DAC chip signature sound?

[stunta]

I've come across several forums where people are saying different DAC chips (Sabre, Wolfson etc.) have different sonic signatures. Sabre DACs are supposed to sound "edgy" for instance.

I am not highly technical in this area, but it is my understanding that there is a lot more than just the DAC chip in a consumer DAC unit that can affect the output. Given this, are such claims of sonic signatures valid?

Thanks!

 

[Purité Audio]

It’s tosh, perhaps if you looked at fundamentally different designs, delta sigma compared to r2r or their implementation, or reconstruction filter.
Remember DACs are digital to analogue converters, they should be transparent.

Keith

 

[restorer-john]

Sabre DACs are supposed to sound "edgy" for instance.

Sabre would be edgy for sure.
Wolfson- probably has a nice bite to it.
Burr Brown- probably detailed but with a smooth chocolaty smooth sound
Asahi Kasei (AKM)- typically Japanese sounding, but great with a beer.
Cirrus Logic- a bit nebulous sounding, cloud-like presentation but pleasant.
Technics MASH- Just a sonic puree
Philips Bitstream- non stop assault of mediocre sound

I could go on, but it's all BS.

 

[stalepie]

Wolfsons generally have a warm, inviting sound.

 

[Wombat]

Wolfsons generally have a warm, inviting sound.

Big, enveloping and overpowering.

 

[restorer-john]

...Big, enveloping and overpowering...

Or sneaks up behind you and takes your arm off...

 

[restorer-john]

It's like the capacitor brand names and physical presentation:

Here's a few. Tell me, without testing them, how will these 'sound'?

Silmic- uses supposedly silk fibres, but a simple brown sleeve.
Cerafine- uses powdered fine ceramics powder and a beautiful metallic cherry red and gold livery or black and gold in larger values...
Starget- electric blue and gold IIRC- very pretty, reminiscent of the night sky?
Blackgate- a special carbon/graphite formulation- black and gold- very manly.
Muse- a gorgeous emerald green sleeve and magical properties.
Fine Gold- a lovely metallic gold sheath, black writing and audiophile pretensions.

Go back to the 1960s and 1970s. Electrolytic capacitors were (sleeve/print) - blue/black, grey/black, black/white, orange/black, yellow/black etc.

 

[Wombat]

The most dogged proponents of this stuff are usually ignorant of electrical, electronic and audio fundamentals let alone a moderate level of understanding. Their sources of information tend to be marketers and internet gossipers.

It can be frustrating for those who have put in the effort to attain knowledge and experience in the field to deal with 'instant experts'.

P.S. Electrons don't have feelings.

 

[RayDunzl]

P.S. Electrons don't have feelings.

I checked to see if they expire.

Current estimate for minimum lifetime is 66,000 yottayears.

 

[Jorj]

I've come across several forums where people are saying different DAC chips (Sabre, Wolfson etc.) have different sonic signatures. Sabre DACs are supposed to sound "edgy" for instance.

I'm just as ill informed as the next guy, so as an American, I will not scruple to give you my opinion.

I think that there is something to the Sabre (and AKM and Burr-Brown, etc.) sound, but that has more to with the 'approved implementation' hardware that often comes with them. If you get an ESS DAC chip, there are a few other bits and bobs that come with it in order to make it work well, and rather than muck about with those, most companies just pop them right on the PCB in an approved configuration.

There are some out there that do it differently, I'm told, and that will affect the sound profile, perhaps. I have listened to a lot of different DACs, and while I have found some to be more resolving than others or to have a house sound, component nervosa is kinda like chasing your tail. If the spec is sufficient, and the implementation is good, it will produce an analog signal so purely that your ears will not be able to tell it from one by another maker that is equally good. Do they sound different? Sure, I guess, but the bigger question is how to quantify that when the graphs all look pretty much the same.

I wish I knew what metrics could even be used to define empirically all those glowing terms we love to toss about. Soundstage. Bright vs warm. Instrument placement. Separation. Detail. Even if we could do that, you still have to take the totality of the signal chain into account, and the measurements would have to be taken from the listening position, with all the issues that come with that. At some point, it's like philosophy. Lots of arguing and hand waving for not much, if any, result (and possibly hurt feelings). My advice? Build your system with components that test well and fit your budget. Spend more on your speakers (or headphones). Don't get taken in by snake oil cable salesmen. Buy quality recordings. Listen. Enjoy. Listen some more. If it does not sound right, or grates your nerves, try some other components. See if it sounds better. If you are someone whose auditory system is one of those that is connected to your internal accounting system, buy some Dan D'Agostino monoblocks and Wilson speakers. It'll make you happy. Me, I'm good with the occasional ChiFi signal processor box, early 90's mid-fi gear, and lossless Redbook.

All that said, I have never once heard someone walk into a listening room or clamp on some cans and say, "That is unquestionably an AKM DAC" without seeing the box first. There's a name for that...

 

[garbulky]

The only answer I can give to that is my experience from non double blind non level matched subjective listening. But since you asked "are such claims of sonic signatures valid?", I believe you are looking for a more objective type approach. So I don't think my answer will be useful to you. Hopefully somebody can refer you to some studies.

 

[stalepie]

The sound of ESS Sabre:

Or is it this?

probably the former, but who knows.

 

[RayDunzl]

probably the former

I find electronic music to generally be lacking a little soul.

Here's a possible exception

 

[stalepie]

wehhhh if that has soul I don't want one!

 

[maxxevv]

Interesting but if you were to ask Rob Watts, the guy who design most of Chord's stuff, as I attended one of his talks, I think he would say there are differences.

Its not just a straightforward D to A conversion, as the algorithm and approaches used for different chips vary . Chord's approach being different from either DS or R2R ladder.

If I understood his talk correctly, he was explaining that a lot the details we get to hear are a result of the final signal reproduction amidst the need for noise/ jitter rejection in the D to A process. And different rejection rates and different reconstruction methods as well as resampling / upsampling methods can alter the final waveform that's reproduced. Thus affecting the "signature" that is ultimately audible to our ears.

And of course, a lot of what we hear has more to do with the endstage output ( such as the opamps and related circuitry) than the digital front end too. So there is no direct comparison to say that an AKM AK4497SQ chip sounds 'mellower' than a ESS ES9038PRO chip unless the circuitry and endstage amplification designs and components are identical.

Heck, on my portable DAC Amp that allows opamp chip rolling, simply swapping the bundled 5534 chips to a pair of AD797 made a very discernible difference to the signature in terms of bass and mid frequency output.

 

[rebbiputzmaker]

Interesting but if you were to ask Rob Waites, the guy who design most of Chord's stuff, as I attended one of his talks, I think he would say there are differences.

He need come here. I have learned a great deal being here.

 

[Wombat]

Eschew tone controls and EQ, then engage in tube and IC rolling, before or after hunting for similar but 'different' components that sound 'right'.
Seems like an expensive form of audio self-flagellation to me.

 

[maxxevv]

Eschew tone controls and EQ, then engage in tube and IC rolling, before or after hunting for similar but 'different' components that sound 'right'.
Seems like an expensive form of audio self-flagellation to me.

You lost me there. Completely.

 

[Beherit]

Many of the chips do seem to have a specific timbre across designs. All of the ESS 9018/9028 and newer AKM chips are rather colored in their own ways. The 9018/9028 DACs have warm bass and treble timbre issues while those based around newer AKM chips like AK4490 have veil/compression in the midbass punch and low treble. They all have this, even the designs that bump up the bass subjectively a bit like the Grace and Schiit DACs.

These chips do a ton internally and these IC manufacturers would be idiots to share their audio dac trade secrets. They're basically black boxes beyond the supplied spec sheets, chip layouts, and example implementations. ESS even makes people sign non-disclosures to view those.

 

[Wombat]

Many of the chips do seem to have a specific timbre across designs. All of the ESS 9018/9028 and newer AKM chips are rather colored in their own ways. The 9018/9028 DACs have warm bass and treble timbre issues while those based around newer AKM chips like AK4490 have veil/compression in the midbass punch and low treble. They all have this, even the designs that bump up the bass subjectively a bit like the Grace and Schiit DACs.

These chips do a ton internally and these IC manufacturers would be idiots to share their audio dac trade secrets. They're basically black boxes beyond the supplied spec sheets, chip layouts, and example implementations. ESS even makes people sign non-disclosures to view those.

How can you isolate the chips from the overall, specific, circuit designs and interconnection parameter differences by subjective listening to a device. Are you expressing this viewpoint from a knowledge-based and practiced experience or internet opinions.