Hello fmak,
this is exactly the way my modded DV-575 is working.
I am doing an upsampling to 96k with a CS8421, no matter what the input signal is. This saves me from the need for additional logic and gives me exactly the sample rate my D/A is using. The only drawback is that the current version doesn't support 192k as input.
Regrettably I'm a bit short of time this week, if you're interested we can discuss this later.
And Charly, thank you very much for the inspiration of modding the DV-575!
Holger
this is exactly the way my modded DV-575 is working.
I am doing an upsampling to 96k with a CS8421, no matter what the input signal is. This saves me from the need for additional logic and gives me exactly the sample rate my D/A is using. The only drawback is that the current version doesn't support 192k as input.
Regrettably I'm a bit short of time this week, if you're interested we can discuss this later.
And Charly, thank you very much for the inspiration of modding the DV-575!
Holger
Hello fmak, what you want is impossible!
The AD1896 runs in master mode and in this mode the highest output rate that could be achieved is 96KHz. This depends on the maximum clock input of 30Mhz which means a maximum output rate of 117187,5 KHz.
This means upsampling is only possible up to 96kHz with the AD1896 :-((
The only chance to get 192kHz out of the AD1896 is to let it run in slave mode. This means you have to supply LRCLK and BCLK into the AD1896. Therefore major changes on the board have to be done (e.g. an additional generator creating LRCKL and BCLK from an external clock generator) which is not possible with this board.
But why upsample so high? You can’t get more information out of a low frequency signal by upsampling.
----------------------------------------------------------------------------------oehlrich said:
My own experience is that upsampling to 176.4 and 192 with 44.1 and 48k material is very beficial to sonics. This is based on my dCS setup where 176.4 sounds as good and natural as analogue.
Part of the reason is, I believe, the shifting of the low pass filter well beyond 44.1k. On my dCS, changing the filter slope and type makes quite a difference.
Even on a computer using Foobar, upsampling sounds better.
As the AD1896 is 192k capable, why not explore it?
Seems I have to use your board w/o AD1896 and then perhaps feed the signal into the AD 1896 EB development board from Analaog Devices. Pity.
Can you describe the sonic changes to the Pioneer DV575 using your board. The unit itself apparently doesn't sound too hot.
You are right, the analogue outputs of the 575 are bad, and the digital output has too much audible jitter to be good.
Regarding your question…..
Well my setup is somehow different from normal setups and difficult to compare.
The digital signal from the 575A is fed digital (44.1 to 192) into a Behringer DCX2496. Originally the input of the DCX2496 is equipped with an old SRC CS8420. Not a good SRC and also not capable of 192kHz. So I looked around what to do and discovered the DAC1 from benchmark.
It has the unique ultrlock technique (pure maketing) but despite of this it is a very remarkable unit because it eliminated jitter nearly completely. I studied that ultralock technology and found out it works with an AD1896. The incoming signal can be all speed, up to 192kHz. The output samplerate for the DACs of the DAC1 is aprx. 110kHz.
Well the DCX2496 works with constant 96KHz internally. Inside my DCX2496 I installed a tentlabs clock for the D/As and a small PCB with a 192kHz receiver (CS8416) and a AD1896 like the DAC1 has http://freerider.dyndns.org/anlage/Behringer-Input-Stage-E.htm. So I have that ultralock technology inside my DCX2496!
Connecting the 575 to this modded unit is really an experience, probable caused by the upsamped input signal and the internal reference clock (tentlab says 3ps!). Never heard that before and I am very happy with his. The result is satisfactory at all and the jitter from the 575 does not matter any more. I tested different digital wires with no audible effect. I even tested transmitting the S/PDIF via telephone lines (really bad S/PDIF signal) http://freerider.dyndns.org/anlage/telefon.htm about a long distance with no audible effect. This means ultralock really works great. I also tested some very expensive CD players digital connected to the modded DCX2496 and got no audible difference to 575A.
So your question about sonic changes I could not really answer. In my setup the result is really satisfactory and I didn’t want to miss it anymore. Also digital SACD and DVD-A is much better than 44.1 and much better than the analogue output of the 575A.
oehlrich said:
You are right, the analogue outputs of the 575 are bad, and the digital output has too much audible jitter to be good.
It has the unique ultrlock technique (pure maketing) but despite of this it is a very remarkable unit because it eliminated jitter nearly completely. I studied that ultralock technology and found out it works with an AD1896. The incoming signal can be all speed, up to 192kHz. The output samplerate for the DACs of the DAC1 is aprx. 110kHz.
Well the DCX2496 works with constant 96KHz internally. Inside my DCX2496 I installed a tentlabs clock for the D/As and a small PCB with a 192kHz receiver (CS8416) and a AD1896 like the DAC1 has http://freerider.dyndns.org/anlage/Behringer-Input-Stage-E.htm. So I have that ultralock technology inside my DCX2496!
-------------------------------------------------------------------------------------Thanks, interesting. What about a pcb for this, or is this part of the enhanced DV575 board, outputting at 96k?
The 1896EB board with 192k dac is reasonably priced and I am planning to get one to try.
It could not be part of it because the ultralock has to be near the D/As so the player is the wrong place.
I made a board and all information about this you will find here http://freerider.dyndns.org/anlage/Behringer-Input-Stage-E.htm . At the bottom of the page in the box you find a link to download the PCB board data.
Please give me a link to that AD1896EB board. Sounds interesting!
----------------------------------------------------------------------------------oehlrich said:
http://www.analog.com/UploadedFiles/Evaluation_Boards/Tools/61313665AD1896EB.pdf
is where you find it. Can be bought on the web.
Huh? On another forum, where Carlos showed his mod to analog part of this player he claimed that 575 has great clock and there's no need to replace it. Can you explain me, what do you mean by "audible jitter"? For me, through digital out, the unit plays perfectly. I can clearly hear the soundstge and instrument placement on Checky and ECM records... :->
Technically the 575A derives its clock signal for the audio section from a 27MHz crystal clock. 27Mhz because this clock is the reference for the video section. A crystal (no low jitter oscillator!) is already bad enough. To decrease this quality further the audio clock is derived from this 27Mhz by a PLL which adds additional jitter to the audio clock. This quite high jitter was measured and published by the German Stereoplay magazine. So far so good.
Now the whole thing starts to become interesting. The quality of what you will hear strongly depends on what will be done with this jittering digital audio signal?
- Recording it via S/PDIF at your PC the jitter will be gone completely. At playback from PC you will only hear the jitter of the PC audio clock. In general this is a clock generator with much less jitter than the 575A has. So the S/PDIF output of the PC has less jitter than the output of the 575A and in general will sound better.
- Connecting to a DAC like the DAC1 from Benchmark Media Systems which resamples the digital signal the jitter will also be gone completely. The same thing with my moded DCX2496 and its AD1896 input stage.
- Connecting the signal to a conventional DAC which regenerates the clock for its D/A chip via a PLL from the jittering input signal the amount of jitter you will hear strongly depends on the quality of the PLL inside the connected DAC. Using a DAC with a very good (maybe multistage) PLL, you will get a perfect result. Using a DAC with a weak PLL the jitter of the 575A will be audible.
So if you say your result is perfect only means you have a very good DAC with a good jitter reduction. It did not mean the digital output of the 575A has no jitter!
But be happy, you are one of the lucky guys like me who have no problem with the jitter produced by the 575A ;-))
First transmitting I2S is not that simple. You need line drivers and 75Ohm wires to get the signal out. On the other side you need receivers to get the I2S back to TTL level. These problems could be solved.
As far as I know the DEQ2496 uses the CS8420 as an input ASRC. This has the side effect, that there is no interface where to connect the incoming I2S. The 8420 has input for S/PDIF which is routed inside the CS8420 to its SRC section. Its output is run in slave mode so LRCLK and BITCLK were supplied by the DEQ (DEQ runs internally at a fixed rate of 96kHz), Same signals you also supply! This means transmitting I2S saves nothing because you have to install the line drivers and a new SRC like the AD1896 which could be supplied with your signals on one side and those of the DEQ on the other side.
I would recommend starting with a S/PDIF output for the player and (only if you really need it) you could think about an improved input stage like this http://freerider.dyndns.org/anlage/Behringer-Input-Stage-E.htm .
No, the hires board has no own clock. It uses the clock of the player and there will be the same jitter as the unmodded player has.qus said:
It makes sense to add a Tentlab clock here and output at constant 96kHz. This will completely kill jitter of the player. But transporting the signal via the S/PDIF to the D/A will again add jitter by the wire and the receiving PLL. So I more recommend a D/A converter which is jitter resistive like the benchmark DAC1 or my modded DCX2496 and output the hires signal as the kit does.
But how knows, just try the Tentlab mod and look if there will be any audible improvement.
Well - I am still looking for the two extra components, as I didn't know there's the expanded kit, when I ordered mine. They're very hard to get for hobbyists, it seems... The prices also differ much from yours. Anyway meanwhile I found another, very interesting and looong thread on 575, including fitting the clock you mentioned:
http://www.worldaudiodesign.com/forum/showthread.php?t=18703&page=1&pp=10
http://www.worldaudiodesign.com/forum/showthread.php?t=18703&page=1&pp=10
I received my kit quickly, the quality is indeed excellent. After assembling and installing the board I was only getting output at 44.1kHz although the LEDs were functioning correctly for higher sample rates. A bit of poking around with a scope revealed that a couple of the pins of the surface mount chip had failed to connect to the solder pads. After retouching the solder joints everything worked perfectly!
At the same time as doing this upgrade I completely disassembled the player and did some other mods. I cut out the bottom of the case underneath the drive mechanism with a dremel tool, and made up a new platform to support the drive consisting of a 5mm lead sheet supported by a 3mm sheet of phenolic. This new sub-chassis was then suspended from the existing case using soft rubber mounts. The front bezel of the player and the lid are partially supported by the drive mechanism as well, so some small modifications were done there so that the drive mechanism was entirely free-floating and the case remained structurally sound. The case, lid, and transport mechanism were then mechanically damped by liberal application of heavy putty and closed-cell foam. Finally, four large metal pointed feet were installed.
Using a Benchmark DAC1, the difference between the sound of regular CDs and SACD/DCD audio has gone from hardly noteworthy to breathtaking. Comparing the Santana "Supernatural" regular disc to the DVD audio version makes them almost sound like different albums, the regular one sounds broken. This was certainly a very worthwhile upgrade for me!
Take care,
Doug
After retouching the solder joints everything worked perfectly!At the same time as doing this upgrade I completely disassembled the player and did some other mods. I cut out the bottom of the case underneath the drive mechanism with a dremel tool, and made up a new platform to support the drive consisting of a 5mm lead sheet supported by a 3mm sheet of phenolic. This new sub-chassis was then suspended from the existing case using soft rubber mounts. The front bezel of the player and the lid are partially supported by the drive mechanism as well, so some small modifications were done there so that the drive mechanism was entirely free-floating and the case remained structurally sound. The case, lid, and transport mechanism were then mechanically damped by liberal application of heavy putty and closed-cell foam. Finally, four large metal pointed feet were installed.
Using a Benchmark DAC1, the difference between the sound of regular CDs and SACD/DCD audio has gone from hardly noteworthy to breathtaking. Comparing the Santana "Supernatural" regular disc to the DVD audio version makes them almost sound like different albums, the regular one sounds broken. This was certainly a very worthwhile upgrade for me!
Take care,
Doug
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.