So you don’t understand THD either. Much studying needed before you can talk sensibly.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
DAC types and their sonic signature
- Thread starter Sparky
- Start date
Maybe that's the best way to put it...because I am the exact stupid person you mention.
My hobby is listening to music. I prefer to buy hardware that allows me to come the closest to the originally recorded input as possible.
For others, the hobby seems to be the endless pursuit of gear, that may or may not render a signal that comes as close as possible to the original. You're then buying someone else's subjective preference...who didn't make the music, they made an assembly of electronic components that adds whatever form of coloration that they happen to like. I would much prefer to have what goes into my amplifier be as close to what got put on the original recording as possible. You know...like...'high fidelity.'
I used to be part of that cult. Fortunately, I actually did some blind listening before I wasted too much money, and was shocked to find I just couldn't hear a difference once I couldn't see the source.
Oh, and for those who think it's because of what must be crap gear, that's not part of the equation...
Do yourself a favor and at least try. Could save a soul a ton of money...for those who care about such mundane things...
Aha! A measurements purist, I see.
Obviously, most of us do not want actually recreate what you would hear live - standing next to the photographer at a recording session like the one below. Nor do we want to hear the tracks as recorded.
Rather I want my stereo system to accurately reproduce the final product that the mixing/mastering engineers created by manipulating the "originally recorded input." That final mix/master is what gets sold to buyers of recorded music, although the mastering engineer is likely the one who takes the final mixdown and decides on those dreaded compression levels that can reduce dynamic range and squeeze the life out of some recordings. Those stages are where the actual "creation" of what we loosely call a "recording," and can take place many years after the actual recording sessions if the raw tracks are available for re-mixing and mastering.
I am also aware that some music lovers who are audiophiles have their favorite mixing/mastering engineers, just like film buffs have their favorite directors.
(I'm not an expert on this, so if any of our resident engineers would like to correct or embellish my ramblings, please do so - and my learning about where my "recordings" come from will continue.)
good riddance to him.
- Joined
- Feb 23, 2016
- Messages
- 20,690
- Likes
- 37,415
Sorry, no can do because he isn't wrong. You can't correct someone who is correct.
Yes, That's a much better way to say it.
Whoever had the final say in the recording/engineering process, I want what HE signed off on when that final master was deemed finished.
I'm not particularly interested in how some DAC designer looking for the best sounding transistor likes his distortion, and whether or not he can convince someone else that his additions to the signal as originally signed off on by...say George Martin... REALLY open up the soundstage. I don't know if there's much I could be attracted to less.
Just give me the music...I'll take it from there.
I'd say I'm a threshold measurement purist. If I can't hear it, and am never going to hear it, I can't be convinced I have to care more about it. But, paying EXTRA for added distortion? Pass...
Prove that signal (below 20kHz) amplitude has stabilized due to applying a LPF above 22 kHz or so (aside of anti-sinc roll-off correction which obviously intentionally corrects the upper end, and aside of LPF's added roll off).
In other words, prove the existence of varying signal amplitude because of reconstruction waveform.
You can't write crazy things.
Last edited:
manisandher
Addicted to Fun and Learning
Yep. I upsample 44.1 content 16x in HQPlayer or XXHighEnd before sending to my NOS filterless DAC.
Last edited:
solderdude
Grand Contributor
I can, if I have to, but you surely do.
That proof has already been delivered with 10kHz and 16kHz plots which are both below 20kHz.
You just don't believe it.
I have another frequency for you that doesn't have a varying amplitude (you see, I CAN write crazy things) ...
20kHz (44.1kHz sampling frequency) filterless R2R DAC signal output.
Oh darn ... the 20kHz varies in amplitude, about 6dB no less so this is 3dB drop-off 'averaged' over time.
Further more I already specifically told you that only the higher frequencies show a roll-off due to amplitude variations and not due linear filter roll-off (which behaves differently as does the hearing)
Frequencies below 5kHz (due to the available amount of samples in 1 period) will not have any amplitude variations.
That's why 4 times oversampling (4x 5kHz = 20kHz) is already enough for the roll off to not be there.
Below a 20kHz properly up-sampled to 176kHz sample rate (= 4x) do note that 5kHz at 44.1kHz will look (and be) exactly the same
You see.. no roll-off no amplitude variations, no problems ... for lower frequencies.
and finally 440Hz with 44.1kHz sample frequency ... nice no problems ... no roll-off and as mentioned before enough sample values per period.
Perfectly usable ...
I hope you finally get it now... the rest of 'us' already did,
With NON filtered DACs it is all about the number of available samples per period of the frequency that is to be reproduced.
Last edited:
The potential trouble with upsampling is that it is a processing, hence it can induce the error as well, change the sound on its own. So it's not only about benefits. Besides, DACs internal architecture must support filtering at a higher frequency threshold to allow its benefits.
Flailing.
I find this discussion quite interesting because it made me realize I know less about DAC output stages that I thought I did.
Just to confirm I understand this correctly: my intuition tells me the reason why the 20 kHz tone varies in amplitude is because it ends up beating with its image frequency component at 24.1 kHz. If we use a proper reconstruction filter (LPF) that removes that above-Nyquist 24.1 kHz component, that component is gone and therefore the beating is gone. Is my understanding correct?
solderdude
Grand Contributor
The question that follows is what one considers an output stage.
Is it the output buffer circuit ?
Is it an I/V converter ?
Is it a post filter ?
meant to attenuate for which frequencies ?
Is it a combination ?
Do you mean stage as in a certain 'block' in an DA chain ?
Is it the output buffer circuit ?
Is it an I/V converter ?
Is it a post filter ?
meant to attenuate for which frequencies ?
Is it a combination ?
Do you mean stage as in a certain 'block' in an DA chain ?
And then there's the part where those three are often combined in some fashion.
Been reading this thread and since you seem to be an expert on ADC/DAC and I am definitely not would you say that it would be recommended to have a NOS R2R DAC in NOS mode feeding such DAC with pre processed DSD (or PCM) content on external application (aka HQP)? I am changing the input "digital" form with pre processsing and would like to confirm if I'm doing the right thing
Also if this DAC accepts for example DSD256 and it is fed with DSD128 even though in NOS mode will it still most likely apply "well designed filters" as part of the internal conversion (I understand the answer could be IT DEPENDS ON THE DAC) but just curious. This is a fascinating subject
solderdude
Grand Contributor
Hardly, it is just electronics, both AD and DA conversion. Electronics is what I know something (not everything) about. Not a math guy either and smart guys doing all the filter stuff (should) know their math and what matters when it concerns audio.
There are many many ways to achieve the same goal, Digital to Analog conversion.
And even when using one 'technique' there is more than one way to skin a flerken*.
Some are technically better (in the sense of reproducing whatever they should) some less so.
Some have audible consequences some not (so much).
I think it is not easy to separate the wheat from the chaff when one does not know what filter does what and has which drawbacks.
Also it is hard to estimate what the audible consequences are.
People like Zalive don't even care about it.. they listen and when they like it simply claim it is better.
Nothing wrong with that except that when something SOUNDS better to them it doesn't mean it is 'technically better'.
You can 'safely' use a filterless R2R DAC when you are using HQP while set to up-sample (using its excellent filters) to at least 176/192 or even higher if the DAC supports it. This will move the crap all the way up where speakers/headphones and ears don't even do anything any more.
It should have a relatively clean area between 24kHz and 100kHz.
It would still be wise to (not steeply) filter above say ... 50kHz or so to ensure the HF crap that remains will be attenuated enough not to (possibly) have some negative effects with some rare amps that don't like to reproduce that.
When someone likes the sound of a NOS filterless R2R DAC that's fine too, they just should not claim it is technically a better solution because it lacks the (always sound degrading ?) filter.
When an R2R DAC device also accepts DSD it probably has a different DAC 'section' for it. I don't think you can toggle the ladder bits that fast...
R2R DACs are for converting PCM and are usually pretty limited in max. switching frequency so to reproduce DSD one would have to convert to PCM first.
In DA conversion you just need to use the right filters. Some of the 'special' filters usually can produce nice square-waves and needle pulses but these are only test signals for tech guys... they do not exist in music nor in any recording nor is it an advantage of a DAC design if it can do this.
* (go see captain Marvel
)No. You still don't understand what those plots actually show. I'll explain (again) below.
Oh darn ... the 20kHz varies in amplitude, about 6dB no less so this is 3dB drop-off 'averaged' over time.
Do you realize this is not a graph of a 20kHz signal, but instead a graph of a complex waveform consisting of more than one frequency?
Why is in bold a false statement? Because you don't see a 20kHz waveform on the graph. And you can't actually tell the amplitude on the 20kHz part of the waveform on graph either. Because it's mixed with the rest of frequencies in single graph from which you can't visually separate a 20kHz from the rest.
But what we do know? Physics. Well, at least I know it
and physics tells the continuous tone of 20kHz, which is a fundamental tone in this graph, well...won't have its amplitude changed because of presence of the rest of frequencies (which are higher than a 20kHz tone). Because amplitudes of various frequencies don't add.@DonH56 : can amplitudes of waves of various frequencies add or subtract in any manner? Or is this something that can happen exclusively if their frequency is the same?
No, I really can't advocate listening of any particular sound to anyone. I advocate to everyone listening to the sound which they personally prefer the most, regardless of what specs show. Because if you like it the most, specs should not matter.
If a DBT is someone's personal way to determine which sound he likes the most...so be it.
Using dithering, I'm following
Agreed and that is what most people doesn't know, I didn't some time ago but I suspected it was still filtered
You are correct, it is my understanding (I'm not an expert) there is a stage for DSD conversion (seen 6 bits) and another separate stage / path for PCM conversion 24 bits I have seen) and yes resistor ladder with the R2R designation for the blocks they use to convert on the output are mainly used for PCM, IMO if you are feeding DSD you are better off using "DSD DIRECT" DS DAC
IME you need also the right modulators
go see HQP 4.11 for the EC (extended compensation mods) filters IMO are negligible at least on my system.I have seen the "flerken" did not know you can skin them (who would know ;-)
Pretty funny reference.
@solderdude I appreciate your candid answer, as it confirms kind of what I was thinking already, thank you for sharing.
Similar threads
