Differences in sound: DAC vs. DAC + Pre-amplifier

[Miska]

On 1/24/2020 at 1:26 AM, SoundSparks said:

• And another silly question: what separate pre-amplifier would be required in order to at least correspond to the specs of this DAC (SNR >-129dB 20Hz-20kHz A-Weighting and THD+N <0.00030% at 20Hz-20kHz A-Weighting)?

 

Benchmark LA4 would at least fit the bill, but there are likely others.

 

[Miska]

1 hour ago, bodiebill said:

Not saying that I tried very many, but the first preamps I was happy with for my Lampizator DAC were LDR (Light Dependent Resistor) passive preamps: Lightspeed Attenuator and Stereo Coffee DIY. Both really affordable. I never achieved that quality with software VC.

 

Challenge with passive ones is that they drastically modify output impedance seen by the amplifier following the passive attenuator. In addition this impedance is not constant by varies depending on volume control position. This in turn increases the impact interconnect cable has. In addition power amps typically have first order low pass filter in their input that assumes low output impedance from the source. Varying output impedance means varying frequency response of this filter.

 

This is why active buffer stage (impedance converter) is usually good idea after passive type volume control, such as potentiometer. In addition, the active buffers I've been building have an additional cable capacitance compensation circuitry.

 

Additional resistance on the path also increases noise. Higher the series resistance, higher the noise. ESS Sabre chip for example has 600 ohm worth of resistance/noise. So if you want to use analog volume control that has lower self-noise than good digital one combined with a good low impedance DAC. You need to have all series resistances below 1 kOhm.

 

[Miska]

2 hours ago, bodiebill said:

One practical difference between

 

(1) software VC into DAC (no preamp), versus (2) DAC + preamp

 

is that -- as DAC's often produce a slight hum -- this hum can become intolerably loud with (1) as the attenuation is before the DAC and does not affect the hum, but is mitigated with (2) as in that case there is attenuation after the DAC.

 

So (2) should be quieter. Correct?

 

Hum is usually due to ground currents. So one needs to check out the overall system grounding layout at that point...

 

[Miska]

7 minutes ago, bodiebill said:

Indeed. However I guess, even without hum, DAC into preamp will have a 'blacker background' than software VC into DAC (no preamp).

 

Why would it be? For most cases, from technical perspective that shouldn't be the case.

 

When I output data to a DAC, the digital noise floor in audio band is roughly 100 dB below the analog noise floor.

 

But I don't have hum or other noises, just dead silence if I put my ear next to the speaker.

[Miska]

6 hours ago, Kimo said:

I have talked with Benchmark and they indicated that the best measured performance was obtained by using the DAC at a fixed volume and employing their preamp.

 

At least they can sell more devices by saying so...

 

[Miska]

On 3/1/2020 at 9:47 PM, craighartley said:

Thanks for this. I do have a dedicated server (Roon on a Mac Mini streaming to HQPLayer on a dedicated machine running the HQPlayer OS), and I'll have another try. My problem occurred when I changed some settings in HQPlayer and (without me noticing) this wiped the volume limit I had set in Roon.

 

When you use HQPlayer, it is better to set volume limit in HQPlayer. If you enable "Direct SDM", volume control becomes disabled because you are asking bit-perfect pass-through for DSD data, but this is indicated in the main window because the volume knob becomes disabled as result too.

 

[Miska]

15 minutes ago, craighartley said:

I can’t see where to set a volume limit in HQP Embedded; only where to set startup volume. 

 

On Embedded volume configuration needs manual adjustment of the configuration file as there are some extra parameters that can be set to deal with software volume and software + hardware combination volume.

 

There are "volume_min" and "volume_max" attributes in the configuration file under "engine" element. Default values are min=-60 and max=0.

 

[Miska]

8 minutes ago, motberg said:

Is it required to oversample the file (say Redbook 16/44) to get the best performance for using a digital software volume control, using HQPlayer for example ?

 

It depends on the DAC too, but it certainly helps since it allows more resolution and especially with noise shaping curves increase of dynamic range beyond that of output word length within the original bandwidth.

 

[Miska]

1 hour ago, motberg said:

For HQPlayer, how is best to manage the "gain staging" of the digital volume control? I am upsampling to 384, within the rate family, and have Min Vol set at -60 and Max Vol set at -3.. would this be optimal ?

 

Suitable range setting depends on how you use the volume control...

 

[Miska]

4 minutes ago, ro7939 said:

A state of the art preamp w/analog volume control allows users to swap DACs without consideration for the DAC's output current capacity, without consideration for the existence of nor quality of digital volume within or in front of the DAC, and without consideration for any existing or potential sum total input impedance of the load amplifier(s), which might include sub amps with very low input impedance (the norm for such amps).

 

Preamp can only worsen the quality that comes out of the DAC, not improve it. Generally it increases noise and distortion, more or less.

 

Input impedance of a pre-amp shouldn't differ from input impedance of a power amp.

 

Pre-amp is of course useful when you want to switch between multiple inputs.

 

[Miska]

5 minutes ago, Kimo said:

ESS will tell you analog volume controls are better than digital.  I think there is a paper online from them with measurements.

 

Problem is that for example ESS has inherent series resistance of 600 ohm which defines the thermal noise limits. With analog volume control you are going to have hard time finding one that has equal or less inherent resistance. In addition you'll need additional analog buffer/impedance converter sections which also will have some noise and distortion.

 

7 minutes ago, Kimo said:

Benchmark will tell you that you will get better performance using their DAC 3 at a fixed setting and inserting their preamp.

 

Sure, because they can sell two devices and get more than twice the amount of money that way!

 

[Miska]

22 hours ago, Kimo said:

My DAC employs a log ladder network to create a constant impedance passive attenuator, for whatever that is worth.  I can't tell the difference between fixed mode or the attentuator.   I tried but failed.

 

My amp has a pretty high SNR, so apparently it is possible to make an analog volume transparent to me at least with an ESS chip.

 

You can get some additional transparency by omitting that ladder network and extra buffer stages and going with a good digital volume control instead.

 

[Miska]

10 hours ago, motberg said:

But I think to use software volume control, you need modify the source file digitally. So isn't the choice digital manipulation of the source vs. analog manipulation of the DAC output?

 

Digital manipulation can have precision that far exceeds any analog capabilities. In digital domain, the errors can be for example a millionth of thermal noise achievable at room temperature.

 

[Miska]

8 hours ago, Kimo said:

I am pretty sure Mr. Linnenberg would not agree.  Dynamic range is specified as 138 db.  Anyway, I use the line level inputs for the vinyl rig, so I need an analog control somewhere.

 

I don't know who is Mr. Linnenberg, I can let him disagree, no problem.

 

Even at 32-bit PCM output, without noise shaping, dynamic range is 192 dB. And with noise shaping you can push it to for example 240 dB or more. And then if the digital volume control errors are at or below -300 dB, I would say that is quite a bit better than you can do in analog domain.

 

[Miska]

10 minutes ago, Kimo said:

But once it hits the analog realm, does any of this matter?  If you built a room quiet enough to hear or not hear any of this, could you actually sit in it very long?

 

Limiting factor is the analog domain. Since some people continue to worry about digital volume control, I just gave some example figures to compare analog things against.

 

Of course implementation of the digital volume control (or any other DSP) matters, because that's where the figures come from. Just like analog implementations matters too. Not all analog volume controls are the same, and not all digital volume controls are the the same either.

 

[Miska]

38 minutes ago, CG said:

To me, a much bigger question is the system topology with cables of various lengths connected in various ways in all sorts of combinations, that often are additionally coupled to other nearby electronics gear.  That could be due to electromagnetic coping or plain ole direct connections.  Amazing, but true.

 

On my DAC designs, the output buffer stage has cable capacitance compensation circuits and such, to ensure that for example square wave outputs don't end up having overshoot or other unwanted effects. This also applies to design of the analog reconstruction filter.

 

For example 7 kHz square wave at DSD128 on DSC1:

 

And with same input on a typical DAC:

[Miska]

27 minutes ago, CG said:

Most people don't realize that the input and output impedances affect the performance.  Basically, almost every amplification device is unstable at some point.  Fundamentally, that's because the gain drops over frequency for a bunch of reasons.

 

And this is related to what? And relation to this topic is?

 

Preamplifiers essentially look exactly the same input and output wise as any other line level device. Or these days you should actually call it preattenuator, because it provides only negative gain.

 

[Miska]

16 hours ago, sandyk said:

 For anyone interested in learning more about many of the amplifier distortion mechanisms, a good reference is

Audio Power Amplifier Design Handbook 5th ed - D. Self (Focal, 2009) where he details 11 Different types of Distortion

 

He also had nice series of articles on these topics on the Electronics & Wireless World magazine back in the days.

 

My current amplifiers are actually designed by him (pre- and power)...

 

[Miska]

1 minute ago, barrows said:

It would be interesting to hear @Miska's analysis on this, as he is one who could certainly offer a more considered and detailed view, and he could add his thoughts via a vis LeedH vs HQPlayer VC.

 

There is nothing special in the LeedH volume control, I find it pretty funny if it really is granted a patent somewhere. Quite typical though, explain a simple thing in very complex way to make it sound complex and fancy and in a way fool the application examiner (and anybody trying to read it), such happens a lot in patent field. But it is notable that it doesn't work when you have any other DSP involved that produces samples that have higher precision than the original integer samples. So the integer way they do volume control means that it is useless for example when upsampling, digital room correction, eq, headphone processing or any other DSP is involved.

 

The way they do volume control with integer data is nothing new and it works only when you have max 24-bit input, do volume control and have 32-bit output. Nothing else. Volume control range is the amount of extra headroom you have due to extra number of bits.

 

[Miska]

4 hours ago, audiobomber said:

I agree with all that. All I'm saying is that you need at least a 32-bit DAC if you are using the DAC as a preamp. A 24 bit DAC is insufficient.

 

It is not just about number of bits, it is about how you use the bits. But you can certainly beat anything you can make in analog domain with 24 bits.

 

[Miska]

2 hours ago, Reg19 said:

@Miska: given that one needs a preamp while using HQ Player Desktop with the RME ADI-2 DAC to upsample music to DSD256 with DSD direct enabled (no volume control in DAC), which one do you recommend?

 

It depends on what kind of features you'd like. At the moment I'm using Cambridge Audio Azur 851E (feeding 851W power amp over balanced) these two talk to each other over CA's serial bus. It fills my requirements of good technical performance, enough inputs (3 balanced and 5 unbalanced) and possibility to name each input.

 

But now for another setup I've been eyeing Schiit Freya+ in solidstate buffer mode or if budget allows Benchmark LA4 which has very nice performance although even more short of inputs.

 

But also largely depends on your budget... McIntosh has nice C53 with nice amount of inputs and C49 is not bad either. And Accuphase has number of very nice preamps: http://accuphase.com/pre_amp.html

 

[Miska]

58 minutes ago, MikeJazz said:

Hello!, let me take the opportunity to ask if CA's claims have proven right, concerning the amplification.

"sonic purity of Class A and the efficiency of Class AB for the best of both worlds"

https://www.cambridgeaudio.com/row/en/products/hi-fi/851/851w

Sorry for the side-topic...

 

At least I consider it pretty much true. It is one of those moving bias amps. It is very powerful and can go very loud, very clean. Driving my Elac speakers (with their JET AMT-style ribbon tweeters). I've been very happy with this combo.

 

It does produce some amount of heat. Not quite as much as the Harman/Kardon HK990 amp I have in my Dynaudio speaker set though.

 

[Miska]

Digital volume is certainly one option too. I just have 8 DACs connected to my preamp... 😅

 

My ADI-2 Pro AE is on my office headphone system, connected to Mac Mini M1 and T+A HA200 over balanced. Then the USB input of HA200 is connected to my Windows 10 for Workstations machine, and Denafrips Ares II to my Linux machine and to HA200 over unbalanced. So three different DACs, one headphone amp and two headphones.

 

The regular ADI-2 Pro is on the listeningroom loudspeaker system and sometimes serving as analog/digital input and sometimes as output. Connected over unbalanced to the preamp tape in/out.

Then in addition I have three other systems. Two loudspeaker and one headphone systems. But those were the two involving ADI-2.

[Miska]

2 hours ago, EvilTed said:

I know what the digital only, no preamp sounds like.

Digital sounds close to analog in my system when played through the preamp.

Without the preamp, digital sounds dull, flat, uninspiring and just meh digital.

 

You need better digital then, not coloration with a preamp... ;)

 

[Miska]

18 minutes ago, Jean Paul D said:

does this include adding gain ?

 

In most cases preamp has negative gain, IOW works as attenuator. I have not yet heard of anybody needing gain with digital sources between the source and power amp...

 

Quite typical gain of preamp between digital source and power amp is between -40 dB and -20 dB (from 1/100th to 1/10th of the source level). Which is the case with my preamp + power amp combo too. And -20 dB setting is really loud. At late evening volume can be as low as -60 dB (1/1000th) to -50 dB (3/1000th).

 

So it should be actually called pre-att and not pre-amp. Gain is only needed for vinyl sources...