A brief history of DAC and ADC chips...

[AnalogSteph]

This started out innocuously enough in a post around these parts, now it has become a fairly comprehensive overview of high-end and midrange converter chips with some performance figures from the '90s into the present day, and even a few links to papers and the like are included along with some "family relations". Oh my.
A Brief History of Audio DAC and ADC Chip Performance
Let me know whether I missed any important parts, got something wrong (almost certainly the case), or what I could improve; maybe some of you will be able to dig up more papers as well. My notes are a bit of an unstructured mess at this point, I may have to tackle that.

This has to be one of the largest entries on this page, which has now become my biggest one (459K at this point) - I may give it a page of its own eventually. Thank goodness I turned on gzip compression for various file types years ago, and HTML happens to compress well.

 

[JSmith]

There were DAC/ADC's well before 1987;

https://en.wikipedia.org/wiki/Analog-to-digital_converter#Wilkinson

The Wilkinson ADC was designed by Denys Wilkinson in 1950. The Wilkinson ADC is based on the comparison of an input voltage with that produced by a charging capacitor. The capacitor is allowed to charge until a comparator determines it matches the input voltage. Then, the capacitor is discharged linearly. The time required to discharge the capacitor is proportional to the amplitude of the input voltage. While the capacitor is discharging, pulses from a high-frequency oscillator clock are counted by a register. The number of clock pulses recorded in the register is also proportional to the input voltage

Especially considering digital audio... clearly there was an ADC/DAC involved;

https://en.wikipedia.org/wiki/Digital_audio#History



JSmith

 

[AnalogSteph]

There were DAC/ADC's well before 1987;

There had to be, considering that the first commercial digital audio recording systems date from the late '70s (1977: Sony PCM-1, 1978: PCM-1600, followed by the well-known PCM-1610 and 1630 systems) and the first CDs were released around 1982. (SHForums thread on old digital recording systems)

But I mean, you have to start somewhere, right? I'm already covering about 3 decades as-is. The further back you go into the pre-Internet age, the harder it is to even find the information and datasheets (and even if so, e.g. the old TDA1541 datasheet doesn't even say anything about dynamic range and THD+N), not to mention it gets more and more nonstandard. It is obvious that tons of research was done on digital audio and converters in the '80s alone, and converter quality improved by leaps and bounds in this decade. It's just hard to quantify without measured data at hand.

 

[solderdude]

Even before 1950 there was encoding/decoding in DSD style. It preceded PCM.
They weren't available in single chips though.

 

[earlevel]

This started out innocuously enough in a post around these parts, now it has become a fairly comprehensive overview of high-end and midrange converter chips with some performance figures from the '90s into the present day, and even a few links to papers and the like are included along with some "family relations". Oh my.
A Brief History of Audio DAC and ADC Chip Performance
Let me know whether I missed any important parts, got something wrong (almost certainly the case), or what I could improve; maybe some of you will be able to dig up more papers as well. My notes are a bit of an unstructured mess at this point, I may have to tackle that.

This has to be one of the largest entries on this page, which has now become my biggest one (459K at this point) - I may give it a page of its own eventually. Thank goodness I turned on gzip compression for various file types years ago, and HTML happens to compress well.

This is really helpful—as someone who doesn't keep close track of these things, but has a reasonable chance of designing something with such chips if I ever have the time.

(I hate to be so picky, but "hex speed" is 6x. You use "octuple" in one place, so "sexdecuple" would follow, but that's "awkward". Maybe just 16x?)

 

[AnalogSteph]

This is really—

You're already speechless, I see?

(I hate to be so picky, but "hex speed" is 6x. You use "octuple" in one place, so "sexdecuple" would follow, but that's "awkward". Maybe just 16x?)

I was vaguely aware that the term didn't seem quite correct, but this is literally what AKM calls it... somewhere. Ah yes, the AK5397 datasheet. They're also calling 8X "Octal Mode", I got the other term from Cirrus Logic, who I think also call single speed "normal speed" or vice versa. (Chaos, as usual. In the words of ABBA, "That's Me". ) Maybe they noticed that it wasn't correct at some point or another designer was responsible, as more modern parts just state the sample rates instead.

 

[Frank Dernie]

This started out innocuously enough in a post around these parts, now it has become a fairly comprehensive overview of high-end and midrange converter chips with some performance figures from the '90s into the present day, and even a few links to papers and the like are included along with some "family relations". Oh my.
A Brief History of Audio DAC and ADC Chip Performance
Let me know whether I missed any important parts, got something wrong (almost certainly the case), or what I could improve; maybe some of you will be able to dig up more papers as well. My notes are a bit of an unstructured mess at this point, I may have to tackle that.

This has to be one of the largest entries on this page, which has now become my biggest one (459K at this point) - I may give it a page of its own eventually. Thank goodness I turned on gzip compression for various file types years ago, and HTML happens to compress well.

Fascinating.
Further evidence that DAC technology was pretty well "solved" decades ago and what has changed has been the price and hence the cost of implementation of an audibly transparent decoder.
Splendid, thanks for compiling.

 

[earlevel]

You're already speechless, I see?

I was vaguely aware that the term didn't seem quite correct, but this is literally what AKM calls it... somewhere. Ah yes, the AK5397 datasheet. They're also calling 8X "Octal Mode", I got the other term from Cirrus Logic, who I think also call single speed "normal speed" or vice versa. (Chaos, as usual. In the words of ABBA, "That's Me". ) Maybe they noticed that it wasn't correct at some point or another designer was responsible, as more modern parts just state the sample rates instead.

Ah, maybe be a translation issue, in any case a programmer is clearly involved ("octal" being synonymous with base-8, "hex" as short for hexadecimal, base-16).

(And yeah, "helpful" disappeared in editing...)

 

[AnalogSteph]

I think I should have most everything from the Gearslutz thread covered now, I'm just not sure how to incorporate codecs, of which the most popular ones should definitely be covered (think AK4620A/B, CS4272, CS4270 etc.). I've mentioned some that had an obvious standalone DAC or ADC counterpart, but some just don't. What do you guys think, should I just add these to the DAC and ADC tables or make a new one entirely, just for codecs? By the time I have both ADC and DAC specs in there a codec table would get awfully wide, or else I would have to go vertical, not ideal.
EDIT: I did the former now, not too bad integrating just a handful of parts. Oops, missed the AK4621, that one still has to go in...

 

[Katji]

So, looking at the DAC specs, I got to think about how I have downgraded my DAC [from ES9018K2M to PCM5242] and my amp [I suppose] from AX-497 to TPA3251 chip amps, yet everything sounds much better.
3 other speakers with everything else the same - including the cheap cables, but now the DAC and the amp have "disappeared" to inside the speakers and there are no cables.
The room is exactly the same, WMP exactly the same, even the USB cable is the same.
The power cables have changed, this is like monoblocs, but plugged in the same powerstrip.
...I'm not sure why it seems funny.