Best DAC?

[MrBurns84]

17 minutes ago, Assisi said:

With an Ideon separate Streamer, Clock and DAC you would a little change out of A$150k.  I am reliably informed that the performance is very serious.  I will not be listening or trialing as the three units are not affordable for me.  Just a fantasy dream.

John

If this is a benchmark ..and just reading the specs on the dac - it uses the ES9038PRO DAC chip.  

[BugPowderDust]

Choosing a DAC is a highly personal decision, and your experience may vary. I own DACs ranging from a $200 Teac headphone amplifier to an MSB Premier DAC worth around $60k, and in my view, the differences between them aren't always dramatic.

 

While the MSB is enjoyable to listen to, I can't help but feel that some of my preference might come down to confirmation bias.

[garyjac]

1 hour ago, Assisi said:

With an Ideon separate Streamer, Clock and DAC you would a little change out of A$150k.  I am reliably informed that the performance is very serious.  I will not be listening or trialing as the three units are not affordable for me.  Just a fantasy dream.

John

I would consider a Steinway D274 first, for that kind of money, but the rest of the band would run into very serious cost over-runs.  I also have my drinking habit to support.  In the end, it is nice, but it is not a "reference".  The only correct reference is real music unamplified played by real people (however poorly).  That is what we try to approach in our listening spaces. An old idea, but a good one.

[garyjac]

1 hour ago, MrBurns84 said:

If this is a benchmark ..and just reading the specs on the dac - it uses the ES9038PRO DAC chip.  

Which tells us,perhaps, that if we want the ultimate device then we need to have our own chip custom developed.  Going too far?

[aussievintage]

9 minutes ago, garyjac said:

Which tells us,perhaps, that if we want the ultimate device then we need to have our own chip custom developed.  Going too far?

 

Doesn't it, instead, tell us that off-the-shelf DAC chips are all that is necessary for benchmark performance.  You can buy one of those DACs from Mouser for less than $100

[garyjac]

Both...and a reasonable call...well designed auxiliary circuits ... same as most solid state devices.

[POV]

10 hours ago, Assisi said:

Are you suggesting  that all  DACs regardless of their respective expense and quality, will provide the same outcome?

Not at all saying the outcome will be the same John, but to me DACs are one area where the correlation of price and performance is sometimes difficult to rationalise.  I have a heap of DACs at my place and reality for me is that the differences between them are typically not that significant.    I also think that diminishing returns hits pretty hard, pretty early when it comes to DACs.  
 

I trialled the Weiss DAC502 after much encouragement from my dealer and I found that it was in no meaningful way any better than the DAC in my Technics G700 SACD/DAC despite being 4 times more expensive.  In fact they sound extremely similar in almost all areas.  The Weiss is 4 times the price of the Technics.  I could give a heap more similar examples, including how I actually preferred my $600 Topping E70s Velvet to the PS Audio DirectStream DAC that a member here lent me to try.  Something very odd about the way the PS Audio Decodes DSD that gave me listener fatigue.

[gemini07]

Further to the above, I also preferred using my Audia Flight One CDP as both a player and a DAC for a very long time, because of the integration of the CS DAC chip, power supplies and of course the Philips drive mech. 

 

But as a standalone DAC, the Audia Flight was bettered only by the Doge7 (usual caveats applicable).

[garyjac]

Another piece of string...CD transports...at the moment, the interesting new kid on the block is the StreamUnlimited 8 drive and Blue Tiger CD-84 servo card. There are three implementations of this out there.  Just three. These units cost $16,000 and Gryphon Ethos @ $39,000 compared to Pro-Ject's sane price of $3,000!  Guess which one I would try.

[aussievintage]

25 minutes ago, garyjac said:

Another piece of string...CD transports...at the moment, the interesting new kid on the block is the StreamUnlimited 8 drive and Blue Tiger CD-84 servo card. There are three implementations of this out there.  Just three. These units cost $16,000 and Gryphon Ethos @ $39,000 compared to Pro-Ject's sane price of $3,000!  Guess which one I would try.

 

I would play the data from any CD from the file I ripped it to.   I converted all my CDs. 

[option-up]

3 hours ago, garyjac said:

The only correct reference is real music unamplified played by real people (however poorly).  That is what we try to approach in our listening spaces. An old idea, but a good one.

 

It really isn't (a good idea). Just an old one. All kinds of music (the majority by amount, I'd guess) has no "acoustic" reference.

 

Let's face it: any music we listen to with our gear is recorded and therefore electronically processed, meaning it is quantitatively different to unamplified music heard in person.

 

But we're a long way from talking about DACs now.

Edited November 15, 2024 by option-up

[garyjac]

Indeed, a long way.  However, the point of acoustic reference is not that we listen to all acoustic music all the time. That would be silly.  Establishing that system reproduces music that we have actually played or listened to inperson, without intervening electronics is a reference to two things.  One of these, as you say is the quality of recording.  The other, applying the principle of uniformity of physical phenomena is that we then can reasonably expect the non-acoustic "instruments" being heard through the system are what the player/composer/recording engineer intended.  If not, you accept the hifi system as a production implement in its own right.  How much qualitative and quantitative difference will we accept?  How much will the neighbours accept?

[garyjac]

1 hour ago, aussievintage said:

 

I would play the data from any CD from the file I ripped it to.   I converted all my CDs. 

In that case there is further loss.  The Reeves algorithm and its variations are not perfect, in spite of what you are told by interested parties.

[POV]

16 minutes ago, garyjac said:

In that case there is further loss.  The Reeves algorithm and its variations are not perfect, in spite of what you are told by interested parties.

What do you mean by this?

[garyjac]

People who sell computers and software are "interested parties" in the CD transfer game.  They will say their software is perfect, or nearly.

[aussievintage]

9 hours ago, garyjac said:

In that case there is further loss.  The Reeves algorithm and its variations are not perfect, in spite of what you are told by interested parties.

 

The data in the file is identical to the data on the CD.  There is no loss.

Edited November 15, 2024 by aussievintage

[aussievintage]

9 hours ago, garyjac said:

People who sell computers and software are "interested parties" in the CD transfer game.  They will say their software is perfect, or nearly.

 

You are drawing a VERY incorrect and paranoid conclusion about ripping CDs to, say, a flac file

[garyjac]

This is stock standard advertising procedure.  Nothing paranoid involved.  Everyone casts their product in the best possible light, while trying to walk the line of misdescription, which, if crossed leads to a smack on the wrist.9

[aussievintage]

2 minutes ago, garyjac said:

This is stock standard advertising procedure.  Nothing paranoid involved.  Everyone casts their product in the best possible light, while trying to walk the line of misdescription, which, if crossed leads to a smack on the wrist.9

 

You are way off base projecting this onto CD ripping.  All the software I use is free anyway.  Noone is selling anything, and the results are verifiable.

 

Maybe I was wrong about the intent of this thread...

[Hi-Fi Whipped]

Fair to say this thread has peaked and is quickly on the way down…..

[garyjac]

1 hour ago, aussievintage said:

 

The data in the file is identical to the data on the CD.  There is no loss.

Coalson's algorithm for flac is made up of a magic number, maetadata and encoded audio. The encoded audio is, like all network data, composed of a header,  data block and CRC checksum. The presence of the cyclic redundancy check indicates that errors can occur.  Flac is 50 to 70% compressed, but is described as "lossless" just the same.

[aussievintage]

Just now, garyjac said:

Flac is 50 to 70% compressed, but is described as "lossless" just the same.

 

Mathematically compressed like a zip file, and like a zip file, the contents can be restored without error.  No loss.

 

Also, do you imagine a CD is read without error as it plays?  Ripping a CD improves on this as it can reread the data multiple times, and compare it to known good data.  This is a more correct result than a simple CD player ever achieves.

[garyjac]

5 minutes ago, aussievintage said:

 

You are way off base projecting this onto CD ripping.  All the software I use is free anyway.  Noone is selling anything, and the results are verifiable.

 

Maybe I was wrong about the intent of this thread...

OK, it's free software, so the ad motivation is not there.  How do you verify the results?  For example when there is a tramsmission/ decode error how do you know it is being recovered?  What method do you suggest to correct this entropy?  How do you verify/measure that the recovery was lossless/ inaudible.  They may not even be the same thing, I admit that.

[aussievintage]

Just now, garyjac said:

OK, it's free software, so the ad motivation is not there.  How do you verify the results?  For example when there is a tramsmission/ decode error how do you know it is being recovered?  What method do you suggest to correct this entropy?  How do you verify/measure that the recovery was lossless/ inaudible.  They may not even be the same thing, I admit that.

 

 

Multiple reads, reject the outliers, compare checksum to those available on the net (e.g. Accuraterip  https://www.accuraterip.com), you need to read up on these things.

 

 

[garyjac]

1 hour ago, aussievintage said:

 

Mathematically compressed like a zip file, and like a zip file, the contents can be restored without error.  No loss.

 

Also, do you imagine a CD is read without error as it plays?  Ripping a CD improves on this as it can reread the data multiple times, and compare it to known good data.  This is a more correct result than a simple CD player ever achieves.

CD is not perfect by far.  Good CD systems do this in the same manner these days.  However, where is the "known good data" on a CD or in a file?  Zip files contain errors also.  You may not see/hear them, but they do.  So, what is our means of verification and how is an error identified?  One common means is to interpolate data with a quick calculation of what should be between two other data points.  This works very well provided the error rate is very low.  Precision read heads and so on assist with that problem.  

Edited November 16, 2024 by garyjac