Has anybody, besides the Star Pure people, made a working controller for the CT7302PL in order to convert PCM to DSD outside the computer. My problem is, that I need 4 channels and at least two of them in DSD format.
I can make J-river put out 4 channel DSD but there are no 4 channel USB to DSD interface that can handle DSD512.
I have a dac that needs DSD for the main speakers , and I need 2 more channels for the subwoofers. That does not need to be DSD.
The speaker system is EQ'ed and room corrected and crossed over by Audiolense.
I can make J-river put out 4 channel DSD but there are no 4 channel USB to DSD interface that can handle DSD512.
I have a dac that needs DSD for the main speakers , and I need 2 more channels for the subwoofers. That does not need to be DSD.
The speaker system is EQ'ed and room corrected and crossed over by Audiolense.
I got one of the eval boards, programmed it with Arduino, and tried out the part in its highest quality external clock mode (among other modes). Bottom line: forget it. AK4137 is still the best choice, followed not too far behind by SRC4392 (all have different feature sets, I know).
Last edited:
Thanks for your reply.
It puzzeles me a little that you found the AK4137 better than the CT7302PL.
The guys behind the very good sounding direct DSD dac the DSC 1 and 2(this was first made by the man behind HQ player , Signalyst) says the precisly opposite:
PureDSD
If you scroll way down, there is a description of the AK4137 and in that, they say :
NEWS: the CT7302 chip used like PCM to DSD converter give better performance.
And they refer to this :
STAR Pure DSD DAC
Looks like they never finished the job , though.
Did you use it with a direct DSD dac after?
No. Used an AK4137 eval board.
Regarding performance of CT7302PL verses AK4137, those other guys may be right based on THD+N measurements. I thought CT7302PL sounded pretty bad compared to the best I know how to get out of AK4137. That's for both for upsampling and or for conversion to DSD at any sample rate. Since I was in contact with Comtrue I checked with them to make sure I was using CT7302PL correctly. They confirmed I was. However, they seemed pretty clueless about how it sounded so I suppose it was probably designed almost exclusively by measurements. Personally, I am one of those who believes listening tests and measurements should both be used in audio design. That said, people can certainly fool themselves when listening if not very careful. I use multiple trusted listeners to help me evaluate sound quality in many cases, and did so when comparing the sound of CT7302PL and AK4137.
The comments was based on listening experience:
The AK4137 could not compete with a PC doing the conversion from PCM to DSD and using a direct DSD dac. The CT7302PL , on the other hand , was doing as well or even better.
Maybe the difference in perceived performance is that you use a AK4499. This is a PCM DAC and converts DSD to PCM before it makes the D to A conversion.
The hole quest is to get any format converted to DSD so you can use a Direct DSD DAC. Have you ever heard such a DAC. I will urge you to do that. Better than anything I have heard of PCM DAC s. And in quite a different way. There are differences between the various High Performance DAC's ( After very long time with all sorts of different DAC's I ended up with a very special way of using the old TDA1541A as my favorite) DSD direct just puts you closer to the music and in a very analog way. IMHO
AK4499 was set in Volume Bypass mode which disables conversion to PCM, so I don't think that's the explanation.
SD dacs are more sensitive to clock jitter than some other types of DACs. That might have something to do with it since CT7302PL sure sounded jittery to me. I can avoid that with AK4137. No fix for CT7302PL
Also, there are tips and tricks for using AK4137 to best effect. AKM does not recommend it for going directly from 16/44 to DSD256, for example. Also, its DSD CLK is inverted from what most of the rest of the world uses (but its configurable in a control register). Also, AK4137 is sensitive to power supply noise from digital circuitry in the vicinity especially if on a 2-layer PCB. In addtion, it needs a very low jitter reference clock. If one does everything right, it can be very good indeed, pretty close to the best of HQ Player.
I can't imagine how CT7302PL could sound as good as HQ Player under any circumstances if a well designed USB board with very low jitter clocks is used to drive the dac.
In any case, I stand by my own results. Don't know how the other guys got their results, so can't speak to why they might be different.
SD dacs are more sensitive to clock jitter than some other types of DACs. That might have something to do with it since CT7302PL sure sounded jittery to me. I can avoid that with AK4137. No fix for CT7302PL
Also, there are tips and tricks for using AK4137 to best effect. AKM does not recommend it for going directly from 16/44 to DSD256, for example. Also, its DSD CLK is inverted from what most of the rest of the world uses (but its configurable in a control register). Also, AK4137 is sensitive to power supply noise from digital circuitry in the vicinity especially if on a 2-layer PCB. In addtion, it needs a very low jitter reference clock. If one does everything right, it can be very good indeed, pretty close to the best of HQ Player.
I can't imagine how CT7302PL could sound as good as HQ Player under any circumstances if a well designed USB board with very low jitter clocks is used to drive the dac.
In any case, I stand by my own results. Don't know how the other guys got their results, so can't speak to why they might be different.
Last edited:
From the data sheet for the AKM 4499:
The AK4499 is a 32-bit 4ch Switched Resistor DAC which adopts newly developed technology,
achieving the industry’s leading level low distortion and low noise characteristics.
This is not Direct DSD dac. Direct dsd dac is in its simplest form only a lowpass filter. The DSC 1 and 2 uses shift registers to reduce the noise in such a DAC , but it is not a conversion DAC as the AKM 4499 is AFAIK...
Have you ever heard such a DAC?
Still I value very much your opinion and thoughts and will have that in mind in further investigations.
The AK4499 is a 32-bit 4ch Switched Resistor DAC which adopts newly developed technology,
achieving the industry’s leading level low distortion and low noise characteristics.
This is not Direct DSD dac. Direct dsd dac is in its simplest form only a lowpass filter. The DSC 1 and 2 uses shift registers to reduce the noise in such a DAC , but it is not a conversion DAC as the AKM 4499 is AFAIK...
Have you ever heard such a DAC?
Still I value very much your opinion and thoughts and will have that in mind in further investigations.
Take a look at this thread:
https://www.diyaudio.com/forums/dig...ar-pure-dsd-dac-signalyst-17.html#post6286367
https://www.diyaudio.com/forums/dig...ar-pure-dsd-dac-signalyst-17.html#post6286367
I know, saw it before.
Look, one can't talk about a DAC chip or an ASRC chip in isolation. To say that ES9038Q2M sounds bright and harsh really speaks to the implementation more than it does the dac chip. Same for ASRC chips like AK4137 because their performance may be very dependent on the circuity around them.
How about let's see some pics of their test AK4137 board, so we can see how its clocked, powered, what feeds its I2S inputs, etc.? Also, how are the I2C registers configured? What are the HEX (or decimal) codes written to each register? If that doesn't solve the mystery then one might start looking more at the characteristics of the dacs AK4137 is tested with.
Look, one can't talk about a DAC chip or an ASRC chip in isolation. To say that ES9038Q2M sounds bright and harsh really speaks to the implementation more than it does the dac chip. Same for ASRC chips like AK4137 because their performance may be very dependent on the circuity around them.
How about let's see some pics of their test AK4137 board, so we can see how its clocked, powered, what feeds its I2S inputs, etc.? Also, how are the I2C registers configured? What are the HEX (or decimal) codes written to each register? If that doesn't solve the mystery then one might start looking more at the characteristics of the dacs AK4137 is tested with.
You are wrong here.. When in direct DSD mode,
Ak4499 is hardly different from that DSC type structure.
All that left in circuit is the output stage.
The difference is that from the DSC architecture one will never get such a low noise & distortion result of what the 4499 chip is capable of..
Ciao, George
Is this really so?
Wonder why AKM dosent tell anything about that.
From the block diagram they call, what you are referring to as outputstage, for SR DAC. And the output is a current output from a switched resistor network. Dosen't sound like anything like the DSC No-Dac topology, but I might be wrong....
Wonder why AKM dosent tell anything about that.
From the block diagram they call, what you are referring to as outputstage, for SR DAC. And the output is a current output from a switched resistor network. Dosen't sound like anything like the DSC No-Dac topology, but I might be wrong....
Look.. I understand You had fallen for the DSC concept.
So let's take that as a starting point.
Now: Who is the original author of that concept? Miska. (at least for this diy community..)
And what does he suggest to use as an appropriate configuration to use with his HQP software?
"or use an AKM chip in direct DSD configuration.."
And in effect, he uses an RME ADI pro, in direct DSD modality..
Other dac he is using is a Holo Spring, which is kind of based on the DSC concept.
He does not suggest to use ESS dacs, given that those are 'closed format', and do manipulate the input DSD stream..
Ciao, George
So let's take that as a starting point.
Now: Who is the original author of that concept? Miska. (at least for this diy community..)
And what does he suggest to use as an appropriate configuration to use with his HQP software?
"or use an AKM chip in direct DSD configuration.."
And in effect, he uses an RME ADI pro, in direct DSD modality..
Other dac he is using is a Holo Spring, which is kind of based on the DSC concept.
He does not suggest to use ESS dacs, given that those are 'closed format', and do manipulate the input DSD stream..
Ciao, George
I am using cometrue board. With some of CT model chip (I dont know byhart which one is on the board.)
It have in the same time I2S (or some other formats) AND DSD output. You can use them with same song playing. PCM or DSD computer play. The chip doing auto hardware conversion to DSD and PCM and output simultaneously booth formats, so yes You can have 2+2 output with dedicated separate converters after ofcourse
George, that analog FIR unity weighted bits LPF concept is very old.
If you look at Burr Brown DSD1700 'DAC' from 1999, it is basically same
thing.
TCD