• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Mr White's "Opus", designing a simple balanced DAC

Jump to Latest
#681
ezkcdude said:
Russ, do you have some preference for SRC4192 over AD1896? The two are pin-compatible, so your design will work for both. SRC4192 is quite a bit cheaper.
Click to expand...


Yes, I do prefer the SRC4192 as it is more flexible in XO choice(up to 50mhz), and has better specs. Even if the AD were cheaper the price difference would probably not be enough to warrant the cheaper part, and it would be a compromise I can't see a good reason to make, especially over a few dollars. I also like the fact that with the SRC4192 when using more than one the phases always match without doing anything special.

Cheers!
Russ
 
#682
ezkcdude said:
Russ, do you have some preference for SRC4192 over AD1896? The two are pin-compatible, so your design will work for both. SRC4192 is quite a bit cheaper.
Click to expand...


Ah yes, the SRC4192 is less expensive that is correct. It is also better, so I guess its a double win for BB. 😉

What is most important to me is that the SRC4192 seems to have a clear edge on THD+N.

Price is not always a good indicator of quality in ICs. A stellar example is the LM4562 which is dirt cheap, but easily outperforms opamps at three times its price.

Cheers!
Russ
 
#687
When I did some basic searching for info about ASRC chips, the consensus seemed to be that the SRC4192 had excellent specs in the datasheet, but fell short in the real world. The threads I found were from 2005, so have these deficiencies been resolved?

As an anecdote, I talked with Justin of headamp yesterday, and he said he tried both and greatly preferred the sound of the AD1896.
 
#689
Hi Russ,

For those of you who wish to use the I2S out from the CDPRO modules into Opus this re-clocker module would certainly fit the bill, as you would no longer need to worry about obtaining a master clock. All you need as input into the ASRC is bitclock, wordclock, and data.
Click to expand...

This is great news for an all I2S upsampling CDP.
I'm glad you are planning a SRC like I suggested on
Post #66 😉

Cheers,
M
 
#690
jar said:
When I did some basic searching for info about ASRC chips, the consensus seemed to be that the SRC4192 had excellent specs in the datasheet, but fell short in the real world. The threads I found were from 2005, so have these deficiencies been resolved?

As an anecdote, I talked with Justin of headamp yesterday, and he said he tried both and greatly preferred the sound of the AD1896.
Click to expand...


Yes, in my research I have read probably a dozen reviews of the three major players, each getting the upper hand at different times.

I have talked to the folks at TI and am very confident about the SRC4192. It just a much more flexible chip with better specs. I have no reason not to try it.

Cheers!
Russ
 
#691
As an example, the sharp folk at Apogee(Lucas van der Mee) ditched the AD1896 in favor of the SRC4192. I am sure they did not do so haphazardly.

Lucas van der Mee: Indeed, for digital I/O, we have been using it. We used the AD1896 from Analog Devices because that was the first SRC chip with acceptable performance. Later we switched to TI's SRC4192 when that came out, because of its better performance.
Click to expand...

That from "Pro Sound News".

I will also test both for myself.

Cheers!
Russ
 
#692
maxlorenz said:
Hi Russ,



This is great news for an all I2S upsampling CDP.
I'm glad you are planning a SRC like I suggested on
Post #66 😉

Cheers,
M
Click to expand...


See I always knew you were forward thinking. 😉

It has always been in the back of my mind, but I like to break things down in to manageable chunks I can (hopefully) do well, not just do... 🙂 Thats really why I decided on a modular approach, so I could add capabilities as I go.

Thanks for sticking with me. 🙂

Cheers!
Russ
 
#693
Also Bob Bauman from Lynx also likes the SRC4192

Bob Bauman: Yes we do offer SRC in the digital I/O of some our products. The user has the choice of enabling this function or not. We have used the TI SRC4192 because of its excellent performance at any sample rate conversion ratio.
Click to expand...

That from the same article as before.

I own a couple Lynx Studio products ("Lynx Two" is my sound card), and they are top notch.

Cheers!
Russ
 
#696
Re: Further reading

Russ White said:
This post sums up what some think, despite the detractors, for the most common set of scenarios the SRC4192 would appear to have the upper hand. In cases of incredibly bad input jitter the AD1896 may be slightly better.

http://www.diyaudio.com/forums/showthread.php?postid=528312#post528312

The CS8421 certainly does look interesting though.

Cheers!
Russ
Click to expand...

Interesting post. Still seems important to keep input jitter down to modest levels, but with the Wolfson DIR that shouldn't be a problem.

Dan
 
#699
Re: Re: Further reading

Spartacus said:


Interesting post. Still seems important to keep input jitter down to modest levels, but with the Wolfson DIR that shouldn't be a problem.

Dan
Click to expand...


I believe that ALL of our I2S sources should be more than adequate for the SRC4192. You would have to have massive jitter to reach the high PLL rate for the SRC4192. Now I suppose if you had some ridiculous jitter coming from your source (or PC) then there is no SRC that can save you. Time for a better source. 🙂

So it comes back to overall performance and I am with the majority pros(at least that I can find so far) on that one. I am making some phone calls today to some of my engineer/producer friends here in Nashville just to get some other opinions.

Some of my gear(Lynx) uses the SRC4192 and I have been extremely happy with it.

I am still researching the CS8421 as I have no practical experience with it. On paper it looks fine. 🙂 The circuit would be more complex though, and perhaps not worth the extra trouble and cost of the additional parts.

Cheers!
Russ
 
#700
Russ White said:
The AK4125 actually looks like a respectable part too, but I am not sure how hard they would be to source. Its specs are no worse than the AD1896. I have no idea what they cost either 😀
Click to expand...

The AK4125 work in a different way from the AD1896 and SRC4192. It uses a PLL to extract the internal clock from the incoming data, it does not use an external clock or crystal. The advantage of this is that the output port can be run in master mode. This allows you to add an external master clock for the output data, which means that the output port is almost completely jitter free.
 
You must log in or register to reply here.
Top Bottom