It is being reported by DAC reviewers that the I2S input of a Topping D90III Discrete DAC sounds more relaxed and 3 dimensional than when fed through its USB input. My question is, how is this possible?
The jitter performance of the DAC is -157db whereas the same parameter is specified as 75 picoseconds. I have no idea how the jitter performance expressed in decibels vs time in picoseconds is comparable. But my assumption is that the clock in the DAC is better in specification and implementation. So how can USB input, I2S output of a DDC sound better than the USB of DAC? Thanks in advance.
The jitter performance of the DAC is -157db whereas the same parameter is specified as 75 picoseconds. I have no idea how the jitter performance expressed in decibels vs time in picoseconds is comparable. But my assumption is that the clock in the DAC is better in specification and implementation. So how can USB input, I2S output of a DDC sound better than the USB of DAC? Thanks in advance.
I don't think even the creators of it know the answer to your question. 🙄
Also, those data from the measurements, like this parameter is -130dB and this one is -160db, it has nothing to do with how it sounds, it's just fooling around with numbers.
Also, those data from the measurements, like this parameter is -130dB and this one is -160db, it has nothing to do with how it sounds, it's just fooling around with numbers.
Thanks NIXIE62 for your answer. But if the DAC is using I2S protocol internally and it has better jitter performance than the DDC under reference, should not the DAC fed directly through USB sound better than USB - I2S via DDC?
Unless there is something wrong with USB input of Topping D90III there should not be an audible difference. Expectation bias is a strong possibility especially if the reviewer has been brainwashed by audiophiles.
A good example of such is this review of a device that in reality does not make any difference to audio:
https://www.hifipig.com/innuos-phoenix-usb/#more-174377
A good example of such is this review of a device that in reality does not make any difference to audio:
https://www.hifipig.com/innuos-phoenix-usb/#more-174377
Its not impossible although its not common. Shouldn't be that way.
However, seems to me that all digital input to this dac has to go through sample rate conversion since there is only one clock on the dac daughterboard. Could be the clocks being used to send I2S into the dac are pretty good. Could also be the clocks used inside the dac for USB are not as good. In that case its possible that jitter differences between the clock sources are affecting the sample rate conversion. Another possibility is that radiated EMI/RFI noise from the USB circuitry is causing problems that don't exist when the I2S input is used. Of course whatever may account for the reported differences, it isn't necessarily true that bias on the part of listeners is the only plausible explanation.
EDIT: Regarding jitter performance numbers, it depends on how they are measured. If measured with J-test and if the jitter is random noise jitter, then it will not show up very well in typical jitter measurements. Therefore, I would discount the jitter numbers as being non-conclusive in this case.
Last edited:
Quite so. But PCM has to be converted to DSD for this dac to work. Therefore there must be some SRC.
When I2S input is used then it has to be the clock master unless there is ASRC.
And there is the question of what the clock on the DAC daughterboard is for? I can't think of any purpose but to clock the FIRDAC output arrays. Maybe someone else has another idea?
When I2S input is used then it has to be the clock master unless there is ASRC.
And there is the question of what the clock on the DAC daughterboard is for? I can't think of any purpose but to clock the FIRDAC output arrays. Maybe someone else has another idea?
The PSRM module responsible for PCM-to-DSD conversion is used also for I2S input so it is unlikely that is the cause for these alleged audible differences. Pointless to speculate further about review based on subjective sighted listening.
Who will know what is happening there, those devices are very complex. If it sounds much worse over a USB connection, then they've screwed something up. The power supply of the circuits inside the device has a lot of influence on the sound. Subjectively, a lot depends on the cables and the source of digital signals. They are just throwing out some magic numbers trying to attract audiences and customers. I remember amplifiers from the golden age of HiFi, it was a race to have the lowest THD and highest damping factor. Those were their main selling points, and if they practically had little to do with the sound. Most of those devices sounded mediocre, regardless of the fantastic data from the datasheets. Such things are still common today. If a device has been tested more than once, and shows some obvious flaw, it should simply be skipped as an option.
Last edited: