That's a bold claim to make, if I was either Soren or Ian I would probably assume you're trying to sell something by trying to discredit the competition, a tactic that's somewhat slanderous and doesn't paint a very favorable picture of your motivations or your moral for that matter.
As I mentioned in a previous comment, which you conveniently choose to ignore: extraordinary claims require extraordinary evidence. If you think you have a superior approach you illustrate that with factual evidence, math, or measurements, not assumptions and hearsay. But if you lack that, which seems to be the case, your next best approach is to discredit others?
Take note that without a 1:1 comparison with your proposed solution on the basis of actual measurements you have nothing but your baseless claims and endless, poorly worded (also from a technical point of view) posts. All I can say at this point is buyer beware ...
In our case it's a little different, the FIFO and the DAC are separate boards so the DAC has no input selector and so on.
I will open a dedicated thread for the DAC as soon as we have completed the test.
The thread for the FIFO Lite is already open.
The Top versions are much more complex and it will take a long time to develop.
So, limited to the Lite version of the FIFO and the DAC:
Oversampling
The DAC itself does not provide oversampling, neither the FIFO lite, you have to oversample in software before feeding the FIFO if you want.
We use lower clock frequency as possible in our new devices to get the best phase noise from the clock, so there is no room to run the DAC at crazy speed. For this there are already a lot of PWM DAC on the market.
R2R
The DAC Lite has segmented architecture and uses sign magnitude notation.
The first 3 MSBs are thermometer decoded (3 to 7), the remaining 20 bit are classic R2R.
So 2 x 23 bit DAC in sign magnitude operation to get a 24 bit DAC.
Balanced Output
The DAC Lite provides single ended output only
Voltage Output
The DAC Lite is voltage output, around 2V rms and around 400 ohm output impedence.
I'm bulding a tube buffer if someone was interested.
Output selectable
No options for this, but you can use whatever you want after the DAC.
Digital Volume Control
Not provided, we have a different thought on this, sorry.
Input Selectors
The FIFO Lite provides 4 X I2S inputs, selectable by a optional user interface board.
For S/PDIF, USB and so on you have to use I2S converters, there are plenty of these devices on the market, although maybe we will provides USB to I2S and S/PDIF to I2S converter boards
What are your thoughts on the topic of filters ("brickwall")? NOS ... and?
I would be happy if you would share this thought with me 🙂
Yes - I think we have the same view on the the general aspects on how it works, but:
- No jitter kan stay alive if payload data is written to memory and clocked out by a different clock. Period. This is an indisputable technical fact. Now, what happens after data is taken out of the memory might still be affected by what the DAC was connected to - but it is not jitter/pn coming from input side. No way.
- If it where possible to stop the changing of the clock in the DAM and still keep the 44 sample buffer (e.g. by feedbacking of clock to source - i.e. making DAM in same sense master) you would actually listen to the Si clock only - no jitter/pn would emanate from incoming side that could affect the SQ. You can argue that the pn from Si is to poor to make a good sound - shure, but that would be what you listen to in terms of jitter/pn
I have now said all I have to express in this matter 🙂
//
- No jitter kan stay alive if payload data is written to memory and clocked out by a different clock. Period. This is an indisputable technical fact. Now, what happens after data is taken out of the memory might still be affected by what the DAC was connected to - but it is not jitter/pn coming from input side. No way.
- If it where possible to stop the changing of the clock in the DAM and still keep the 44 sample buffer (e.g. by feedbacking of clock to source - i.e. making DAM in same sense master) you would actually listen to the Si clock only - no jitter/pn would emanate from incoming side that could affect the SQ. You can argue that the pn from Si is to poor to make a good sound - shure, but that would be what you listen to in terms of jitter/pn
I have now said all I have to express in this matter 🙂
//
We have a different point of view on this, I respect your opinion but I have not yet found a so called modern DAC sounding better than a properly implemented NOS TDA1541A DAC, despite it's a old 16 bit DAC.
We have a different approach, we privilege the timing aspect in a DAC rather than the accuracy.
A true 18 bit DAC is 108 dB, more than enough.
Maybe the only digital manipulation we could provide will be a one time DAC calibration, but I'm not sure.
Originally Posted by andrea_mori:
"And the starting condition of the DAM is far worse than that of Ian's FIFO buffer."
I agree to this. The total working of the DAM fifo is less ideal than the Ian one. Already on the drawing board the DAM is doomed. Ian is theoretically isolating. Then comes implementation and the real world.
//
"And the starting condition of the DAM is far worse than that of Ian's FIFO buffer."
I agree to this. The total working of the DAM fifo is less ideal than the Ian one. Already on the drawing board the DAM is doomed. Ian is theoretically isolating. Then comes implementation and the real world.
//
So almost all useful features of the DAM DAC will be stripped. With the hope of achieving an (audibly) lower jitter. By the way: do you have any idea of the jitter introduced in the shift registers? This may limit the scale of improvement possible...
That's a bold claim to make, if I was either Soren or Ian I would probably assume you're trying to sell something by trying to discredit the competition, a tactic that's somewhat slanderous and doesn't paint a very favorable picture of your motivations or your moral for that matter.
As I mentioned in a previous comment, which you conveniently choose to ignore: extraordinary claims require extraordinary evidence. If you think you have a superior approach you illustrate that with factual evidence, math, or measurements, not assumptions and hearsay. But if you lack that, which seems to be the case, your next best approach is to discredit others?
Take note that without a 1:1 comparison with your proposed solution on the basis of actual measurements you have nothing but your baseless claims and endless, poorly worded (also from a technical point of view) posts. All I can say at this point is buyer beware ...
You are totally out of the way.
We have nothing to sell, we are hobbyist without commercial interests.
I work in IT, I don't own a company focused on electronic devices like someone other.
We develop audio devices for ourselves, then we share them with the audio community for those who could be interested on them.
We don't claim jitter < 1 ps when the integration bandwidth used for the measurement is 12 kHz to 20 MHz.
We don't call our devices "absolute".
Our oscillators have been all measured and the result have been published.
The results are clear, but if you don't trust our plots you can make the measurements by yourself.
We have measured and published the phase noise of Ian's FIFO and the plots have been published.
We will measure soon the phase noise of the DAM1021 and the results will be published.
Therefore not assumptions but real measurements with the appropriate instruments.
Please, look at the dedicated threads, I will not publish again the measurements in this thread.
And as I have already wrote you are welcome in the oscillators thread if you have something to dispute, even in the FIFO thread and also in the power supply thread.
Is it so difficult to understand for you that we are not competitor since we don't have and we don't look at audio business?
I believe I have already replied: "I have not yet found a so called modern DAC sounding better than a properly implemented NOS TDA1541A DAC, despite it's a old 16 bit DAC".
About the digital volume control my opinion is here
Reference DAC Module - Discrete R-2R Sign Magnitude 24 bit 384 Khz
Unfotunately it will be so, I believe there is a price to pay to improve the SQ performance tweaking the DAM.
As I said I cannot access the firmware source code, so the only possible approach is replacing all the front end (FPGA, micro and VCXO).
About the jitter of the 595s there is nothing to do, I cannot replace them.
However I'm confident the 595s are the least problem if they are clocked by a clean latch.
Cool!
Maybe ground dirt and capacity/inductive coupled RF "sounds" different than *jitter* 😉
//
Maybe ground dirt and capacity/inductive coupled RF "sounds" different than *jitter* 😉
//
#72
"Without manipulations (for example oversampling, dithering, noise shaping, "self calibrating" etc...etc) these parameters (xx bits, yyy THD etc.) remains unreachable."
Holo May is no exception.
#72
"Without manipulations (for example oversampling, dithering, noise shaping, "self calibrating" etc...etc) these parameters (xx bits, yyy THD etc.) remains unreachable."
Holo May is no exception.
I am basically open to the NOS design. But I find it striking that no one shows (or I don't find...) measurement data at just under the 1/2 Nyquist frequency. Especially not when this is at 44.1 kHz. Not even sinusoidal measurements at, say, 10-20kHz.... I only find variants where over- or upsampling was done before to set the Nyquist frequency high to be able to use an analog flat filter - which is no longer a NOS design. Are there any links to such measurement series? I only found this: Picture
I'm not much interested on THD, I play a DHT amplifier with 2% THD, despite I'm very happy with it.
I'm much more interested on timing errors.
However the Holo DAC it's a PWM DAC, here we are talking about PCM DAC.
I don't know if there are more of such measurement, however even the old TDA1541A can be oversampled up to 352 kHz.
Many users have run it at such that frequency.
Simply we don't provide the oversampling inside the FIFO and the DAC, but it could be done externally (up to the input sample rate frequency limit of the FIFO Lite, so no chance to get 3.072 MHz or similar crazy speed).
Then the NOS or not NOS feeling depends on personal taste.
Here Metrum r2r (Multibit) DACS | Audio Science Review (ASR) Forum is another picture from 10kHz Sinewaves of a real NOS-DAC without sampling (Metrum Quad DAC). Some guys are meanwhile using SRC4192 (o.a.) on the PCBs to make a clean output possible. Using high Nyquist frequency and a flat reconstruction filter.