New vs. old DACs - opinions?


I'm on the market for a new DAC. I've noticed that you can find used DACs from, say, 8 years ago that are heavily marked down from their original price. I just saw one sell for $400 that was originally $1500, for example.

So, correct me if I'm wrong, but the progression of DACs seems very different from that of amps... an old amp, like McIntosh, is still highly competitive today... but it seems that newer DACs are more evolved, refined, and use higher quality parts for less money, right?

Another thought is - before DACs were as widely used as they are today, perhaps the mark-up was much greater in the past...? Where-as now, with the influx of foreign manufactured DACs, there is a healthy bit of competition that keeps prices down by limiting the manufacturer mark-up. Correct me if I'm off here as well.

So, overall I'm wondering if I would be better off buying something new like a Keces or MHDT DAC or finding something older that is heavily marked down.
djembeplay
"The specifications for D/A chips have improved over the years". But has the sound improved? Chris Bryant, whom I have already quoted, summarizes it like this.

To my mind there has been a general and continuing downward shift in the sound quality extracted from DACs over the past 20 years. DACs are produced by semiconductor manufactures and require large volumes to make operations economic as most of these will find their way into cheap products, there is little incentive to create good sounding components.
The DACs available for the HIFI designer will also be found in computer sound cards, personal digital stereos and televisions. As a result many modern examples have lost their audiophile verve, as the semiconductor manufactures prefer to make cheaper and lower powered devices better adapted to their principal, high volume customer's needs.
These mass produced low-bit converters do deliver excellent measured performance, but often seem unable to deliver the goods when real high fidelity sound reproduction is required."
This is the other side of the "if it is newer it must be better" argument. Personally, I have not heard enough decoders on my system to have fixed ideas on the subject. I only introduced these ideas to point out that there are differences between informed observers as to the general sound trend in CD. One distressing element is that the big Japanese audio companies who , with Phillips, were the driving force behind CD originally took their best designers off it's further development years ago. I think there are two things we can agree on: One , each DAC has to be evaluated on its own merits and Two, after all this time CD should sound better that it currently does.
Stanwall - you mentioned your DAC playing at 96kHz better than 192kHz. It is because every DAC chip has lower THD at 100kHz than 192kHz update rate. For that very reason Benchmark decided to update only at 110kHz output DAC chip that is capable of 24bit/192kHz.

As for multibit or one bit dilema they have specific sound. Most of converters now are one bit (delta-sigma) and even SACD is a byproduct of delta-sigma just before filtering. Most of people don't complain about sound of SACD or DSD recording maybe because it is not in priciple but in realization. It is very difficult to keep exact timing with high oversampling rates and delta sigma end-up not much better than multibit converter. Multibit is limited by tolerance of component to about 18 bit but there are improvements on the scheme. DCS makes RING-DAC where they shuffle components inside (ring of 5 if remember correctly) to get an average value. ARCAM bought license from them and used RING-DACs in earlier FMJs (FMJ-23 I believe). They had poor production yield and stopped but I believe DCS still makes them.
This is an interesting question that seems a bit too general.

One thing I can say is that in the last 20 years, transports have gotten a lot better. This is one reason I have a difficult time comparing old DACs I have had to the newer ones. I know my transport is tons better than older ones I have had. One does not know what their DAC can do until they hook it up to a killer transport. And 20 years ago, those were way WAY expensive. The cost of great transports has fallen a lot.

I always recommend getting a great transport first (I like a music server with an Empirical Audio Off Ramp Turbo device) using the best ripping and playback software). You hook this up to almost any DAC and you are going to have a great idea how that DAC performs. So many transports back in the day (10 - 20 years ago) were awful... And this did not contribute to the sound of older DACs in the least.

Keith
I didn't mean to insinuate that new DAC's have not improved over older DAC's, Many replies here are simply stating that current DAC's are better than older DAC's, technology has improved, I think there is no argument there. I was under the impression that we were talking costs vs. performance here.

Certainly a $1500 DAC today will sound better than a $1500 DAC from 6 years ago, however, in the marketplace their street prices will be very different. So let's say we compare apples to apples from a street price point of view. Since digital gear seems to depreciate fairly rapidly, let's assume that after 6 years the DAC is now selling for 25% of it's MSRP. So is a $1500 new DAC as good as a 6 year old $6000 DAC? Obviously, as in all things audio, there is no one right answer, only opinions.

A similar comparison might be would you prefer to buy a brand new Ford or a older used Mercedes. Each side will have it's advocates and detractors.

Cheers,
John
"I always recommend getting a great transport first"

- why would you buy transport first not knowing what type of DAC you'll end up with. Fancy transport with upsampling DACs like Benchmark DAC1 or Bel Canto DAC3 is a waste of money since all transports will sound exactly the same.

The only advantage of expensive transport over cheap DVD player is low output jitter. Benchmark DAC1 has jitter bandwidth of 3Hz and at the frequency of interest (kHz) reaches over 100dB rejection. If you take into consideration that jitter caused artifacts were at level of roughly -85dB to start with then it is virtually impossible to hear any difference between transports unless one of them is not bit transparent (DSP processing, digital volume control etc).