RMAF11: Digital Jitter and Volume Controls, Martin Mallison, CTO, ESS Technology

[@antunkatona]

This almost felt like I was back in college! Great stuff!

[@keithlevkoff6876]

Very good presentation - and quite unbiased... I was a little surprised that he didn't mention the other downside of most analog volume controls - physical potentiometers are notorious for having very poor channel tracking down low (switched attenuators avoid this problem; but most lower-cost DACs with analog controls use a plain old potentiometer). There were two other thing he didn't mention when he talked about "how nice it would be if the DAC could get the source device to use it's clock - instead of the other way around)..... 1) When you have a one piece CD "player" instead of a separate transport, then that is usually exactly what you have - the transport drops the data into a buffer, and it is then clocked out by the clock - and sent to the DAC. As long as that clock is very good, you avoid all that stuff about locking onto the clock in another box with a PLL - or ASRC. 2) The same is true when you have an ASYNCHRONOUS USB input. The receiving device (the DAC) tells the sending device when to send data, and uses a small buffer to make sure it never runs out, and so you get to use the clock in the receiving device to clock the data instead of the clock in the sending device. Again, you have his "ideal situation" and the data clock is as good as the clock in the DAC, so you don't need to fix it in the DAC.

[@craigozland]

I suppose they have to cover all bases having the sample rate converter in there. I really hope that sample rate converters have improved. I used one of the AD1890 AD1891 devices when they first came out and the performance was really dismal. Like rubbing sand paper through your ears. Ended up putting my own design low noise butler oscillator at the DAC chips with some ALS logic family D flip flops re-clocking the critical word latch. Sending that master clock buffered back to the CD player. Took that sub standard sample rate converter out and never used one again.

[@vivz1000]

I'm here because my LG G6 quadDAC which has the ESS Sabre 9218+ sounds better than the onboard audio on my PC as well as my ASUS Xonar DG sound card. Manufacturers should start adding better onboard audio chips on their products.

[@Grassy_Gnoll]

Analog volume control plus remote equals awesome Jedi powers! Going to be a sad day when mine dies. Don't even know if they make them like that anymore.

[@henriksrensen3220]

Resistor ladder volume control are the best.

[@rlwainwright]

Yeah, the Checkered Demon has been my "handle" ever since the CB craze in the US back when Carter set the speed limits to their ridiculous 55 MPH. The Demon is my evil alter-ego and I've had to explain him to several gals throuout my dating life after getting divorced in 2010... -CD-

[@craigozland]

And just to add, before the digital era, in the good old analogue days, before audio IC op-amps were around we had a much better chance of quality audio reproduction as designs were using discrete semiconductors or vacuum tubes or a mix of both. However, now all of these fantastic advances in digital are being crippled and compromised by the ubiquitous use of op-amps to handle the audio signal.

[@dedskin1]

I dont know what is wrong with using OP amps in audio , but of course digital noise is present , as is Amp noise , i didnt hear Sampling Noise , i heard Amp noise , that is less in analog controll , but its not that big of a deal , only in quiet room , not something you find in Citys often

[@craigbuckingham4373]

Good, seems progress is occurring. However, if the chip uses op-amp based voltage out it is compromised in meeting the pro-audio benchmark. Op-amps never can come anywhere near good discrete or vacuum tube design. They suffer from geometry limitations, use high forward gain and feedback and so introduce objectionable distortion that can never be removed and makes using open loop no feedback amplifiers in the following stages a worthless exercise. So if the chip has op-amp output that can not be bypassed it creates a weak link in the device that which will be a limiting factor that cannot be overcome. Also what semiconductor family is used? If it is CMOS how is it implemented to overcome the threshold switching sensitivity to supply voltage and resulting jitter generated over better family types such as ALS or better still ECL? Having a low jitter clock and feeding it into noisy CMOS is a waste of a good clock oscillator and counter productive to the aim of reducing jitter. Also how is the final D/A conversion clocked? How many gates, how much degradation to the source clock purity? Seems the compromise for ESS is placing low power consumption higher on the list and pushing audio quality further down the list. Oh, and if you are thinking of quoting THD+N figures being very low and using that as an argument, please don't, you are not even close to the ball park. Some people will understand what I am talking about.