Exactly. The information that is displayed in AS as the output should correspond to what the DAC effectively gets. It’s like that with all the players who display the output.
1 Like
You’re wrong about that. If you upsample it’s not bit-perfect. Same as when you apply software volume control or EQ.
RME are makers of audio equipment from Germany:
RME ADI 2 is great. I heard about a guy pbthal who made rips from famous vinyls using this dac.
You are messing between different things.
R-2R DACs function differently than the other DACs. They do not upsample indeed, but they represent 1% of the DACs in use, while all the DACs, that represent 99% of the Dacs in use, do upsample.
R-2R DACs are great, but expensive and cumbersome. That’s why there are so few of them in use.
So, I have no idea what RME says about bit-perfect. You just gave me a link to its site. If it pretends that only the R-2R DACs DACs are bit-perfect, and that the 99% in the world, who do upsample, are not, it’s more of a commercial argument than a reality.
RME is not R2R, it’s a vanilla sigma delta chip based DAC. If you upsample using any algorithm, you’re changing the sound if even so slightly. That’s why people do it.
The RME will give you a “pass” if it receives the sample file in unaltered form. This is before it internally upsamples. It does that with 16bit, 24bit and 32bit samples.
RME can pretend what it wants. It is a niche maker.
iFi, that builds DACs that do upsample guarantees them as being bit-perfect all along the chain of upsampling.
That’s impossible. They use some sort of algorithm internally, even if that’s the one provided by the DAC chip.
I tell you that 99% of the DACs in use do upsample by converting the PCM signal to hi frequency DSD signal. That’s what they do.
Already explained several times, software upsampling such as r8brain changes the sound, you even have access to parameters to fine tune the result. And if you don’t hear any difference when changing these parameters, then there’s a problem somewhere because they’re really obvious. And that’s why people want it, and that’s why you have access to parameters and different algorithms exist.
When you manually engage Audirvana’s upsampling, you don’t only change the bitrate, you also change the sound.
While upsampling methods are way beyond my understanding, I understand they are not just multiplying bits, they also modify the waveform somehow, introduce good and bad, and use recipes to mitigate issues (such as anti-aliasing).
The way 99% of the DACs function, there is an upsampling because it is necessary for a good digital to analogue conversion.
It doesn’t matter if this upsampling is done by the algorithm of the player and the CPU of the computer or by the cheap of the DAC with the algorithm of the DAC. What matters is which upsampling gives a better SQ. And you hear indeed a difference between them, but it’s not because the sound is not bit-perfect.
The upsampling of the sound is somehow analogue to the increase of a resolution of a picture. Better algorithms and increased computing power, will produce a picture with less aliasing that will look nicer.
It’s the same with the sound, but it’s done in real-time. A given player, may have a better algorithm than the one of the DAC, and may produce a better upsampling of the sound, because it has the advantage to compute its algorithm with the CPU of the computer that is way more powerful than the cheap of any DAC.
Exactly that, you claim that the Audirvna upsampling algorithms are better than DACs internal upsampling algorithms.
It’s true that DAC chips internally convert the PCM to DSD just before the analogue output. That’s why Sony designed the DSD in the way they did. It’s to make the DAC conversion simple. You just need to low-pass the DSD stream to get analog signal.
The main reason why DACs upsample is to apply filtering at higher frequencies and shift the noise beyond the audible spectrum.
Here is what Benchmark thinks about upsampling:
I never claimed that Audirvana’s upsampling algorithms are better than the ones of my DAC. I heard Damien claiming it in the early years of Audirvana regarding the DACs in general. It was the case indeed at the time, because the DACs were not as good as the DACs that we use today, and the DACs that performed well were very expensive.
I do not use AS’s upsampling modes, because I don’t hear a tangible benfit with my DAC. The only player that gives me a better sound with its upsampling is HQPlayer. I got a M1 Mac two days ago, and its powerful CPU allows HQPlayer to give me even better results than those thab the ones I was getting with my i7 MacBook Pro.
Now, if you want to know more about how DACs perform, read this article, which speaks also about this topic.
1 Like
I find the Ted Smith’s explanation quite comprehensive:
I watched the video, but it deviates the discussion. The topic was: does upsampling affect bit-perfect?
Regarding this video about DSD in DACs. Even only PCM capable DACs do convert PCM to DSD before the conversion of digital to analogue. That’s how all DACs work.
Most of the DACs do not have a great chip for the conversion to DSD. That’s one of the reasons why SACDs sound better than Hi Res PCM. It’s not because DSD is superior to PCM. It’s because the DAC does not have to make the conversion. it gets directly DSD that sounds better than if it was a PCM that it would have to convert.
Last thing, to have a really good sound from a SACD track, it’s better to re-modulate its DSD 64 to at least DSD 128. It makes unnecessary to filter DSD 64 high frequencies, because with DSD 128 they are out of the audible threshold.
That’s easy to answer, it does. It’s no longer bit-perfect once you upsample it. That’s a fact.
It’s not about having great chip to upsample to DSD, its just the way most chips work internally is similar to DSD (4-5 bit DSD to average out quantisation errors with noise shaping). Many DAC chips can’t even do native DSD, they convert to PCM and go through the same process as with PCM signal (Wolfson is an example).
That’s not really true. Already with DSD64 the high frequency energy is beyond the audible spectrum. Even if it’s not audible it can create issues with some electronics. That’s why it’s filtered out. With DSD128 you have even more high frequency energy.
The truth is that R-2R DACs give the best sound. But they are expensive and cumbersome, and only few of us, though we are audiophiles, use them.
All the other DACs, including costly high-end DACs, are using an inferior technology that is more cost-effective.
However, a descent DAC remains bit-perfect. It doesn’t add bits or doesn’t suppress bits from the source file.
I think that some DSD 64 high frequencies are filtered. With DSD 128 the frequencies are so high that there’s no point to filter anything.