"...never see any benefit..." is what you said. That means not any, none. Most PLL dac clock recovery is significantly jittery. AK4118 is one example. It needs to be followed by ASRC to clean it up.
Look, I get you are 'book smart.' As far as knowing your way around real world dac board design you still have a lot of practical knowledge to gain. It takes years to get really good at it. Some of the people who frequent this part of the forum know a lot more than you do, and I don't see any of them coming out to defend what you say about PLLs in reference to your first quote above.
Moreover, didn't you recently recommend some opamp like LME49720 or LM4562 for use in a dac without any qualification as to where it should or shouldn't be used? Did you know about their sensitivity to EMI/RFI coming out of modern dac? Didn't appear that way: " I'd go with the LM4562 (= LME49720) instead. "
https://www.diyaudio.com/community/threads/do-dacs-benefit-from-op-amp-upgrade.395957/post-7272157
Again, you have a ways to go. There are people here who would be happy to help speed up the process if you didn't come off like you already know it all.
Last edited:
Article on music and memory: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2776393/
IMHO memorization of the sound of some reproduction aberration in recorded music is IME pretty the same thing as memorizing the music itself, its one part of the sound that is encoded into memory.
IMHO memorization of the sound of some reproduction aberration in recorded music is IME pretty the same thing as memorizing the music itself, its one part of the sound that is encoded into memory.
Actually I haven't been reading any of Benchmark's marketing fluff. I base my opinion on my knowledge of jitter cleaners from my industry experience designing them for a major semiconductor manufacturer.
I don't know. That's a question better answered by Bruno. There's tons of examples of "look! we did something special" in audio. That doesn't mean whatever special sauce is rooted in science or even in a perceived difference. It is often reflected in the price tag, however.
Please read what I write. I said you would not see any benefit. Not that jitter would be eliminated. If the DAC clock jitter is identical with the fancy clock oscillator versus the clock from the SPDIF receiver - as it would be if it goes through a clock jitter cleaner for reasons explained above - I don't see how you would get any benefit from the cleaner clock, except that the cleaner clock buys you a bullet point on the marketing slide.
Sorry that my education level is intimidating you. I enjoy learning and plan to keep doing that for as long as I live.
Perhaps you could use your smarts to explain how an ASRC cleans up any jitter in the PLL DAC clock recovery from an AK4118. What does an ASRC do that a simple D flip-flop reclocker provided with a clean clock couldn't do? Besides, I'm talking about the jitter cleaner that's commonly part of a DAC. Not about the CDR in an SPDIF receiver like the AK4118.
It should be obvious to anyone reading that thread that I suggest using the LM4562/LME49720 in place of the NE5532. That, of course, doesn't prevent someone from taking that out of the context in which it was intended. And, yes. I'm keenly aware of the inner workings of the LM4562/LME49720. I use it a lot and do know how to read documentation.
Tom
Last edited:
LOL. That's hilarious. Did you even read the abstract? In the abstract author claims that music might aid in the encoding of memories associated with the music. They then spend the majority of the review article talking about how music aids in memory recall in particular of autobiographic memories. I fail to see what that has to do with ABX testing in audio.
You can make similar claims about scents and smells. A particular smell often triggers memories. That would be just as relevant as above article in your case.
Tom
It isn't. It's just that I think you are sometimes speaking more from theory than experience. It takes enough of both.
It was already explained well by werewolf. If you had started reading the thread then a few posts down you would have seen his credentials.
Basically, it has to do with what the PLL is used for in a SPDIF receiver is different from what its used for in an ASRC. In former case the PLL is used for frame clock recovery (or else MCLK recovery from noisy embedded frame clock information), whereas in the second case the PLL is used to estimate the correct polyphase filter coefficients for resampling. Two very different applications. As a practical matter given the same incoming jitter, the loop corner frequency can be a lot lower in the ASRC case and remain in lock.
I will stop here for today.
Apologies to come in with a high level/newbie question.
But in relation to PC supplied noise in powering a USB device, doesn't the USB protocol require a checksum on packets/frames transmitted? and if the noise was either corrupting the frame/packet due to supplied noise (from PC) or induced noise (from surrounding equipment getting past any cable shielding), then as long as the receiver applies the checksum, frames can be reliably discarded and resent? (I'm assuming some sort of sliding window protocol).
Therefore cable quality would only matter where there were gross "drop outs" of frames. That is - the evidence would be buffer underruns and gaps in playback, none of the subjective "sounds cleaner / better treble etc..." claims?
But in relation to PC supplied noise in powering a USB device, doesn't the USB protocol require a checksum on packets/frames transmitted? and if the noise was either corrupting the frame/packet due to supplied noise (from PC) or induced noise (from surrounding equipment getting past any cable shielding), then as long as the receiver applies the checksum, frames can be reliably discarded and resent? (I'm assuming some sort of sliding window protocol).
Therefore cable quality would only matter where there were gross "drop outs" of frames. That is - the evidence would be buffer underruns and gaps in playback, none of the subjective "sounds cleaner / better treble etc..." claims?
Hi Dave,
That is fairly common question. Basically its true if problem is as you describe then the resulting audible symptom would probably be as you describe. So the question may then become, is the problem that the person listening is imagining something, or else could there be some other physical issue with the USB connection interacting with the dac in some way to produce a real audible symptom?
To understand the latter possibility requires knowing a little more about how a dac works. A lot of what goes on in a dac is analog more than digital. In particular, a dac needs an analog time reference and an analog voltage reference. It should probably be explained at this point that the time reference, a clock, is more analog than digital. That's because clock edge timing is affected by analog phase noise, and of that noise some of it initially starts as amplitude noise, then more or less gets mode converted.
Anyway, the problem with USB powered dacs usually has to do with noise associated with the USB connection. It could be related to power, ground, and or data lines. What connects most directly to the analog parts of the dac are the USB power and ground lines. They are typically noisy, so the question is how much immunity does the dac have to noise. The short answer is improving noise immunity often involves increased cost and complexity of design.
Another problem which can occur is related to USB cable wire resistance or wire impedance, along with varying dac load current, that then modulates the USB power voltage that analog parts of the dac can be sensitive to.
The above examples are the sorts of things might plausibly cause a true audible symptom, is all.
Does that help at all or are there more questions?
Mark
That is fairly common question. Basically its true if problem is as you describe then the resulting audible symptom would probably be as you describe. So the question may then become, is the problem that the person listening is imagining something, or else could there be some other physical issue with the USB connection interacting with the dac in some way to produce a real audible symptom?
To understand the latter possibility requires knowing a little more about how a dac works. A lot of what goes on in a dac is analog more than digital. In particular, a dac needs an analog time reference and an analog voltage reference. It should probably be explained at this point that the time reference, a clock, is more analog than digital. That's because clock edge timing is affected by analog phase noise, and of that noise some of it initially starts as amplitude noise, then more or less gets mode converted.
Anyway, the problem with USB powered dacs usually has to do with noise associated with the USB connection. It could be related to power, ground, and or data lines. What connects most directly to the analog parts of the dac are the USB power and ground lines. They are typically noisy, so the question is how much immunity does the dac have to noise. The short answer is improving noise immunity often involves increased cost and complexity of design.
Another problem which can occur is related to USB cable wire resistance or wire impedance, along with varying dac load current, that then modulates the USB power voltage that analog parts of the dac can be sensitive to.
The above examples are the sorts of things might plausibly cause a true audible symptom, is all.
Does that help at all or are there more questions?
Mark
Last edited:
