Burr-Brown vs Cirrus Logic vs Analog Devices: What are the differences between these three DACs

[tonyvdb]

Right now there are three main companies developing DACs for Receivers and such.
Texes interments, Burr-Brown
Cirrus Logic
Analog Devices

With the recent switch from using the Bur-Brown DACs in the lower end Onkyo receivers to the Cirrus Logic DACs I wonder if this really makes an audible difference.

Pioneer and Denon have been using the Analog Devices DACs in some of there receivers
Yamaha and the higher end Onkyo's and Denon are Using Burr-Brown
And Marantz and the newer lower end Onkyo's are using the Cirrus Logic DACs

[nova]

Good question,... to which I have no answer. For some unknown reason I have always assumed the BurrBrown DACs were the best

[drf]

DACs, like op amps, are only as good as the supporting circuitry. To the end user it doesn't really matter which dac is used, so long as the unit as a whole is properly designed. In other words a good chip that is poorly implemented will sound worse than a poor chip properly implemented. I would buy based on overall performance rather than what chips they use.

You are significantly more likely to hear a difference in the program algorithms and fx processing than you are dac performance.

my 2c
Dr F

[tonyvdb]

drf wrote:

DACs, like op amps, are only as good as the supporting circuitry. To the end user it doesn't really matter which dac is used, so long as the unit as a whole is properly designed. In other words a good chip that is poorly implemented will sound worse than a poor chip properly implemented. I would buy based on overall performance rather than what chips they use.

You are significantly more likely to hear a difference in the program algorithms and fx processing than you are dac performance.

Thats a very good point, And one to seriously keep in mind when buying an new Receiver or PrePro.

[Sonnie]

Here are a few specs for various DACs:

Burr-Brown PCM1791 (SNR 113dB / THD 0.001%)

Burr-Brown PCM1796 (SNR 120dB / THD 0.0005%)

Burr-Brown PCM1792 (SNR 129dB / THD 0.0004%)

Cirrus Logic CS4398 (SNR 120dB / THD 0.0007%)

Analog Devices AD1955 (SNR 120dB / THD 0.0006%)

Wolfson Microelectronics WM8741 (SNR 128dB / THD 0.001%)

[eugovector]

In addition to chiming in with "DRF's comments sound spot on to me", I would add...

Well, those SNR and THD numbers might make a difference if you were really cranking the system (like, stadium levels), but I'd be curious to see study where at 85dB in your Home Theater someone was able to hear the difference. Even the worst spec'd of those DACs (SNR 113dB / THD 0.001%) still has great numbers, though as DRF said, it's only part of the equation.

[bobgpsr]

Used to be that Burr Brown and Analog Devices had the best reputations for quality performance and truly meeting their specs. Now it looks like Cirrus Logic and Wolfson have arrived, as far as specs go, for consumer audio playback.

[SteveCallas]

Nowadays there won't be an audible difference between DACs, circuitry, algorithms, or anything else unless the manufacturer has purposefully intended to achieve a non flat FR or the amp section is way underpowered. Look to power, inputs/outputs, and features rather than DACs or claimed sound quality.

[avaserfi]

Looks like it has been covered. No difference unless something is broken or the difference is done on purpose. With the types of specifications DACs have humans would need more than golden ears to hear a difference!

[thxgoon]

Sonnie wrote:

Here are a few specs for various DACs:

Burr-Brown PCM1791 (SNR 113dB / THD 0.001%)

Burr-Brown PCM1796 (SNR 120dB / THD 0.0005%)

Burr-Brown PCM1792 (SNR 129dB / THD 0.0004%)

Cirrus Logic CS4398 (SNR 120dB / THD 0.0007%)

Analog Devices AD1955 (SNR 120dB / THD 0.0006%)

Wolfson Microelectronics WM8741 (SNR 128dB / THD 0.001%)

Agreed with all said above. According to these specs we're talking about a ten-thousandth of a percent or .000000# difference between DACs. Hardly measurable let alone audible.

[smorse]

i think a more relevant specification, though hardly available in a standardized measurement, would be the jitter rejection of the various DAC implementations. S/N for these probably already exceed the analog noise level of the surrounding circuits, leaving jitter as the more likely differentiator of sound quality.

[lcaillo]

Really? Is jitter a significant problem in the DAC itself?

[bobgpsr]

smorse wrote:

i think a more relevant specification, though hardly available in a standardized measurement, would be the jitter rejection of the various DAC implementations. S/N for these probably already exceed the analog noise level of the surrounding circuits, leaving jitter as the more likely differentiator of sound quality.

I would think that the jitter that you talk about is related to the conversion clock edge given to the DAC by circuitry outside of the DAC.

This clock recovery is an important issue for AVR's getting audio from a S/PDIF (incl Toslink) or a HDMI audio input. But outside of the DAC itself, normally a higher quality AVR has special PLL circuity, etc. for digital audio input interfaces to get better conversion clock signal to send to the DAC chip.

[smorse]

I guess I'm not sure exactly what is on-board these VLSI DAC's and what isn't -- i.e., are they monolithic chips which contain on-board crystal oscillators (clocks), or do they require the manufacturer independently implement the clocks -- or can they be made to work either with an on-board clock or an independent clock? My assumption, possibly erroneous, is that the jitter performance would be affected by an onboard clock, by whether there is onboard resampling/upsampling, etc. Does anyone know for sure?

[lcaillo]

My assumption is that most clocks would be very stable and that other components would contribute more to jitter. Without more direct knowledge of how a circuit is designed, however, I have learned from experience that basing an opinion on such assumptions is high unreliable. Are there any measurements that support the notion that jitter is a signifficant problem at all these days?

[bobgpsr]

The big problem with S/PDIF (whether coax or optical Toslink) is that the clock must be recovered from the information signal. Years ago this was not always done well and audible jitter effects could be heard.

Much discussed over the years. Just do a google search on "spdif clock recovery". Here is one article that was found that way: Digital audio transmission, SPDIF, and Jitter



More recent IC's have addressed this issue. For instance the TI DIR9001 96kHz Digital Audio Receiver:

TI wrote:

The industry-leading jitter performance of the DIR9001 makes it an ideal data and clock source for TI's high-performance digital-to-analog converter (DAC) products, such as the PCM1792A, to provide the highest performance DAC output available today.
...the DIR9001 integrates a high-performance phase-lock loop (PLL) to not only improve jitter performance but also remove the need for an external clock source when the sample frequency calculator is not used. Both features reduce the overall system cost by eliminating additional system components that competing digital audio interface receivers typically require.

• One-Chip Digital Audio Interface Receiver (DIR) Including Low-Jitter Clock-Recovery System
  
• Compliant With Digital Audio Interface Standards: IEC60958 (former IEC958), JEITA CPR-1205 (former EIAJ CP-1201, CP-340), AES3, EBU tech3250
  
• Clock Recovery and Data Decode From Biphase Input Signal, Generally Called S/PDIF, EIAJ CP-1201, IEC60958, AES/EBU
  
• Biphase Input Signal Sampling Frequency (fS) Range: 28 kHz to 108 kHz
  
• Low-Jitter Recovered System Clock: 50 ps
  
• Jitter Tolerance Compliant With IEC60958-3

[lcaillo]

So are the one chip solutions more stable and reliable in recovering the clock signals than previous designs, or does integrating the process make it more of an issue?

[bobgpsr]

lcaillo wrote:

So are the one chip solutions more stable and reliable in recovering the clock signals than previous designs, or does integrating the process make it more of an issue?

I would think that S/PDIF reception IC's now doing better clock recovery internally, just lets the mid to high end AVR's skip having to implement a bunch of discrete (multiple parts), PLL clock recovery circuity with newer designs.

Cost savings.

[smorse]

lcaillo wrote:

Are there any measurements that support the notion that jitter is a signifficant problem at all these days?

Jitter is easily (well, not easily, but readily) measurable -- just pick up any issue of stereophile magazine -- it varies even among the modern top-tier compnents they tend to review. It can be measured at home too (at least to the point of being able to determine which of two setups exhibits less jitter). The question is whether the jitter levels that exist in modern receivers is audible. I have only one AVR, so I can't compare directly -- but I have compared various modern digital sources (sony DVD player, squeezebox, etc) and found sound quality to track well with lower jitter levels. I also was able to clearly see that coaxial S/PDIF connections exhibited less jitter than toslink connections.

Also, see if you can find robert harley's excellent explanation of how jitter affects signals in an article of The Absolute Sound -- published sometime over the past 12 months or so I think. In it, he explains how jitter causing the conversion of a signal sample at the wrong time actually is the same as the signal being reproduced at the right time with amplitude errors, which one can easily believe as being audible.

[Lordoftherings]

There are so many other variables that affect the audio not only from the Dacs in the analog circuitry,
but in their implementation also, that we can talk about this till the end of times.

Balanced Differential mode, Dual Balanced Dual Differential mode, Capacitors, Resistors, Digital Filters, etc., etc., etc... Dual layer epoxy circuit boards, copper quality, silver, gold, short circuit paths & wiring,...

* And more magic extra tweeking...

** Oh, and what about AKM Dacs too, and ESS Dacs? And of course Wolfson Microelectronics Dacs?

1. Anthem Pre/Pros like to use AKM Dacs in Dual Differential mode.
2. Oppo BDP-83SE Universal Blu-ray Player with ESS Sabre32 Premier Dacs.
3. Pioneer Elites like WM Dacs, in their A/V Receivers and Blu-ray players.