Looking in the data AK4499 data sheet and search for the term 'THD', I find the table attached below. There is no specified minimum THD, there is description of how THD was determined, and just a blank for DSD512.
There are some graphs for PCM modes in the eval board manual, but didn't see any for DSD.
Even if DSD512 THD+N @1kHz did happen to look worse than for DSD256, that's still far from the whole story. A lot of what comes out of dacs does not show up nicely on typical FFTs IME. Too much RF making some distortion only partially stationary relative to FFT acquisition timing. That's my opinion for now, as always subject to revision as new information becomes available.
No. Not what I said.
No need to resort to straw man arguments.
Regarding your 'appeal to authority type arguments' in knocking down your straw man, please see: Appeal to Authority
Finally, my point that seemed to go right over your head was that nothing is recorded at native DSD256 or DSD512. There is always DSP involved in upsampling, conversion, whatever. If an particular upsampling and DSD conversion algorithm sounds better at DSD512, then that's part of what is affecting what comes out of the dac, not just 'typical 1kHz THD+N.' In some cases DSD512 has sounded better than DSD256, and in other cases it has been exactly the other way around. One should listen carefully and not jump to conclusions either way.
Attachments
Last edited:
Attachments
Interesting, my datasheet is named: ak4499_p02e-1.pdf
Just found a newer one online named: AK4499_Feb2019.pdf
Just found a newer one online named: AK4499_Feb2019.pdf
Last edited:
DAC blind test: NO audible difference whatsoever
DAC blind test: NO audible difference whatsoever
(UPDATES BELOW)
That blind test was long due, it's now done.
Started with a 30$ unit (Fiio) against a 3000$ one (Forssell) and once SPL-matched (massive gain difference), no one could tell the difference in a ABX test.
Then, we switched to a different set-up, using a pair of B&W CM9 speakers and the Forssell against a Eximus DP1 (3500$ or so). Same result: impossible to spot them in a ABX.
We were only 4 participants, but regardless it didn't feel like day & night difference to start with... ''Eyes opened'' we FELT differences, but couldn't prove it in the ABX.
Cables, amplifiers, Lossy v.s. Lossless/HD, EQ'd mid drivers, DAC... Nope. Nothing is passing a ABX blind test.
I'm pretty sure, now, that the human auditory capacities are very, very, overestimated.
The good news is: we can probably save a LOT of money.
given that modern DACs are now highly evolved with extremely high performance I see no reason why they should sound differently unless implementation issues have not been addressed, or as is common in audio, we simply prefer a little flavour and sometimes an old wine is best
I plan to save a lot of time and money myself
DAC blind test: NO audible difference whatsoever
That blind test was long due, it's now done.
Started with a 30$ unit (Fiio) against a 3000$ one (Forssell) and once SPL-matched (massive gain difference), no one could tell the difference in a ABX test.
Then, we switched to a different set-up, using a pair of B&W CM9 speakers and the Forssell against a Eximus DP1 (3500$ or so). Same result: impossible to spot them in a ABX.
We were only 4 participants, but regardless it didn't feel like day & night difference to start with... ''Eyes opened'' we FELT differences, but couldn't prove it in the ABX.
Cables, amplifiers, Lossy v.s. Lossless/HD, EQ'd mid drivers, DAC... Nope. Nothing is passing a ABX blind test.
I'm pretty sure, now, that the human auditory capacities are very, very, overestimated.
The good news is: we can probably save a LOT of money.
given that modern DACs are now highly evolved with extremely high performance I see no reason why they should sound differently unless implementation issues have not been addressed, or as is common in audio, we simply prefer a little flavour and sometimes an old wine is best
I plan to save a lot of time and money myself
Okay, so now the truth comes out.
Did you receive any training in ABX? If so, what did it consist of?
Also, did you really buy a D90?
If so, and if you are convinced all dacs sound the same, why would you do that?
Last edited:
Hahaha
If you actually believe that why would you be wasting your time reading this thread about a state of the art DAC chip which sounds the same as any other one? And please do not bother trying to bait us further, been there done that and will not respond again.
DAC blind test: NO audible difference whatsoever
DAC blind test: NO audible difference whatsoever
(UPDATES BELOW)
That blind test was long due, it's now done.
Started with a 30$ unit (Fiio) against a 3000$ one (Forssell) and once SPL-matched (massive gain difference), no one could tell the difference in a ABX test.
Then, we switched to a different set-up, using a pair of B&W CM9 speakers and the Forssell against a Eximus DP1 (3500$ or so). Same result: impossible to spot them in a ABX.
We were only 4 participants, but regardless it didn't feel like day & night difference to start with... ''Eyes opened'' we FELT differences, but couldn't prove it in the ABX.
Cables, amplifiers, Lossy v.s. Lossless/HD, EQ'd mid drivers, DAC... Nope. Nothing is passing a ABX blind test.
I'm pretty sure, now, that the human auditory capacities are very, very, overestimated.
The good news is: we can probably save a LOT of money.
given that modern DACs are now highly evolved with extremely high performance I see no reason why they should sound differently unless implementation issues have not been addressed, or as is common in audio, we simply prefer a little flavour and sometimes an old wine is best
I plan to save a lot of time and money myself
If you actually believe that why would you be wasting your time reading this thread about a state of the art DAC chip which sounds the same as any other one? And please do not bother trying to bait us further, been there done that and will not respond again.
So a properly carried out blind test with a solid sample size doesn't provide one with any knowledge and is purely based on imagination? Not saying this one was carried out properly, but since you generalised it, I will too.
Also, I assume using objective measurements in combination with sighted listening test to have measurements fuel subjective opinions is the path of enlightenment then? I am sorry, but you must be joking.
Also, I assume using objective measurements in combination with sighted listening test to have measurements fuel subjective opinions is the path of enlightenment then? I am sorry, but you must be joking.
Erm.... if my memory serves me correctly, didn't this thread split to accommodate this discussion and to avoid diluting experiences and observations of the AK4499?
Simon is correct IIRC. People who want to discuss whether or not new dac designs are still needed were directed to take that discussion to: If it's purely an engineering challenge why bother designing yet another DAC?
That (in bold blue) has been the case for at least 20 years.DAC blind test: NO audible difference whatsoever
DAC blind test: NO audible difference whatsoever
(UPDATES BELOW)
That blind test was long due, it's now done.
Started with a 30$ unit (Fiio) against a 3000$ one (Forssell) and once SPL-matched (massive gain difference), no one could tell the difference in a ABX test.
Then, we switched to a different set-up, using a pair of B&W CM9 speakers and the Forssell against a Eximus DP1 (3500$ or so). Same result: impossible to spot them in a ABX.
We were only 4 participants, but regardless it didn't feel like day & night difference to start with... ''Eyes opened'' we FELT differences, but couldn't prove it in the ABX.
Cables, amplifiers, Lossy v.s. Lossless/HD, EQ'd mid drivers, DAC... Nope. Nothing is passing a ABX blind test.
I'm pretty sure, now, that the human auditory capacities are very, very, overestimated.
The good news is: we can probably save a LOT of money.
given that modern DACs are now highly evolved with extremely high performance I see no reason why they should sound differently unless implementation issues have not been addressed, or as is common in audio, we simply prefer a little flavour and sometimes an old wine is best
I plan to save a lot of time and money myself
Returning for a moment, if I may, to the new Chinese USB boards with Accusilicon clocks: Modded my USB board to use clean +5v power rather than USB power. The mod is almost exactly the same as some people use with Amanero USB boards.
Anyway, hooked it up as master clock source for AK4499 eval board to see how it sounds. Not very good, at first that is. Left it running overnight and it smoothed out and cleaned up a lot. So much better, in fact, that I set up the Topping D90 next the eval board so I could do some quick A/B comparisons. Sound difference between the two dacs is now quite small. Sounds to me like it can be accounted for almost entirely by differences in the output stages: there is a little tiny bit of RF leakage coming out of the eval board after the passive post-I/V filters. On the other hand, D90 has more opamps and what look to be some solid state mute switches at the analog outputs after the last opamps. Surprising result to me that the two come out so similar. Neither JL Sounds clocks or selected NDK SDA clocks I tried sounded to good as Accusilicon.
Also, don't think there is any question now that the clocks on the USB board I received are real. Don't know if all the Accusilicon clocks selling on Tabao and Aliexpress are real too. Someone will have to try them and find out, I guess.
Anyway, hooked it up as master clock source for AK4499 eval board to see how it sounds. Not very good, at first that is. Left it running overnight and it smoothed out and cleaned up a lot. So much better, in fact, that I set up the Topping D90 next the eval board so I could do some quick A/B comparisons. Sound difference between the two dacs is now quite small. Sounds to me like it can be accounted for almost entirely by differences in the output stages: there is a little tiny bit of RF leakage coming out of the eval board after the passive post-I/V filters. On the other hand, D90 has more opamps and what look to be some solid state mute switches at the analog outputs after the last opamps. Surprising result to me that the two come out so similar. Neither JL Sounds clocks or selected NDK SDA clocks I tried sounded to good as Accusilicon.
Also, don't think there is any question now that the clocks on the USB board I received are real. Don't know if all the Accusilicon clocks selling on Tabao and Aliexpress are real too. Someone will have to try them and find out, I guess.
Markw4: This is very interesting news, Mark. I am also looking for a 1.536 MHz/DSDD1024 USB-to-I2S so hearing that they are actually quite good as they are is good news. Hmmm... I remember reading in one of the threads that you had some reservations regarding ASIO and and another driver option. If I remember correctly: Can I ask you if you did solve this (I'd prefer using Win7 or Win10 with the digital interface)?
Also, I reckon there are a couple 1.536 MHz USB boards available by now - I've found this one ... Is that the one you have tried?
USB digital interface AS318B PCM1536 DSD1024 compatible with Amanero Italy XMOS | eBay
Cheers & enjoy listening ;-)
Jesper
Also, I reckon there are a couple 1.536 MHz USB boards available by now - I've found this one ... Is that the one you have tried?
USB digital interface AS318B PCM1536 DSD1024 compatible with Amanero Italy XMOS | eBay
Cheers & enjoy listening ;-)
Jesper
Last edited:
Hi Jesper, Yes that looks like it.
The driver package with manuals seems to exist in possibly more than one version. The (apparent) manufacturer has a package for download at their website. Inside are some PDFs: a user manual and driver setup instructions (google translate can be your friend.) Support – XING
Also, I trust you are familiar with the Amanero clean power mod which involves removing L1 near the USB connector, and bringing in +5v power there. In this case L1 is actually an 0R resistor. There are some holes for a 2-pin header to be soldered in so as to allow quick enable/disable of USB power. I soldered another 2-pin header on the bottom side of the board, one pin of it to the load side of top pin header, and the other bottom side pin to the adjacent ground plane. Hope that makes sense to interested parties, if any.
Last edited:
diiyralf, Measurements are fine as far as they go, but they don't inform one as to exactly how something sounds. Even Amir over at ASR has started to figure that out. Notice how people over there are now allowed to talk about sound quality aside from measurements?
And must add: I value your testing, but maybe I will wait for some measured results (Like in: somebody with a Timepod) for the differences, characteristics of the Accusilicon clocks..
They could be good, but datasheet comparisons do not point to a clear advantage.
Like a 22MHz NDK clock, referred to a best value 45Mhz Accusilicon clock. (they are not available less than that).
Ciao! G
They could be good, but datasheet comparisons do not point to a clear advantage.
Like a 22MHz NDK clock, referred to a best value 45Mhz Accusilicon clock. (they are not available less than that).
Ciao! G