Mhdt Labs Pagoda Review (R2R Tube DAC)

[tvrgeek]

Taiwan is very different from Mainland China and they have been leading silicon production for years now. We are not talking about Chinese clones here. In fact, for the good or the worst, mhdt design is quite unique and innovative.

Sorry, I did not realize it was from Taiwan. I agree, good solid history. Same logistical pipe though.

My personal view though is even though tubes can be quite linear as a VAS, they are darn hard to couple with transistor circuits so not really worth it in the long run. There is a big market for them though. Recently I have seen them as a "cure" ne "bandaids" for poorly implemented class D modules and sloppy DACs. Good marketing, but dubious engineering. I am not saying this unit is either. As I don't care to swap piles of tubes until I got a good one, or swap them every fey years, I am not doing them again.

 

[fcdvpds]

It really feels awkward to see so much discussion about a product by people that even did not listen to it. The mhdt Atlantis (which i bet measures as bad as this one by ASR gold standard) is one of the best sounding DACs I ve ever heard. I ve been using it for more than 2 years now. Firstly with my Luxman MQ-88uc and lately with Shindo Cortese-Charlemagne 80 mono blocks. I ve been enjoying hifi since 97 and have regular visits from other hifi hobbyists. So far no one told me that my system doesn't sound good or that there's something wrong with the Dac. Other DACs I have include Audio Note One Signature, Bel Canto DAC 1, Audial S4, and AK300. I also have a modern Luxman cd player. I knew this dac would measure horrible before I purchased it. Anyone knowing a little about digital audio knows this dac cannot measure well. But that's missing the point. The intent of mhdt is to give a different perspective of how digital audio can sound. Many users love it including myself. The price point seems correct to me and my unit worked flawlessly so far. I highly recommend it.

 

[aj625]

It really feels awkward to see so much discussion about a product by people that even did not listen to it. The mhdt Atlantis (which i bet measures as bad as this one by ASR gold standard) is one of the best sounding DACs I ve ever heard. I ve been using it for more than 2 years now. Firstly with my Luxman MQ-88uc and lately with Shindo Cortese-Charlemagne 80 mono blocks. I ve been enjoying hifi since 97 and have regular visits from other hifi hobbyists. So far no one told me that my system doesn't sound good or that there's something wrong with the Dac. Other DACs I have include Audio Note One Signature, Bel Canto DAC 1, Audial S4, and AK300. I also have a modern Luxman cd player. I knew this dac would measure horrible before I purchased it. Anyone knowing a little about digital audio knows this dac cannot measure well. But that's missing the point. The intent of mhdt is to give a different perspective of how digital audio can sound. Many users love it including myself. The price point seems correct to me and my unit worked flawlessly so far. I highly recommend it.

Then AM radio will sound even better.

 

[BDWoody]

Anyone knowing a little about digital audio knows this dac cannot measure well.

Why?

There are well engineered r2r DAC's out there that show very good measurements.

This is a fuzzbox...

 

[Jim Taylor]

It really feels awkward to see so much discussion about a product by people that even did not listen to it.

People use preliminary investigation to judge further involvement. IF the water isn't cold, they'll go swimming. IF the bicycle is strong enough, they might ride it. IF the house is well-built, they might make an offer on it. IF a table saw has the proper safety features, they might use it.

They need to satisfy themselves regarding the physical or electrical basics before they'll go any further, and if their standards are not met then they reject whatever item is under discussion. Some people are like that, and some aren't. Jim

 

[voodooless]

It really feels awkward to see so much discussion about a product by people that even did not listen to it. The mhdt Atlantis (which i bet measures as bad as this one by ASR gold standard) is one of the best sounding DACs I ve ever heard.

If you hear a DAC something is wrong. These things should be transparent and not bring any sound signature of their own. Measurements clearly show this thing has a signature and chance is high that this will be audible. Hell it can’t even resolve more than 8 bits:

So no, I don’t need to listen to it to dismiss it.

 

[fcdvpds]

Then AM radio will sound even better.

there you go

 

[fcdvpds]

Why?

There are well engineered r2r DAC's out there that show very good measurements.

This is a fuzzbox...

damn...you guys comment on this stuff with very little knowledge about digital audio...this dac does not use a reconstruction filter, each sample presented to the DAC chip results in a DC output voltage that is sustained until the next sample...also all noise at high frequencies is not eliminated. This results in horrible measured performance but within the trash, the actual audio signal is less manipulated and somehow better representative of the original data. Digital audio was simply not designed to work like this but some argue that actually the low-pass filter is not needed and omitting this obligatory filter can actually deliver a more honest presentation of the music. This dac is exactly this, a broken device that offers a perspective of what digital sounds like if a few steps of the digital process are omitted. Measuring this device and commenting about the bad measurement performance is like commenting that unpolished stone is not smooth. Most people that buy this dac know what they are buying.

 

[fcdvpds]

If you hear a DAC something is wrong. These things should be transparent and not bring any sound signature of their own. Measurements clearly show this thing has a signature and chance is high that this will be audible. Hell it can’t even resolve more than 8 bits:
View attachment 174001
So no, I don’t need to listen to it to dismiss it.

dude, everything has a sound signature...you think a musician wants their music to sound as it is without any manipulation? well you are wrong. You can't imagine how much most singers pay to record and master engineers to make their voice sound as good as possible which is completely different than as real as possible!

 

[fcdvpds]

 

[aj625]

damn...you guys comment on this stuff with very little knowledge about digital audio...this dac does not use a reconstruction filter, each sample presented to the DAC chip results in a DC output voltage that is sustained until the next sample...also all noise at high frequencies is not eliminated. This results in horrible measured performance but within the trash, the actual audio signal is less manipulated and somehow better representative of the original data. Digital audio was simply not designed to work like this but some argue that actually the low-pass filter is not needed and omitting this obligatory filter can actually deliver a more honest presentation of the music. This dac is exactly this, a broken device that offers a perspective of what digital sounds like if a few steps of the digital process are omitted. Measuring this device and commenting about the bad measurement performance is like commenting that unpolished stone is not smooth. Most people that buy this dac know what they are buying.

Wow what a joke, digital signal less manipulated means ? Do you mean digital audio is all about having spurious noise ? Did they have it originally in mind to not to have reconstruction filter ? Is not all digital processing a kind of manipulation ? Even your good old calculator does this manipulation. And how a multi tone test without a reconstruction filter with all kind of grass will not affect music, if you say 1khz test does not represent music ? Somehow these grass will not affect magically real music but only multi tone test ? My friend please stop spreading this misinformation that a digital signal manipulation while removes noise but somehow degrades the actual music by some kind of manipulation.

 

[voodooless]

dude, everything has a sound signature...you think a musician wants their music to sound as it is without any manipulation? well you are wrong. You can't imagine how much most singers pay to record and master engineers to make their voice sound as good as possible which is completely different than as real as possible!

I'm not a musician, nor is my DAC. They can add signatures all they want... It's their art. How dare you defile that with your own whims

It's like you buying the Mona Lisa and then deciding that you don't like the colour of her eyes, and go changing them

And no, not everything has a sound signature. SOTA DAC's are all indistinguishable in blind tests, they all sound the same. They totally faithfully reproduce the sound of the digital samples they were given for our poor and flawed ears and brain.

 

[fcdvpds]

Wow what a joke, digital signal less manipulated means ? Do you mean digital audio is all about having spurious noise ? Did they have it originally in mind to not to have reconstruction filter ? Is not all digital processing a kind of manipulation ? Even your good old calculator does this manipulation. And how a multi tone test without a reconstruction filter with all kind of grass will not affect music, if you say 1khz test does not represent music ? Somehow these grass will not affect magically real music but only multi tone test ? My friend please stop spreading this misinformation that a digital signal manipulation while removes noise but somehow degrades the actual music by some kind of manipulation.

i understand your view. put it this way, most design choices comprise trade offs. think about a digital camera. The standard is to have anti aliasing filter. Ricoh GR and a very few other cameras omit the filter. The result is increased resolution but artefacts occur in many situations that need to be removed in post processing. In general, low pass filter is the best option and that's why it's the standard choice in dacs. that being said, it is good to have alternative design options that may suit the taste of some people better. the non-sense is to complain about poor measured performance on a dac without filter!

 

[fcdvpds]

I'm not a musician, nor is my DAC. They can add signatures all they want... It's their art. How dare you defile that with your own whims

It's like you buying the Mona Lisa and then deciding that you don't like the colour of her eyes, and go changing them

And no, not everything has a sound signature. SOTA DAC's are all indistinguishable in blind tests, they all sound the same. They totally faithfully reproduce the sound of the digital samples they were given for our poor and flawed ears and brain.

sure. go buy a SOTA dac and enjoy it.
if they are indistinguishable I wonder why they sell different models...

 

[solderdude]

. the non-sense is to complain about poor measured performance on a dac without filter!

The non-sense is to not use a reconstruction filter as only with it you adhere to the sampling theorem and the output signal is closest to the intended signal.

Of course, everyone is free to do whatever they think is 'better' and it does not hurt to provide options for this.
Just know these options do not improve signal quality/fidelity but deviate from it. That may have audible consequences some may even prefer.

if they are indistinguishable I wonder why they sell different models...

different connectivity, looks, options, bit-rate/bit-depth, price, also wanting a piece of the market (to make money), purpose coloration, aiming for audiophools and their wallets ... you name it.
Plenty reasons other than technical performance.

 

[aj625]

i understand your view. put it this way, most design choices comprise trade offs. think about a digital camera. The standard is to have anti aliasing filter. Ricoh GR and a very few other cameras omit the filter. The result is increased resolution but artefacts occur in many situations that need to be removed in post processing. In general, low pass filter is the best option and that's why it's the standard choice in dacs. that being said, it is good to have alternative design options that may suit the taste of some people better. the non-sense is to complain about poor measured performance on a dac without filter!

You can't compare music with photo. The analogy is only comparable if you interpolate the pixels like upsampling digital audio but since photo requires three colors to produce a picture therefore simple interpolation won't work in real time unlike audio where you have to only interpolate the amplitude values wrt time. Photos can be interpolated in pc though but still algorithm is not like dac which has refind it over the years by using lot more processing power in real time. Hq player pggb and m scaler are the example.

 

[voodooless]

Oe, I missed this one

damn...you guys comment on this stuff with very little knowledge about digital audio...this dac does not use a reconstruction filter, each sample presented to the DAC chip results in a DC output voltage that is sustained until the next sample...also all noise at high frequencies is not eliminated. This results in horrible measured performance but within the trash, the actual audio signal is less manipulated and somehow better representative of the original data.

"somehow"? How then? By definition, you are always wrong with this approach. The chance that a sample value would be the correct average of the whole sample period, is very slim. Therefore it always introduces errors, as is very evident from the data.

If you were to actually make this work a little bit better, you would need to do the following: properly upsample the values to a higher sample rate (let's say at least 8x), and then average together those 8 samples to get back to the original sample rate. Then play that without a reconstruction filter. This would actually be a much better approach than just simply thinking that the value you play is the average of the sample outright

Digital audio was simply not designed to work like this but some argue that actually the low-pass filter is not needed and omitting this obligatory filter can actually deliver a more honest presentation of the music.

It measurably can't. You can find it more pleasurable, but that does not mean it's objectively a better representation. Those are two different things.

 

[tvrgeek]

Measurements correctly done are objective measures of only those things measured. They may or may not indicate if a listener likes the result or not and the standard distortion and noise tests are far from compressive. Listening is subjective. There is no such thing as objective listening. Even blind A/B testing still is a matter of how your brain perceives things.

For my ears, every newer DAC I have heard has an ugly harshness to violins, trumpets and female voices. Will a given implementation of an R2R or a tube buffer provide high enough even order distortion to hide it? That is how some are designed.

The fact that you can hear a difference between two DACs, then at least one is wrong. That is a separate question from which sounds best to you. If both sound and measure identically, it still does not prove both are not wrong.

Something not often mentioned is the accumulation of distortions. Every step, from mic to your speakers adds. Even if a given box is cleaner than one could detect on it's own doe it just add enough to raise the accumulation to objectional levels?

 

[BDWoody]

This results in horrible measured performance but within the trash, the actual audio signal is less manipulated and somehow better representative of the original data.

Uh huh...

 

[voodooless]

For my ears, every newer DAC I have heard has an ugly harshness to violins, trumpets and female voices. Will a given implementation of an R2R or a tube buffer provide high enough even order distortion to hide it? That is how some are designed.

Sounds like the DAC is not the problem, but rather the speakers and/or your room is.