usually problems like this are always tackled in an engineering type of way. By this I mean that there are pretty much standard solutions to a large amount of problems, like the PLLs you mentioned, but its not so much of "lessons learned" as it is that engineering solutions build upon each other, so you need not look too far back to find something designed by someone that improves upon the lessons they learned (and in this way incorporating a great deal of lessons learned).
Also, about the phase locked loops, they tend to have a lot of components that themselves depend on clocks. The signal path for these loops also includes filters, which, since were talking about jitter, complicates things, due to the group delay, which is frequency dependent. In the overall view, I'm not sure if these things have any audible effect on sound, but they do technically influence the timing of the signal. Anyways, I don't mean to turn this into a discussion on how to best conquer jitter, but I just don't think the problems are dissapeared as quickly as droping a PLL into the overal signal path.