Had developed at least 5 different types of DAC in the past few month, from old TDA1305, TDA1541 to most up to date 32bit 384k XMOS. The specifications of these systems has huge difference. The most expensive system is at least 5 times of the cheapest ($105 vs $22). But to be honest, do they really sound that different?
We made side to side comparing with all those systems. The XMOS is really better than TDA1541 and TDA1305, especially in mid range. But is it really because of the digital performance of the system? The capacitors and other components in xmos are much better than TDA1541 system. We tried to upgrade the capacitors and power system of TDA1541. The sound quality was really improved.
So, which factors will affect the sound quality of a digital source according to your experiences?
We made side to side comparing with all those systems. The XMOS is really better than TDA1541 and TDA1305, especially in mid range. But is it really because of the digital performance of the system? The capacitors and other components in xmos are much better than TDA1541 system. We tried to upgrade the capacitors and power system of TDA1541. The sound quality was really improved.
So, which factors will affect the sound quality of a digital source according to your experiences?
To my knowledge, XMOS only make processors not DACs so your post is a bit confusing. You're using XMOS as a USB -> I2S interface?
To answer your question - lowest noise on the audio signal at the DAC output sounds best. So TDA1541, being a bipolar multibit DAC will have lower noise than TDA1305 which is a higher operating frequency CMOS-process DAC chip.
To answer your question - lowest noise on the audio signal at the DAC output sounds best. So TDA1541, being a bipolar multibit DAC will have lower noise than TDA1305 which is a higher operating frequency CMOS-process DAC chip.
sorry the system is XMOS+PCM1794...We are used to name it "xmos board" in the company.
My question in simpler words: does sampling rate (above 44.1KHz) and word length (>16bits> really make sense to sound quality? I feel a good PIO capacitor can do much better job than 24 bits.
About the noise, actually non of those systems has audible noise with headphone amp. The noise of the amplifier system is usually much larger than DAC.
My question in simpler words: does sampling rate (above 44.1KHz) and word length (>16bits> really make sense to sound quality? I feel a good PIO capacitor can do much better job than 24 bits.
About the noise, actually non of those systems has audible noise with headphone amp. The noise of the amplifier system is usually much larger than DAC.
Last edited:
I am very surprised then that you have found PCM1794 sounding better than TDA1541. To my ears even the lowly TDA1543 sounds better than PCM1794 - more accurate instrumental timbres just to mention one way its better.
In my experience with multibit DACs (primarily TDA1545), increasing the sample rate decreases the SQ. I reckon this is due to there being more glitches the higher the update rate, along with settling time being a greater proportion of the total when running faster.
On noise, I wasn't talking about audible noise, rather noise which masks fine details (especially hall acoustics) in a recording. At least I reckon that's what makes the difference, I don't have any proof. Whether an amp's noise exceeds a DAC's - it depends on the DAC and amp in question. I reckon my modified Xindak amp is adding more noise than my DAC, but the DAC I have is so highly optimized now its hard to find an amp which'll match it.
In my experience with multibit DACs (primarily TDA1545), increasing the sample rate decreases the SQ. I reckon this is due to there being more glitches the higher the update rate, along with settling time being a greater proportion of the total when running faster.
On noise, I wasn't talking about audible noise, rather noise which masks fine details (especially hall acoustics) in a recording. At least I reckon that's what makes the difference, I don't have any proof. Whether an amp's noise exceeds a DAC's - it depends on the DAC and amp in question. I reckon my modified Xindak amp is adding more noise than my DAC, but the DAC I have is so highly optimized now its hard to find an amp which'll match it.
Dan Lavry has a few good papers on sampling rates and recording: Lavry Engineering
But your questions is very unspecific: do you mean higher sampling rates for the recording or running the dac chips as high as you can, through oversampling ? The last case at least makes your life easier when it comes to analog filtering. Some dac chips also have varying performances depending on the sample rates and bit depth (it's quite detailed in the pcm1794 datasheet). But you probably knew that already ?
Anyway, a lot of the "sound" of a DAC is the analog stage. What are you using as I/V and filtering stage ? Something similar for all your versions ?
Btw, the pcm1794 is really a great chip (even the lower grade pcm1798). I'd pick it over the tda1543 any day.
But your questions is very unspecific: do you mean higher sampling rates for the recording or running the dac chips as high as you can, through oversampling ? The last case at least makes your life easier when it comes to analog filtering. Some dac chips also have varying performances depending on the sample rates and bit depth (it's quite detailed in the pcm1794 datasheet). But you probably knew that already ?
Anyway, a lot of the "sound" of a DAC is the analog stage. What are you using as I/V and filtering stage ? Something similar for all your versions ?
Btw, the pcm1794 is really a great chip (even the lower grade pcm1798). I'd pick it over the tda1543 any day.
Certainly PSU and output stages are hugely important to the SQ - yet if you start out with in-band noise modulation from the DAC itself, there's nowt a decent analog stage can do to fix that up. Against the laws of physics...😛
- Status
- Not open for further replies.