[HeyManslowdown97]

Yes but what's is function in the CPU and the I2s design in how it relates to the frequency in the signal being transmitted into the DAC.

I believe that’s what it’s doing.

The I2S is all encompassing with clocks and signals, data. It gets decoded but the I2S the only way to decode DSD without a HDMI cable in SACD because Sony has copyrighted its use. If for instance you get a SACD player it can only transmit the digital signal over HDMI but that’s not the case with I2S

It doesn’t convert DSD only PCM.

 

[Ivan TT]

Yes but what's is function in the CPU and the I2s design in how it relates to the frequency in the signal.

Sorry, can you paraphrase this as I cannot quite understand?

If I understand question correctly, CPU handles reading data from TF card, GUI, EQ and sending control signals via I2C (not to be confused with I2S) bus such as digital filter selection or "high quality mode" enable.

CPLD (as far as I understand) is programmed to convert data to I2S format and selects appropriate clock (45 or 49MhZ) based on instructions from CPU depending on the sampling frequency of the recording and sends data and clock to DAC via I2S bus.

It may or may not turn power on or off controlling power supply, but data is kept separate and independent from I2S and I2C interfaces and is not modulated or otherwise influenced by data stream in any way.

PS: Specifically the picture posted in the first post of this topic clearly identifies Digital and Analogue parts of T1:

Link:
https://www.head-fi.org/threads/zishan-t1-hi-fi-player-thread.911795/

PPS:

It gets decoded but the I2S the only way to decode DSD without a HDMI cable in SACD because Sony has copyrighted its use. If for instance you get a SACD player it can only transmit the digital signal over HDMI but that’s not the case with I2S

It doesn’t convert DSD only PCM.

I2S does not decode anything, it's just a protocol (software and hardware) to send digital data stream (both PCM and DSD) over 3 (or in some implementations 4) wires, usually internally (within DAC/DAP).

If you check Pin Functions table in AK4497 datasheet (page 6) you will see that pins 3, 4 and 5 have different function depending on whether PCM or DSD data is recieved via I2S bus ( this is determined by some DSDPATH bit, controlled either by CPU or CPLD).

So to conclude: I2S allows for both PCM and DSD data transfer and there is a way to switch DAC from reading data in PCM or DSD formats, that are encoded in a different way (PCM is interleaved, DSD separately for L and R channels).

 

[HeyManslowdown97]

CPLD (as far as I understand) is programmed to convert data to I2S format and selects appropriate clock (45 or 49MhZ) based on instructions from CPU depending on the sampling frequency of the recording and sends data and clock to DAC via I2S bus.

Therefore it is being transmitted via the I2S as a digital signal to the DAC there’s decoded at the output as analog. The conversion has been made before it hits the DAC. So it’s a Direct Stream via the signal from the clocks.

 

[Ivan TT]

Therefore it is being transmitted via the I2S as a digital signal to the DAC there’s decoded at the output as analog. The conversion has been made before it hits the DAC. So it’s a Direct Stream via the signal from the clocks.

Not entirely.

Firstly, data is converted to analogue form at the output of DAC, Digital to Analogue Convertor as per block diagram posted above.

The conversion of data itself (from PCM to DSD) is not performed by neither CPU or CPLD, but inside of the DAC. Essentially it means that ANY data is converted to DSD stream inside of DAC before it's converted to analogue domain, even mp3 files! And also it means that's there's no benefit to convert PCM files to DSD using computer and conversion software as this conversion is routinely performed by DAC itself (being delta-sigma DAC it can only convert DSD data stream to analogue domain, it cannot convert PCM directly).

I don't understand what you mean when you say "So it’s a Direct Stream via the signal from the clocks" so cannot comment, sorry.

 

[HeyManslowdown97]

I don't understand what you mean when you say "So it’s a Direct Stream via the signal from the clocks" so cannot comment, sorry.

The Msop switches frequencies and doesn’t just convert voltage. PWM is how the Data travels.

(Digital) Bjt outputs are 250 ma there my be a conversion going on therefore between a digital input into and and analogue output Bjt.

If so there is an advantage of playing DSD files over PCM.

I disagree that there’s audio advantages converting to DSD. They sound better in my opinion!

 

[Ivan TT]

The Msop switches frequencies and doesn’t just convert voltage. PWM is how the Data travels.

MSOP is a form-factor, I understand you mean DC-DC convertor, MP1542 in stock?

It utilises PWM indeed, but it has nothing to do with data or its travel (that would be I2S bus exclusively), modulating width of pulses allows it to maintain constant voltage output under load by increasing width of pulses (thus providing more current) or decreasing if current is not needed thus making them more efficient. There's usually two FIXED frequencies at which they operate: 700kHz for lighter loads and 1.2-1.3MHZ for heavier loads (when high current output is required).

This unfortunately pushes a bit of noise into analogue power supply that then is filtered (attenuated) by a combination of inductors (T1 uses small SMD inductors on DAC/LPF daughtercards by the way) and de-coupling capacitors, removing AC noise as much as possible and making power supply DC.

DSD does not utilise PWM actually, but PDM, Pulse Density Modulation, the difference is that pulses does not change length as with PWM, but same length pulses are placed more densly for higher voltage (well, actually current hence current output DACs such as 4499 or ESS family, it's simply more straightforward to implement).

Regardless, these are completely independent systems and there is no digital data feed into DC-DC converter at all.

(Digital) Bjt outputs are 250 ma there my be a conversion going on therefore between a digital input into and and analogue output Bjt.

There is no data conversion taking place in output transistors either, as they belong to analogue section of T1 as per block digram above and all the Digital to Analogue Conversion is performed by DAC and DAC only. BJT transistors are not DACs, they don't perform digital to analogue conversion, DAC does.

To reiterate and as per T1 block diagram posted earlier: digital data is supplied to Digital to Analogue Convertor (DAC) via I2S interface, is converted to analogue domain and from this point onwards (being DACs analogue output pins) all the way to headphones signal is analogue (including input of transistors), pink background in the diagram.

If so there is an advantage of playing DSD files over PCM.

Sorry, I don't understand, can you speak more clearly please?

I suggest reading a wikipedia article on DC-DC converters (not to be confused with DAC converters), I'm sure there is one there and it may help to clarify what DC-DC converters do, including the fact that they are purely analogue devices converting between voltages.

I disagree that there’s audio advantages converting to DSD. They sound better in my opinion!

Fair enough, no argument from me.

 

[HeyManslowdown97]

Yea, some DSD is played through a PCM Container DoP it’s actually tricks the dac to think it’s PCM but this requires lots of CPU and there absolutely no audio difference I cannot hear to difference (lots of purest fight over this Native vs DoP)

So because you don’t listen to DSD files the lpf in the DSD Pro really is a big loss for you based on how the PDM signal relates to DSD?

“The process of decoding a PDM signal into an analog one is simple: one only has to pass the PDM signal through a low-pass filter. This works because the function of a low-pass filter is essentially to average the signal. The average amplitude of pulses is measured by the density of those pulses over time, thus a low pass filter is the only step required in the decoding process.”

 

[Ivan TT]

Yea, some DSD is played through a PCM Container DoP it’s actually tricks the dac to think it’s PCM but this requires lots of CPU and there absolutely no audio difference I cannot hear to difference (lots of purest fight over this Native vs DoP)

Sorry, but there's a bit of confusion here I'm afraid.
DoP is related to USB interface, not I2S.
It is a way of encapsulating DSD data into PCM container, but it does not covert it to PCM at all.
At the receiving end it is treated as DSD data (which it remains at all the times).

So because you don’t listen to DSD files the lpf in the DSD Pro really is a big loss for you based on how the PDM signal relates to DSD?

I do listen to DSD files, but delta-sigma DACs have built in (internal) digital filter so DACs analogue output may not even require extra filtering, but there is at least 1 1-pole analogue LPF in my mods (approximately 150kHz cutoff frequency), which I find sufficient based both on subjective (listening) and objective (measurements) tests.

 

[fabien32]

Correction: AD8516 not AD8556 (which has nothing to do with audio op amps lol)

I replaced the Dual crystal oscillator with two NDK NZ2520SD in the DSD Pro. I am not sure what frequency are in the T1 yet.

I will also replace the microprocessors in the DAC that control the high frequencies and PWM which I will do again.

The best way I can describe it’s effect on audio in my own words is, have you felt some parts speed up while others are slowed down. That's what it feels like. When you ”fix” it the audio sounds like a metronome that musicians use to keep time but it’s not “clicking” the meter instead it’s keeping time. Some people don’t like it and feel it’s to clinical, “cold” “analytic”. I don’t however.

There’s are especially two PWM versions for audio one being the Class D amplifier and the other SACD encoder which the Zishan DAC are designed using.

“A new class of audio amplifiers based on the PWM principle is becoming popular. Called class-D amplifiers, they produce a PWM equivalent of the analog input signal which is fed to the loudspeaker via a suitable filter network to block the carrier and recover the original audio. These amplifiers are characterized by very good efficiency figures (≥ 90%) and compact size/light weight for large power outputs. For a few decades, industrial and military PWM amplifiers have been in common use, often for driving servo motors. Field-gradient coils in MRI machines are driven by relatively high-power PWM amplifiers.
Historically, a crude form of PWM has been used to play back PCM digital sound on the PC speaker, which is driven by only two voltage levels, typically 0 V and 5 V. By carefully timing the duration of the pulses, and by relying on the speaker's physical filtering properties (limited frequency response, self-inductance, etc.) it was possible to obtain an approximate playback of mono PCM samples, although at a very low quality, and with greatly varying results between implementations.

In more recent times, the Direct Stream Digital sound encoding method was introduced, which uses a generalized form of pulse-width modulation called pulse density modulation, at a high enough sampling rate (typically in the order of MHz) to cover the whole acoustic frequencies range with sufficient fidelity. This method is used in the SACD format, and reproduction of the encoded audio signal is essentially similar to the method used in class-D amplifiers.”

in crystal oscillator what is better higher ppm in frequency stability or lower ?

 

[AxelCloris]

Please refrain from making any personal comments in the discussion. Thank you, everyone.

 

[HeyManslowdown97]

DoP is related to USB interface, not I2S.
It is a way of encapsulating DSD data into PCM container, but it does not covert it to PCM at all.
At the receiving end it is treated as DSD data (which it remains at all the times).

That’s correct I wasn’t sure about it being on portable players.

The PCM filters are way better on the AKM4497eq with more noticeable effects!

I have heard the ES9038 has way better digital filters and PCM sounds better on them than the Ak4497.

I have listen to the ES9018 chip and have mixed feelings about it but those chips have lots of power and the bass response is really tight. The soundstage in was a bit muddled in the DAC that heard it in but I’ve heard some really nice things about the flagship ES9038 and at 50 for the Zishan I think it’s gonna be a steal if it’s well implemented.

You should sell one of your ak4497 players and get the T1. It’s definitely an great upgrade. The T1 Ak4499 will probably be good but it take some time on the market before anyone perfects its implementation. Make no mistake I’m gonna get it though. Haha

Ivan, Sorry things got out-of-hand, moving forward I’m gonna listen more and limit the noise on the thread. I realize things got out of hand this morning when I woke up and saw what was being posted and Icefrosty plea that we get along. It’s not fair to anyone and I regret that it ever happened and I contributed to it!

I do respect your opinions, knowledge and contributions to the thread... just said some stuff out of anger. I did, however, mean for you and encourage you in pursuing your own projects with your ideas and build stuff of your own and who knows maybe we’ll be modding one of your projects on the head fi one day.

Let’s just put all this in the past, move on and start have some constructive discussions about the T1, all things Zishan and mod related.

 

[HeyManslowdown97]

Your post made me realize that 4497 is not using buf634s which drives headphones in "dsd 4495seq". maybe that is the reason for why 4495 was pricier than 4497..

If you wanted to you could probably replace the Buff634 2x with the AD8397 2x?? You may like the way it sounds better if it’s ok to use them as buffer amps??!

 

[fabien32]

ES9038 has way better digital filters and PCM sounds better on them than the Ak4497

i have real good portable dacs with ES9038 like the NX4, and i even have the D50, and the T1 sound better, maybe not soundstage wise and maybe not in all and every tiny aspects, but, overall i prefer the sound of the T1

 

[HeyManslowdown97]

i have real good portable dacs with ES9038 like the NX4, and i even have the D50, and the T1 sound better, maybe not soundstage wise and maybe not in all and every tiny aspects, but, overall i prefer the sound of the T1

I don’t have any dacs with ES9038 and I’m sure there really good when well implemented. I totally agree with you about the T1 it’s special and sounds completely different from the DSD Pro.

Since I am a “why” guy it would be awesome to take a deep dive and see what makes the T1 sound better/different. It sounds to me to be more it’s warm, the vocals sound great and the tremble is perfect! To me it sounds like a DSD Pro with an AD8620 only without the harshness in the higher frequencies.

At times the T1 can sound metallic but that’s most likely the stock amps.

If I had to give one sentence for describing the T1 it’s more intimate with lots more headroom for improvements.

I think the balanced line has a different TRRS?? In the 2.5mm?? I’m having issues with output! I was always unclear about the different in the balanced 2.5 versions in the headphone jack.

I can’t wait to experiment with different op amps.

 

[fabien32]

TRRS?? In the 2.5mm?? I’m having issues with output

i also have a problem with the 2.5 TTRS with one headphone jack, when not pushing all in it was ok. with my others2.5 TTRS no problem