advertisement


Tenson's DAC Bake-Off (09/02/2013)

I'm happy to consider that perhaps an alternative method of level adjustment is needed, if someone can provide data showing the level of these artefacts. My own small experiment showed better than -180dB. If I understand what John means by 'second-order' effects, this is the PSU rails being modulated by the processor. In this case the PC is external to the DACs (unlike any processor doing upsampling or dithering etc.. jobs inside a commercial DAC).

Additionally I'll use transformers to provide galvanic isolation for each DAC from the computer.
 
The trouble is that your argument for the digitally adjusted signals being similar can also be applied to the (competent) DACs themselves, which makes the whole bake-off moot.

I think you need to use analogue attenuation and perhaps just two DACs at a time.

Paul
 
I see your point, but I don't really agree as the effect of digital volume adjustment (using only a few dB) is so far below the distortions of even good DACs.

However if it pleases, I could use a 5K pot after each DAC output. I could also use one per channel to get perfect channel matching.

Why only 2 DACs at a time?
 
Manufacturers are clearly running scared from any form of controlled testing.

Indeed Mike, and not for the first time.

All very sad but not unexpected.

Nevertheless the test will continue despite the whinging.
Bollox to them :)

I am happy to sit back secure in the knowledge that John has forgotten more than you (or Tenson) have ever learnt about digital audio. It is what he does every single day, including today and tonight.

His ability to throw together a competent dac while blindfold is not in question, nor the ability to produce something technically SOTA.
What he claims to be hearing certainly is.
He has specialism in electronics, not in his auditory capabilities.


Or that the sound differences between DACs are vanishingly small, hence the ridiculous list of requirements by certain manufacturers in order to obtain "best results".

Well put.
 
The trouble is that your argument for the digitally adjusted signals being similar can also be applied to the (competent) DACs themselves, which makes the whole bake-off moot.

I think you need to use analogue attenuation and perhaps just two DACs at a time.

Paul

I see your point, but I don't really agree as the effect of digital volume adjustment (using only a few dB) is so far below the distortions of even good DACs.

However if it pleases, I could use a 5K pot after each DAC output. I could also use one per channel to get perfect channel matching.

Why only 2 DACs at a time?
I think I mentioned the number of dacs earlier when I brought up the idea of using the same short musical piece. I think more than 3 dacs is a problem.

Now it's obvious I've never taken part in a test like this so maybe what I'm saying here is obvious to everyone but just humour me as I don't know how these tests work.

Let's label each of these 5 dacs a, b, c, d and e. How do you propose to step through them? Will it be each in sequence, a to e because if so I don't think that is ideal. I think it better to step through them randomly, for example, if each dac is to be used twice the sequence could be c, a, d, d, b, a, e, e, c, b. With anything more than 3 dacs that process would be extremely difficult for the listeners given the minute differences between dacs where differences do exist but I think it's better than going through them a few times in the same sequence. 5 is too many regardless imo.

Also, I know you replied to this earlier but can you explain how the level matching will work with an instantaneous switch of dac plus replay of the same, short musical piece?

Surely connection of each dac to a passive preamp, instantaneous switching and repeating the same musical piece is the only way to do this properly? I'm not sure how that would be easy to do.
 
I think I mentioned the number of dacs earlier when I brought up the idea of using the same short musical piece. I think more than 3 dacs is a problem.

I can't see any problem with auditioning one pair of dacs at a time.
It's operationally easier too. I'll speak with Simon tonight when i see him, preferably before the alcohol kicks in :)

Surely connection of each dac to a passive preamp, instantaneous switching and repeating the same musical piece is the only way to do this properly? I'm not sure how that would be easy to do.

Brian, the problem is that whatever variations to the test are proposed, certain people will raise objections. Now I'd have no problem with that, nor I expect would Simon, if these same objections were ever mentioned during a regular sighted listening demo, be it a factory, home or dealer demo.

If someone comes onto a forum and states that they've bought say a Weiss dac and that it sounds better than the incumbent Cyrus, you won't see manufacturers barking questions about digital attenuation, gain structures, listener fatigue etc etc.
The opinion is more or less accepted.
Yet the minute some controls are introduced, certain manufacturers start raising objections.

Franky it stinks.
 
If we reduce the number of DACs in each comparison, say to 2, then we need to adopt an ABX comparison. This is A, this is B - is X a or B? This is possible but I'd have to make sure each DAC gets a fair number of plays. With 5 DACs to compare, it would take many comparisons to make sure each one gets compared to all the others on the list. (A to B, A to C, A to D. B to C, B to D. etc..) and that only gives each comparison one trial.

My original plan was to randomly sequence through each DAC (playing each only once) and ask the listeners to identify each, and do this 5 times. Between each of the 5 runs I could refresh their memory of how each one sounds, letting them know the identity.

I prefer the second approach as it seems like it would take less time to do, and not bore people too much.

Brian, for restarting the tracks I would have to hit a key on the computer each time I swap input. Not overly difficult with 2 hands :)
 
I can't see any problem with auditioning one pair of dacs at a time.
It's operationally easier too. I'll speak with Simon tonight when i see him, preferably before the alcohol kicks in :)



Brian, the problem is that whatever variations to the test are proposed, certain people will raise objections. Now I'd have no problem with that, nor I expect would Simon, if these same objections were ever mentioned during a regular sighted listening demo, be it a factory, home or dealer demo.

If someone comes onto a forum and states that they've bought say a Weiss dac and that it sounds better than the incumbent Cyrus, you won't see manufacturers barking questions about digital attenuation, gain structures, listener fatigue etc etc.
The opinion is more or less accepted.
Yet the minute some controls are introduced, certain manufacturers start raising objections.

Franky it stinks.

Some of your controls are likely to alter the sound of the DACs.

It is a foregone conclusion - they will all sound the same!
 
Steven, the issue I take with your suggestion is that simply because we can detect one of the controls introduces a change or distortion, doesn't automatically mean it will completely flood the differences between the DACs.

How can a music producer ever make a decision about whether a vocal is better in the mix with a little attenuation, if the act of attenuating it completely floods the transparency of his monitoring system?
 
Steven, the issue I take with your suggestion is that simply because we can detect one of the controls introduces a change or distortion, doesn't automatically mean it will completely flood the differences between the DACs.

It is obvious that this "DAC bakeoff" is to 'prove' a particular point.
 
It is obvious certain people are utterly desperate not to have "proven" a particular point Steven.

Etc Etc Etc. Ad in finitum............

It's amazing what some people claim is transparent - particularly when it comes to speakers and their room.

Yet question the transparency of 1db of digital attenuation instead of 0.5db and the same people are up in arms.
 
...I could use a 5K pot after each DAC output. I could also use one per channel to get perfect channel matching

Really? Surely that could well raise the output Z and, as such, cause rather more damage than the problem it's supposed to solve ;)
 
Having spent years and lots of money on hifi it has become clear to me that the differences between well designed components are very small. I use analogue as my main source and I regularly use three different turntables, all with different arms and cartridges. My conclusion is that though they have subtle differences in presentation, I could easily be happy with any single one of them.

I am really interested in this bake off and I am disappointed that I cannot attend. I use a squeeze box touch in my second system through a Beresford Caiman streaming lossless files. I always thought that I would eventually integrate my digital into my main system when I could afford a comparable DAC to my analogue front end. I really hope that major differences are identified and a stand out performer is found as it is an area that I want to be more involved with.

For what it's worth, my decks are a 401 and a Goldring G99 both in slate plinths. The Garrard uses a Nottingham 12" Ace Anna arm and an Art-1 cartridge. The G99 has an Origin Live Conqueror and Denon DL-S1. I have a TD 125 with audiomods V arm in a solid wood plinth made by Russ Collinson and an AT33ptg.
This all runs through a Leema Agena into Bryston amps and Shahinian Obelisks.

A DAC of comparable quality would definitely be on my shopping list for 2013 and I really hope that you find one. I would always try to get a home demonstration anyway but a short list would be a lovely thing!
 
It is obvious certain people are utterly desperate not to have "proven" a particular point Steven.

Etc Etc Etc. Ad in finitum............

It's amazing what some people claim is transparent - particularly when it comes to speakers and their room.

Yet question the transparency of 1db of digital attenuation instead of 0.5db and the same people are up in arms.

If the object of the exercise is to 'prove' a particular point it can hardly be referred to as a bakeoff.
 
I really hope that major differences are identified and a stand out performer is found as it is an area that I want to be more involved with.

This will not be allowed to happen.
 
Steven
Are you trying to imply that this exercise is being rigged to suggest all dacs sound the same irrespective of price?
 
This will not be allowed to happen.

I know that there are two schools of thought on this, including your own, but I hope that the actions of the organiser are genuine so that we benefit.

If it is found that there is little difference then so be it. If there is a difference then I will arrange a home demo.
 
Perhaps the resident dac designers should devise their own "bake off" , with Steven Toy deciding, in the event of a failure to agree, which is best?
 
This will not be allowed to happen.

Come off it Steven, Simon is doing whatever he can to provide a truly level playing field here. I can understand technical folk criticising certain minutiae of his methodology, but assigning motives due to one's own biases or paranoia or whatever is a bit much IMO. It is also very rude. The whole point here is to remove the usual volume-riding, dancing salesmen etc, etc. If you were a real subjectivist and only interested in what you can hear you'd be all for it!
 


advertisement


Back
Top