Is Sabre ES9018 the best DAC chip right now?

If you're talking about the affordable range, then the ES9018 is one of the better choices. But the Ref 5.2 is based on the PCM1704UK, which is considered a high-end chip, and is supposed to be in a different league altogether. If budget is not your concern then you made the right choice.

You're worrying too much. 

Quote:

Originally Posted by PleasantSounds 

If you're talking about the affordable range, then the ES9018 is one of the better choices. But the Ref 5.2 is based on the PCM1704UK, which is considered a high-end chip, and is supposed to be in a different league altogether. If budget is not your concern then you made the right choice.

Yes it's different. It has a completely different design paradigm. Rather old-school actually.

Different as in better is quite another matter. A matter of taste and implementation really.

As far as cost per piece they are pretty similar. The thing is though the Sabre will do a lot more, so by the time you are done the 1704 design will cost more.

Quote:

Originally Posted by econsumer666 

Can you say that DACs based on Sabre ES9018 should be favoured over other configurations like WM8741 and PCM1704UK? I kinda keep reading over and over again that this Sabre chip is currently the best affordable DAC-chip.

In my view the Sabre is overpriced for what it offers in subjective SQ, its a digital chip so can't cost nearly as much to produce as the PCM1704 yet they're similarly priced. If your primary objective is SQ then you're better off with the PCM1704, or, if you don't need hi-res, a decent implementation of a TDA1545A is just fine and saves a lot of money.

The Sabre ES9018 is not solely a DAC. Its an ASIC with patented upsampling and jitter reduction algorithms, embedded S/PDIF receiver and so on... It's also a modular 8 channels DAC.

The technically advanced ES9018 is a great departure from the simplistic PCM1704 r2r mono DAC. One could say, they are diametrically opposite in the DACs realm, in a technical standpoint.

Saying which one is better, quality-wise, is not simple and may not have an unanimous answer (since it depends on pretty much everything else in the gear). One might be easier to implement though, but I'll let the audio designers answer this.

As an audio designer myself the PCM1704 is by far the easier of the two to implement. That has a bearing on the SQ as designers don't have limitless time to optimize their layouts and component choices.

Yes don't be to concerned as to which chip you have the implementation is far more important. The Sabre32 9018 is a great chip and I love my Sabre 9018 chipped NFB7 but don't expect ANY chip to be considered the absolute best and that's before implementation is taken into account.

Quote:

Originally Posted by Sapientiam 

As an audio designer myself the PCM1704 is by far the easier of the two to implement. That has a bearing on the SQ as designers don't have limitless time to optimize their layouts and component choices.

That's an interesting comment. I've heard the 1704 is hard to implement because of the sensitivity of the grounding of the I/V stage.

All in all, 1704 based DACs tend to be expensive because you need at least 4 DACs to get a balanced design, and the DACs cost about $80 each, at least in the U-K version. Then add in all the additional components that you need vs the 9018 and you are looking at pretty costly part count compared to a DAC designed with the 9018 which since it's 8 channel you only need one at about $40 for a balanced stereo design.

I don't know how big a factor designer time is in the cost of a DAC, I suspect that will depend a lot on the number of units sold. 

ES9018 is affordable with all in one design, Many many high-end use 1704 and AKM as well.

It is how they use the chip, chip itself is not a huge factor.

As a somewhat related note, do you guys think there's such a thing as a "common sound" to all DACs using the same chip?

I'm a believer in the implementation, output stage, etc. being key, and that some 1704 DACs are very musical while others are very clinical and lifeless. Yet it seems that a lot of the "hot new" DACs tend to group together and use the same chip, and I've wondered if the reviewers then to just prefer a certain "DAC sound" when I'm trying to interpret their reviews...

Just musing.

Quote:

Originally Posted by justin w. 

Implementation of the DAC chip is more important than the DAC chip itself.

Please add this on sticky thread.

It is discussed since long time ago, and people didn't make it a common sense yet?

Quote:

Originally Posted by ehlarson 

 

That's an interesting comment. I've heard the 1704 is hard to implement because of the sensitivity of the grounding of the I/V stage.

 

All in all, 1704 based DACs tend to be expensive because you need at least 4 DACs to get a balanced design, and the DACs cost about $80 each, at least in the U-K version. Then add in all the additional components that you need vs the 9018 and you are looking at pretty costly part count compared to a DAC designed with the 9018 which since it's 8 channel you only need one at about $40 for a balanced stereo design.

 

I don't know how big a factor designer time is in the cost of a DAC, I suspect that will depend a lot on the number of units sold. 

Grounding is always a sensitive issue, and yes on the PCM1704 its necessary to pay attention to it for sure. But grounding at higher frequencies is a much bigger issue - compare the symbol rates of the two DAC chips in question. The 1704 can be run NOS, i.e. 44k1. The Sabre runs oversampled by a large factor - 64X is it? Thus if grounding is sensitive on the 1704 (and it is) then its 64X more sensitive on the Sabre. Of course running balanced helps, but getting good CMRR (to reject ground noise) at frequencies above 1MHz is difficult. Add to this the finer process geometry of the Sabre part compared to the TI one - fine geometry CMOS generates much more substantial glitching on the supplies.

A bigger issue than grounding is linearity of the amplifier stage and getting good linearity at 64X OS is that much harder than getting it at 44k1. Which is why you see (hear I suppose!) so much difference in SQ between implementations of the Sabre. A third issue is jitter sensitivity, running NOS means its an issue, but running 64X OS in a noise-shaped system makes this issue a huge one.

I have heard that jitter is more of an issue with DS chip designs.

Grounding with the 1704 is supposedly particularly difficult because of a servo amp that continually attempts to null the output signal wrt ground.