Hey all!
I'm new to the forum as of today. I've been reading some of the posts the last few days, and have been very pleasantly surprised
Last year, I developed an outboard Dolby Digital and DTS decoder with two other guys for our Senior Design project. We based it around the older Cirrus DSP family, using the far less complex CS4926, and the DIR portion of the CS4226 multichannel CODEC. On the output, we used Burr-Brown PCM1720 DACs and some Burr-Brown OPA134 series for the analog filtering. It was an awesome experience! I also took two DSP classes senior year. The combination really got me interested in digital audio. Up to that point, I had been a real analog junkie, hoping to design amplifiers.
I currently have a Denon DCM-370 5-disc carosel, which has HDCD decoding using the Burr-Brown DACs for those who don't know. When I bought the player... (gosh, 3 years ago now!), I was simply amazed with the sound quality. To this day, I still love the "smoothness" of the HDCD DACs.
My college roommate and I constantly tried out new CD players and DVD players, trying to find a better sound. Last year, he bought the new Outlaw 950 preamp. HOLY COW! We actually found some DACs that came near the fluidness of the HDCD DACs and yet offered additional dynamics, clarity, and punch. I'm not exactly sure which chipset is used in the Outlaw, but I'm guessing it's the CS4383 DAC. Depending on the day and the particular CD I was listening too, sometimes I like the Denon's DACs, and sometimes I liked the Crystal DACs in the Outlaw. This has spurred me on to attempt to find the best sounding DAC system (of course, this is subjective).
So... my thoughts/ideas right now cosist of the following. I obviously need a DIR, and DAC as the bare minimum. But I have decided that I want to be able to obtain more than what is offered from such a system. I'd like to recover a more stable master clock than what a standard DIR can deliver in order to reduce jitter, so I'd like to either use something like the Analog Devices Sample Rate Converter AD1890, or develop my own reliable PLL circuitry. I also have been thinking of separating the oversampling circuitry from the DACs in a configuration similar to using Burr-Brown DF1704 and some PCM1704 DACs. I would also like to try some of the more exotic stereo DACs from Cirrus/Crystal.
I had initially attempted using the CS8414 DIR in our senior design project, but we were never able to get it to even lock on a signal. I spent weeks trying to analyze the darn error outputs from it (I designed a buffer stage with some 3904's driving a series of LED's) during senior design, but we eventually just moved on to using a pre-developed uP program for accessing the CS4226 we had on hand... (time was a HUGE factor). I'm wondering if any of you have some suggestions for an easier-to-work-with DIR... there's also a very good possibility I was doing something very stupid with my prototype board (Senior Design = no sleep for 9 months
)
Also, does anyone here recommend the AD1890? Anyone have any problems or good experiences with it? Recommendations for a similar, better performing device?
And... any suggestions on my selection of DACs?
In the end, I'd like to avoid having to develop a uC interface for any of the IC's I use. I'd like this design to be as simple as can be done to generate some really good sound. If a uC is absolutely necessary, what are some recommendations for some that you have had good luck with in DAC systems?
I REALLY appreciate your time if you've actually managed to make it through all this. I look forward to hearing from any of you!
I'm new to the forum as of today. I've been reading some of the posts the last few days, and have been very pleasantly surprised
Last year, I developed an outboard Dolby Digital and DTS decoder with two other guys for our Senior Design project. We based it around the older Cirrus DSP family, using the far less complex CS4926, and the DIR portion of the CS4226 multichannel CODEC. On the output, we used Burr-Brown PCM1720 DACs and some Burr-Brown OPA134 series for the analog filtering. It was an awesome experience! I also took two DSP classes senior year. The combination really got me interested in digital audio. Up to that point, I had been a real analog junkie, hoping to design amplifiers.
I currently have a Denon DCM-370 5-disc carosel, which has HDCD decoding using the Burr-Brown DACs for those who don't know. When I bought the player... (gosh, 3 years ago now!), I was simply amazed with the sound quality. To this day, I still love the "smoothness" of the HDCD DACs.
My college roommate and I constantly tried out new CD players and DVD players, trying to find a better sound. Last year, he bought the new Outlaw 950 preamp. HOLY COW! We actually found some DACs that came near the fluidness of the HDCD DACs and yet offered additional dynamics, clarity, and punch. I'm not exactly sure which chipset is used in the Outlaw, but I'm guessing it's the CS4383 DAC. Depending on the day and the particular CD I was listening too, sometimes I like the Denon's DACs, and sometimes I liked the Crystal DACs in the Outlaw. This has spurred me on to attempt to find the best sounding DAC system (of course, this is subjective).
So... my thoughts/ideas right now cosist of the following. I obviously need a DIR, and DAC as the bare minimum. But I have decided that I want to be able to obtain more than what is offered from such a system. I'd like to recover a more stable master clock than what a standard DIR can deliver in order to reduce jitter, so I'd like to either use something like the Analog Devices Sample Rate Converter AD1890, or develop my own reliable PLL circuitry. I also have been thinking of separating the oversampling circuitry from the DACs in a configuration similar to using Burr-Brown DF1704 and some PCM1704 DACs. I would also like to try some of the more exotic stereo DACs from Cirrus/Crystal.
I had initially attempted using the CS8414 DIR in our senior design project, but we were never able to get it to even lock on a signal. I spent weeks trying to analyze the darn error outputs from it (I designed a buffer stage with some 3904's driving a series of LED's) during senior design, but we eventually just moved on to using a pre-developed uP program for accessing the CS4226 we had on hand... (time was a HUGE factor). I'm wondering if any of you have some suggestions for an easier-to-work-with DIR... there's also a very good possibility I was doing something very stupid with my prototype board (Senior Design = no sleep for 9 months
)Also, does anyone here recommend the AD1890? Anyone have any problems or good experiences with it? Recommendations for a similar, better performing device?
And... any suggestions on my selection of DACs?
In the end, I'd like to avoid having to develop a uC interface for any of the IC's I use. I'd like this design to be as simple as can be done to generate some really good sound. If a uC is absolutely necessary, what are some recommendations for some that you have had good luck with in DAC systems?
I REALLY appreciate your time if you've actually managed to make it through all this. I look forward to hearing from any of you!
I think if you search on this forum you will find that CS8420, AD189x, PCM1704, DF170x, CS43122, et. al. are the standard choices. Many people also use the Philips and Analog Devices DACs. There are many discussions, schematics, photographs, and so forth.
If I may summarize some discussions about DIRs, it would be this: CS841x have Schmitt trigger inputs which make Jocko Homo angry. CS8420 has a better receiver, but requires either dozens of pull-up resistors or a microcontroller. The AD189x is an ASRC which can be controlled without a µC, but doesn't include a receiver. The DIR1703 has a really vague datasheet but seems promising.
My preference is for the CS8420. If you don't want the sample rate conversion, that stage can be bypassed.
Good luck!
If I may summarize some discussions about DIRs, it would be this: CS841x have Schmitt trigger inputs which make Jocko Homo angry. CS8420 has a better receiver, but requires either dozens of pull-up resistors or a microcontroller. The AD189x is an ASRC which can be controlled without a µC, but doesn't include a receiver. The DIR1703 has a really vague datasheet but seems promising.
My preference is for the CS8420. If you don't want the sample rate conversion, that stage can be bypassed.
Good luck!
I am also planning a DAC project at this time, using the CS8420 receiver/Sample rate converter, the Burr-Brown DF1704 (or 1706) upsampler/digital filter and four PCM1704 DACs for stereo balanced output. I/U converter and output stage will be from the Passlabs D1 (see service manual on www.passlabscom). I hope to have a draft drawing of the circuit finished next week. But I was also consiering the AD1896 SRC together with a different receiver.
But an open question was how to get digital data out of SACD or DVD-A, as there are no commercial products with digital outputs of this high-data rates. I am not sure if it is possible to pick up the data stream just before the integrated DACs.
peter
But an open question was how to get digital data out of SACD or DVD-A, as there are no commercial products with digital outputs of this high-data rates. I am not sure if it is possible to pick up the data stream just before the integrated DACs.
peter
Holy god...outboard DD/DTS is the holy grail here. ^_^ Welcome.
Personally, I would recommend DIR1701/1703 for your DIR. Good jitter rejection and has 96kHz support. If needed, pass this through an AD1896--better specs than 1890.
I've already done some substantial work on a design with DIR1701 frontend and AD1896 to AD1853 DAC; this could certainly be adapted to the DAC of your choice.
You can see some info about the project here
http://www.diyaudio.com/forums/showthread.php?threadid=5069&highlight=1853
or on my website:
http://www.anidian.com/audio/construct/adac_1.shtml
If you would be interested in collaborating send me an email. (Work has been paused due to my recent marriage ^_^)
Personally, I would recommend DIR1701/1703 for your DIR. Good jitter rejection and has 96kHz support. If needed, pass this through an AD1896--better specs than 1890.
I've already done some substantial work on a design with DIR1701 frontend and AD1896 to AD1853 DAC; this could certainly be adapted to the DAC of your choice.
You can see some info about the project here
http://www.diyaudio.com/forums/showthread.php?threadid=5069&highlight=1853
or on my website:
http://www.anidian.com/audio/construct/adac_1.shtml
If you would be interested in collaborating send me an email. (Work has been paused due to my recent marriage ^_^)
Wow, responses already - you guys rule !!
Thanks a bunch now, and look forward to reading some more.
Tiroth, the DD/DTS decoder is going to be impossible for anyone without a few million $$ laying around to pay for the licensing fees. We were granted some educational licenses for our project from both Dolby and DTS... It was a lot of fun though
Tiroth, I would be interested in collaborating... my problem is that I have absolutely no experience with layout or any tecnique with respect to that.
I'll get back to you later tonight after work -
Take care all
Thanks a bunch now, and look forward to reading some more.
Tiroth, the DD/DTS decoder is going to be impossible for anyone without a few million $$ laying around to pay for the licensing fees. We were granted some educational licenses for our project from both Dolby and DTS... It was a lot of fun though
Tiroth, I would be interested in collaborating... my problem is that I have absolutely no experience with layout or any tecnique with respect to that.
I'll get back to you later tonight after work -
Take care all
Tieftoener said:
Have you checked here?
http://www.diyaudio.com/forums/showthread.php?s=&threadid=8056
more options
Hi,
You wrote: "I'd like to recover a more stable master clock than what a standard DIR can deliver in order to reduce jitter, so I'd like to either use something like the Analog Devices Sample Rate Converter AD1890, or develop my own reliable PLL circuitry. "
Two more options:
- async reclocking
- master clock in DAC and feed it back to the transport.
Last option is the way i am going: 2*TDA1541 (one left, one right channel) and async FIFO between DAC and transport.
Greetings,
Guido B
Hi,
You wrote: "I'd like to recover a more stable master clock than what a standard DIR can deliver in order to reduce jitter, so I'd like to either use something like the Analog Devices Sample Rate Converter AD1890, or develop my own reliable PLL circuitry. "
Two more options:
- async reclocking
- master clock in DAC and feed it back to the transport.
Last option is the way i am going: 2*TDA1541 (one left, one right channel) and async FIFO between DAC and transport.
Greetings,
Guido B
>>>Holy god...outboard DD/DTS is the holy grail here. ^_^ Welcome.
I do not want to decode the DD signal, just get the already decoded I2S data stream out of the player to a different DAC. And not even for DD or any multi-channel, only for DVD-A and SA-CD.
And about licence: Don't want to sell this, and I don't care what the industry dislikes to be done. I take the right to use any bit that's on the disk, I paid for it. I don't want to copy anything. Just listen to the recording, with equipment I can trust.
skaara: will do. Just some delay because my newly build Phono preamp makes some trouble, see pass forum. And I am nut sure yet to take the CS842 or the AD1896/DIR1203.
Tieftoener:
I will do a layout for my project, maybe we can get it modular to put a DSP module as an option on it?
I do not want to decode the DD signal, just get the already decoded I2S data stream out of the player to a different DAC. And not even for DD or any multi-channel, only for DVD-A and SA-CD.
And about licence: Don't want to sell this, and I don't care what the industry dislikes to be done. I take the right to use any bit that's on the disk, I paid for it. I don't want to copy anything. Just listen to the recording, with equipment I can trust.
skaara: will do. Just some delay because my newly build Phono preamp makes some trouble, see pass forum. And I am nut sure yet to take the CS842 or the AD1896/DIR1203.
Tieftoener:
I will do a layout for my project, maybe we can get it modular to put a DSP module as an option on it?
pquadrat:
>>I do not want to decode the DD signal, just get the already decoded I2S data stream out of the player to a different DAC. And not even for DD or any multi-channel, only for DVD-A and SA-CD.
>>And about licence: Don't want to sell this, and I don't care what the industry dislikes to be done. I take the right to use any bit that's on the disk, I paid for it. I don't want to copy anything. Just listen to the recording, with equipment I can trust."
If all you want to do is pull of the digital output data stream, that's not too difficult. You just need to get the spec sheets for the processor that's used. The hard part is wading through the spec sheets (for our project we had about 3,000 pages of spec sheets and application notes printed out). Also, figuring out what type of output mode the DSP is set in would be a trick, as that is done through the uC and a serial programmable interface (i.e. you can't tell by just looking at the schematics of a few mode setting pins like on a CS8412/4). Most of the processors we looked at (Zoran, Motorola, Cirrus, Sanyo, and a few others) had an option for I2S, but as with most DACs, there are several other modes that the user can use.
As for the licensing, Dolby, DTS, Microsoft (now owners of HDCD), and the developers of DVD-A format DO care about licensing. There is a HUGE paper trail of forms to fill out just to get some samples or an evaluation board for Dolby processors. On the flip side, DTS was pretty cool about it for our project. You can't even get an evaluation board without signing 2 or 3 waivers. Dolby and DTS both have a licensing fee just to go through the design process/development with a DSP with their algorithm on it. Then on top of that you have to pay licensing fees for each unit that is sold (when the design eventually does go into commercial sales). Yes, its ridiculous, but that's the way it is. So, for those of you out there who are wanting to get your hands on the actual DSP for DD, DTS, HDCD (an interpolation filter in this case), good luck unless you have a couple $million for licensing as I mentioned before (or a good inside source at Cirrus ).
That said, pquadrat, I'm fine with making it modular. I'm just not sure what you mean by adding a "DSP Module" on it.
Hey, I also just had an idea
! I found a MUX that would be pretty cool to integrate also - it would allow a user to switch the 4 clock and data lines (LRclock, master clock, bit clock, and data) between two different DACs... this would be really sweet for direct comparison. The chip is by IDT, model # QS3384. By having a center off SPDT switch mounted on the case, it would be really easy to implement (I think). Cirrus uses this chip in their eval board for the CS4396 DAC. It allows the user of the eval board to switch between the SP/DIF lines extracted from the on-board CS8414 and an external DSD (Direct-Stream-Digital) input from an SACD player module. Pretty slick, I think. Lemme know if you think this has any merit or not.
>>I do not want to decode the DD signal, just get the already decoded I2S data stream out of the player to a different DAC. And not even for DD or any multi-channel, only for DVD-A and SA-CD.
>>And about licence: Don't want to sell this, and I don't care what the industry dislikes to be done. I take the right to use any bit that's on the disk, I paid for it. I don't want to copy anything. Just listen to the recording, with equipment I can trust."
If all you want to do is pull of the digital output data stream, that's not too difficult. You just need to get the spec sheets for the processor that's used. The hard part is wading through the spec sheets (for our project we had about 3,000 pages of spec sheets and application notes printed out). Also, figuring out what type of output mode the DSP is set in would be a trick, as that is done through the uC and a serial programmable interface (i.e. you can't tell by just looking at the schematics of a few mode setting pins like on a CS8412/4). Most of the processors we looked at (Zoran, Motorola, Cirrus, Sanyo, and a few others) had an option for I2S, but as with most DACs, there are several other modes that the user can use.
As for the licensing, Dolby, DTS, Microsoft (now owners of HDCD), and the developers of DVD-A format DO care about licensing. There is a HUGE paper trail of forms to fill out just to get some samples or an evaluation board for Dolby processors. On the flip side, DTS was pretty cool about it for our project. You can't even get an evaluation board without signing 2 or 3 waivers. Dolby and DTS both have a licensing fee just to go through the design process/development with a DSP with their algorithm on it. Then on top of that you have to pay licensing fees for each unit that is sold (when the design eventually does go into commercial sales). Yes, its ridiculous, but that's the way it is. So, for those of you out there who are wanting to get your hands on the actual DSP for DD, DTS, HDCD (an interpolation filter in this case), good luck unless you have a couple $million for licensing as I mentioned before (or a good inside source at Cirrus
). That said, pquadrat, I'm fine with making it modular. I'm just not sure what you mean by adding a "DSP Module" on it.
Hey, I also just had an idea
! I found a MUX that would be pretty cool to integrate also - it would allow a user to switch the 4 clock and data lines (LRclock, master clock, bit clock, and data) between two different DACs... this would be really sweet for direct comparison. The chip is by IDT, model # QS3384. By having a center off SPDT switch mounted on the case, it would be really easy to implement (I think). Cirrus uses this chip in their eval board for the CS4396 DAC. It allows the user of the eval board to switch between the SP/DIF lines extracted from the on-board CS8414 and an external DSD (Direct-Stream-Digital) input from an SACD player module. Pretty slick, I think. Lemme know if you think this has any merit or not.
For SACD, TI has a couple of DSD dacs. I got a couple of samples of their DSD1700 a couple of weeks ago, but I don't see a samples button for it anymore. Maybe it will come back soon. I have 3 of them, and can part with one if anyone needs one to play around with one.
Datasheet:
http://www-s.ti.com/sc/ds/dsd1700.pdf
What about hacking apart a dvd player to get dd/dts/sacd components? I bought a sony NS500V from bestbuy for $120 (open box) and it has all of those chips inside it. A link to the service manual has been posted on the forum:
http://www.diyvideo.com/forums/showthread.php?s=&postid=95089
Would it be possible to modify this player, upgrading all of the components, instead of trying to make your own with impossible to obtain chips?
--
Brian
Datasheet:
http://www-s.ti.com/sc/ds/dsd1700.pdf
What about hacking apart a dvd player to get dd/dts/sacd components? I bought a sony NS500V from bestbuy for $120 (open box) and it has all of those chips inside it. A link to the service manual has been posted on the forum:
http://www.diyvideo.com/forums/showthread.php?s=&postid=95089
Would it be possible to modify this player, upgrading all of the components, instead of trying to make your own with impossible to obtain chips?
--
Brian
skaara:
>>Thieftoner: So the dac chip and the decoder for dvd-a and dts are in the same chip so you cant pull out the digital data to use it with other dac?
Can we use chips that are already in dts players and took their data for our dac? If so, we dont need millions for licensing..
The processor that decodes DD, DTS, MPEG, etc. typically sends out PCM data streams
- master clock (this may be pulled off of a DIR, depending on the processor and if it is operating in a slave mode)
- frame sync (otherwise known as L-R clock)
- bit clock
- and 3 stereo data streams (for 6 channel surround sound)
This PCM stream can be passed in I2S, left-justified, right-justified, or whatever else the processor supports. Often there are several options for this. The problem is knowing (1) which pins on the processor are what, (2) figuring out what mode the processor is outputting the data streams in (left-justified, I2s, etc). The difficulty in figuring out which mode the processor is in is near impossible because it is programmed through an EPROM or uC serial interface upon boot-up.
For the record, I have no experience with DVD-A processors. I can't imagine there's much of a "processor" though.
If one can gain access to the DACs in a reciever, DVD player with multichannel decoding, preamp, etc... then perhaps one can find out what mode the built-in DACs are running in, thus knowing what mode the main DSP is outputting the multichannel information in. Then, theoretically, one could tap these lines and run to a better designed DAC module.
Theoretically, its shouldn't be all that difficult, such as Dorkus' mods to the Sony NS-500V DVD/CD/SACD player DAC outputs. In this case, he knows that the DACs are made by AKM. With some initiative (not that I'm saying he hasn't had any with his previous mods) he could tap the data and clocking lines that are sent to the multichannel AKM DAC. Using these, he could run it to a Cirrus 4226 or equivalent for experimentation. The information is there on the boards, one just needs to find it.
Hope this helps!
>>Thieftoner: So the dac chip and the decoder for dvd-a and dts are in the same chip so you cant pull out the digital data to use it with other dac?
Can we use chips that are already in dts players and took their data for our dac? If so, we dont need millions for licensing..
The processor that decodes DD, DTS, MPEG, etc. typically sends out PCM data streams
- master clock (this may be pulled off of a DIR, depending on the processor and if it is operating in a slave mode)
- frame sync (otherwise known as L-R clock)
- bit clock
- and 3 stereo data streams (for 6 channel surround sound)
This PCM stream can be passed in I2S, left-justified, right-justified, or whatever else the processor supports. Often there are several options for this. The problem is knowing (1) which pins on the processor are what, (2) figuring out what mode the processor is outputting the data streams in (left-justified, I2s, etc). The difficulty in figuring out which mode the processor is in is near impossible because it is programmed through an EPROM or uC serial interface upon boot-up.
For the record, I have no experience with DVD-A processors. I can't imagine there's much of a "processor" though.
If one can gain access to the DACs in a reciever, DVD player with multichannel decoding, preamp, etc... then perhaps one can find out what mode the built-in DACs are running in, thus knowing what mode the main DSP is outputting the multichannel information in. Then, theoretically, one could tap these lines and run to a better designed DAC module.
Theoretically, its shouldn't be all that difficult, such as Dorkus' mods to the Sony NS-500V DVD/CD/SACD player DAC outputs. In this case, he knows that the DACs are made by AKM. With some initiative (not that I'm saying he hasn't had any with his previous mods) he could tap the data and clocking lines that are sent to the multichannel AKM DAC. Using these, he could run it to a Cirrus 4226 or equivalent for experimentation. The information is there on the boards, one just needs to find it.
Hope this helps!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.