the more bits the merrier?

The CD format is 16 bits with a 44.1khz sample rate per channel, period. You cannot get more resolution than this from a CD.

In theory, a high quality 16 bit DAC is fully sufficient to extract all information avaialable from the CD. Making a high quality DAC is not an easy task, though. As digital evolved, it turned out that using a DAC capable of decoding a higher than 16 bit resolution signal could result in improved performance on pure 16 bit datastreams due to better DAC linearity and precision required in general for higher resolution DACs.

Also oversampling and dithering operations added extra bits to the 16 bit data stream that could be conveniently fed to a higher resolution DAC. These extra bits don't provide extra information but can allow the DAC to reveal more of the infomation contained in the original 16 bit data stream (the "how" is a long story...).

In a lot of ways bits beyond 16 are "marketing bits". Manufacturers engaged in marketing wars over the number of bits supported by the DAC.

In the beginning all DACs were multibit. For each word (16, 18, 20, 22, 24 etc bit) input they output a corresponding analog voltage (or current in many architectures). It was hard to ensure linearity across all bits, and many such DACS required complex calibration during both manufacture and assembly. This kept their costs relatively high.

As fast logic became cheaper to manufacture and use, single bit DAC architectures became feasible. A single bit DAC effectively has only one voltage (or current) reference instead of 16 or more. The input datastream is oversampled (64x or higher) into a single bit data stream. This in turn becomes a high speed pulse train. There are variations on the theme but basically the width of the pulses in the train corresponds to the levels of the incoming sample words. This pulse train can be filtered to produce an analog signal.

Of course I'm oversimplifying greatly. This process results in lots of switching noise that typically is relocated above the audio band by a digital process called noise shaping. The clock stability becomes much more critical than with multibit converters. Clock stability is one ot the limiting factors on single bit converter reolution.

Single bit converters usually specify equivalent multibit resolution. Many now claim to provide 24 bit resolution. There was another marketing war in the 90's on multibit vs single bit converters. Multibits claimed to do bass better and single bits claimed to do highs better. This has pretty much died out, and now there are excellent examples of both implementations on the market.

DAC performance depends on many factors, including power supply, filter implementation, DAC architecture, and analog output stage. The artistry of design is the balance of these elements.

So you can't conclude that a 1 bit architecture is inferior to a 20 bit architecture or that an 18 bit architecture is inferior to a 24 bit architecture. You have to listen.

The CD has come a long way in 24 years. Consider that when it was first standardized it had to be theoretically capable of providing 10-20khz frequency response with better dynamic range than analog tape, and it had to be implementable with 1982 technology. It's an amazing achievement.

Again this discussion appliies to CD only. To support higher resolution formats (SACD or DVD-A or HDCD) we are dealing with data streams with greater than 16 bit resolution, but that's a topic for another reply.

Ben,

Unfortunately there is no simple answer. A well designed single bit DAC can be low cost and can be exceptionally good when implemented properly.

A single bit DAC is elegant, very linear, low cost but has the disadvantage of high out of band noise that must be heavily filtered. (Out of band noise is not a problem but heavy filtering or cheaply implemented heavy filtering may affect some of what is in the upper range audibly)

For the above reasons, pro audio and high end favours multi-bit DAC's which are more complex, expensive and have more potential for linearity problems (matching between individual DAC elements). These do have the considerable advantage of making an output that is much closer to an analog signal and therefore much easier to filter, however, they place more constraints on the power supply etc.

If you buy reputable gear then I would not get too hung up on the DAC methodology - the manufacturer will have made trade offs between the chip and the architecture of the entire design including power supply etc.

IMHO, both can produce excellent sound reproduction with hundreds of times less distortion than your speakers.

If you want top quality then you probably want to get a component with multibit delta-sigma and dynamic element matching architecture.

If you want the best but are not sure about how much to spend on this particular component. Then consider avoiding a $5000 purchase on pro audio quality CD / DAC player if you have only $1000 speakers....it simpy does not give you the best sound quality return on your investment... upgrade the speakers first. If you already own $20K speakers installed in a specially treated room, like most studios, then definitely consider the highest end DACS.

...my two cents

You must listen for yourself and decide. Give two engineers the exact same parts and you'll get two different sounding devices. Don't get hung up on what's inside the box or what other people think.

That being said, I think nonos is the bees knees. But I don't know if it exists in a one box player.

Thank you all for your VERY informative responses. I am glad you kept it simple, because if it was any more complex, I fear my brain may have exploded. Isn't it interesting how - in some respects at least - electronics mirror the instruments whose music they reproduce? After all, put a guitar in my hands...or say, those of Eric Clapton...and it's very likely the same "parts" will produce an entirely different sound. Cheers, and thanks again, Ben