I prefer my TeraDak Chameleon over my DAC19DSP for LCD2, would getting a Reference 8 be unwise... - Page 2

Quote:

Originally Posted by realmassy 

Cheers Wood!

Definitely the last thing I want is a sterile and thin sound, so thanks for the advise. I'll look into the galvanic isolator.

Will let you know how I get on with the jkenny 'salamander'  usb direct to i2s board. I am hoping it will improve the sq considerably but not interfere with the 'soul' of the dac.

I see you use the halide bridge, this should sound superior to using a adum by itself. Also I don't know if the adum will work with the halide bridge? However if you are finding direct usb is preferable to the halide bridge, then the adum with usb will be better still.

I own the Mini dac X 4 mutibit phillips dac chips - and prefer it over the Dacmagic and slightly more than the Bryston hi end DAC.  I don't find that it is "musical" or "colored".  I find it sounds neutral with good transients and free from an "artificial sheen" that is rampant on the Dacmagic - and slightly detectable on the Bryston DAC (which has an upsample bypass feature).  The Havana sounds like a good DAC, however I am concerned that the tube output stage will "color" the output.  Is there a highly acclaimed multibit DAC that is free from the "artificial sheen" and solid state in output.

Don't get me wrong, I like tubes - but prefer the DAC solid state, so I can choose a tube amp for the tube goodness, or a solid state amp for the clinical precision - but without the "artificial sheen".  I am after a neutral and precise DAC - one that does not exhibit this unnatural coloration that exists on the Dacmagic and Bryston - is the Audio GD reference 7 the answer with the mutibit R2R 1704 chips?  or does the answer lay in a multibit NOS implementation?  I hope someone also understands the "artificial sheen" which I refer to and can help me find a neutral and accurate dac that fulfills my need.

Your Muse dac does not have an output amplifier to boost the signal from the TDA1543, its output relies on current generated by the 4 chips and the IV resistors which convert the current to voltage. The reason your Muse, the Valab and Chameleon dacs use more chips, 4, 8 and 16, is fourfold.

First, the resolution of the dac improves (this more than compensates for the fact that a single TDA1543 is less resolving than other ladder type dac chips).

Secondly the output impedance decreases to improve loading with downstream amplification.

Thirdly, the output power increases to give improved dynamics and the ability to drive an amp.

So the fourth benefit...... because the output amplifier is the weak link in dacs that use them, such as the Bryston, (as it is the point where coloration could be introduced), by dispensing with an output amplifier, and relying on the combined output of multiple chips  the whole issue of coloration is avoided....or is it?.

No, because on the Muse, Valab and Chameleon the two IV resistors will introduce coloration. After a lot of experimenting with different resistors on the Chameleon and Valab threads we found that carbon comp, the same in your dac, are the least colored but have lower resolving power than all the boutique resistors that were tried, but one. We found that only the Texas Components TX2575 Z foil resistors had more neutrality than carbon comp, and has the transparency, in spades, that carbon comp lacks.

The second factor is that the muse uses two electrolytic capacitors as DC blockers on the output signal that goes to the RCA. These will also introduce coloration, and so would benefit from a change to film caps.

Of course this is all just about the output, the analogue realm, so to speak. The input, the digital realm, is another story.

The big surprise is just how much the the SQ of a dac can be influenced by different types of resistors and caps, its not subtle, believe me,  even the decoupling capacitors for the TDA1543 will assert their own strong SQ flavors.

After a lot of experimenting, I found using Blackgates for decoupling, Teflon for coupling (yet to try Transformers), and TX2575 for IV had the most neutral sound.  My conclusion is that a dac designer will use parts in the output stage amplifier (cheap as possible) to 'tune' it to their approximation of what 'realistic' is to his ears, so frankly, I have given up on hope of buying an of the shelf, truly neutral, dac, and have had to teach myself how to use a soldering iron.

Quote:

Originally Posted by Wood 

Your Muse dac does not have an output amplifier to boost the signal from the TDA1543, its output relies on current generated by the 4 chips and the IV resistors which convert the current to voltage. The reason your Muse, the Valab and Chameleon dacs use more chips, 4, 8 and 16, is fourfold.

First, the resolution of the dac improves (this more than compensates for the fact that a single TDA1543 is less resolving than other ladder type dac chips).

Secondly the output impedance decreases to improve loading with downstream amplification.

Thirdly, the output power increases to give improved dynamics and the ability to drive an amp.

So the fourth benefit...... because the output amplifier is the weak link in dacs that use them, such as the Bryston, (as it is the point where coloration could be introduced), by dispensing with an output amplifier, and relying on the combined output of multiple chips  the whole issue of coloration is avoided....or is it?.

 

No, because on the Muse, Valab and Chameleon the two IV resistors will introduce coloration. After a lot of experimenting with different resistors on the Chameleon and Valab threads we found that carbon comp, the same in your dac, are the least colored but have lower resolving power than all the boutique resistors that were tried, but one. We found that only the Texas Components TX2575 Z foil resistors had more neutrality than carbon comp, and has the transparency, in spades, that carbon comp lacks.

 

The second factor is that the muse uses two electrolytic capacitors as DC blockers on the output signal that goes to the RCA. These will also introduce coloration, and so would benefit from a change to film caps.

 

Of course this is all just about the output, the analogue realm, so to speak. The input, the digital realm, is another story.

 

The big surprise is just how much the the SQ of a dac can be influenced by different types of resistors and caps, its not subtle, believe me,  even the decoupling capacitors for the TDA1543 will assert their own strong SQ flavors.

 

After a lot of experimenting, I found using Blackgates for decoupling, Teflon for coupling (yet to try Transformers), and TX2575 for IV had the most neutral sound.  My conclusion is that a dac designer will use parts in the output stage amplifier (cheap as possible) to 'tune' it to their approximation of what 'realistic' is to his ears, so frankly, I have given up on hope of buying an of the shelf, truly neutral, dac, and have had to teach myself how to use a soldering iron.

 

Thanks for your reply, it has been most informative.  The only reason I bought the Muse dac was because I was initially very interested in the Chameleon DAC, but was concerned that it would be unacceptable in terms of sound quality without modifications.  However this remarkably positive experience with paralleled TDA1543 chips has eased most of these concerns...it appears I must purchase the Chameleon and also an entry level AGD PCM1704UK after selling off both Dacmagic and the Bryston in order to experience the sonics for myself.  It looks like my avoidance of the soldering iron has failed.  Cheers.

If you don't want more technical ability, then you don't want more expensive audio-gd dac's, CountChoculaBot. The musical line is supposed to be "neutral but slightly sweet, slightly warm".

Maybe something you would like is for audio-gd to modify their ST-7 preamp for headphone use. It appears to be more colored than the musical line, and note it was designed to be a "complimentary" pre-amp, meaning you may not want it as your only pre-amp due to the way it deviates from neutral. But I think it might suit your tastes as your only main headphone amp because you said you don't care for anything but a charming sound.

If you consider having this modded for headphones, please keep in mind what I said about it being unneutral hehe. And although you say you don't care for it, chances are it has very good technical ability.

http://www.audio-gd.com/Pro/pre/ST-7/ST7EN.htm

The chameleon is no joke, I you might be hard-pressed to find a real upgrade to it if you want a NOS dac. Not without spending major money anyhow. It's stupid easy to mod too, I threw some of those TX resistors in mine and replaced the output caps with some decent quality caps, improved in spades for about 100$ parts cost. I see people suggesting the Havana, im not sure it's much of an upgrade. 

The Chameleon got knifed in the back by Six Moon big time.  They didn't do the right thing by not comparing to Chameleon as sold with rivals in its price range like the Dacmagic and insisted on carrying out comparisons with expensive dacs.  They didn't even damn it with faint praised, they just plain damned it.  The Chameleon thread is also donig a fine job of killing the Chameleon.  The heading of this thread has been the most positive praise I have found about the Chameleon on the internet.

Hopefully I will have the Chameleon side by side with the Dacmagic, REF 7 and Bryston - and if a a 60 dollar mutibit DAC fared well, I don't see why the Chameleon won't.  I think a lot of people listen with measurements - yet these same people fail to recognise the superior square wave reponse of multibit NOS dacs over 1-bit dacs.
 

Quote:

Originally Posted by MikeW 

The chameleon is no joke, I you might be hard-pressed to find a real upgrade to it if you want a NOS dac. Not without spending major money anyhow. It's stupid easy to mod too, I threw some of those TX resistors in mine and replaced the output caps with some decent quality caps, improved in spades for about 100$ parts cost. I see people suggesting the Havana, im not sure it's much of an upgrade. 

I hope you compare them and post your impressions. I think it's kind of unfair to judge the chameleon without looking at the modding potential. It was designed to be modded in every imaginable way. It's like buying a model and complaining that it comes in pieces. It's a pretty specialized market, and i'd say well over half of chameleon owners knew what they where buying and already had a gameplan to mod it before it ever arrived at their door step.