Technical Design - Decision #001 - Network Interface via open hardware, Ravenna vs WhiteRabbit
62 views
Skip to first unread message
Alexcount
Jun 1, 2012, 11:55:41 AM6/1/12
to open...@googlegroups.com
Greetings and Welcome to the Open-NGDI Development Message Board,
We are in a very early stage of development. Technical specifications and major hardware choices are still being finalized.
I want to discuss a new approach to dealing with Digital Audio in a studio or laboratory situation, in addition showing two possible standards using this approach to transmit data over Ethernet/fiber with accurate synchronization over large areas. These are both "open hardware" projects, but they differ in the licenses used.
The Merging HORUS Converter, still in development, supports the recently announced Ravenna Open Standard, which utilizes Gigabit Copper Ethernet to transmit data, allowing many new interesting ways to wire and route digital audio via standard equipment. This has been demonstrated with compatible tech from other companies, including a pair of Genelec Studio Monitors featuring Ethernet. The Ethernet switch used in demo's was not specialized for Ravenna, but rather a standard gigabit network switch.
The WhiteRabbit network has been released by CERN in Geneva, on the Open Hardware Repository (OHWR.org) website. It promises high fidelity of synchronization of many nodes (1000+), and support for large runs of cable (10km+), over either copper or fiber Ethernet. This has been released under the CERN OHL Open Hardware License. It was developed not for audio, but for data from the large amount of sensors required at CERN for various scientific projects and experiments, many of which involve the Large Hadron Collider, the worlds largest particle accelerator. With the large facility required for the LHC, the WhiteRabbit network is a vital part of interpreting data properly from the experiments with great timing fidelity.
Cited from WhiteRabbit Project:
"White Rabbit is a fully deterministic Ethernet-based network for general purpose data transfer and synchronization. The aim is to be able to synchronize ~1000 nodes with sub-ns accuracy over fiber and copper lengths of up to 10 km. The key technologies used are physical layer syntonization (clock recovery) and PTP (IEEE 1588)." http://www.ohwr.org/projects/white-rabbit
Advantages of WhiteRabbit
* VHDL Code for programming an FPGA available at no cost.
* Generous overhead for current requirements, allows further expansion beyond current goals
* CERN OHL is a robust open hardware license.
* Supports any novel formats of data such as 5, 6 or 8 bit at 25~Mhz, 12-bit at 100Mhz, etc.
* Supports Fiber or Copper Ethernet.
* Generous overhead allows expansion of channels/resolution for uses beyond audio frequencies, allows utility for ultrasonic/RF.
* Would be the highest specification for a audio system as far as synchronization to bit accuracy at beyond DSD rates.
Disadvantages
* Might Require expensive dedicated hardware for switching for maximum specifications
* cost might be prohibitive if requirements are too demanding otherwise.
*If using Fiber Optics*
* Fiber optics are not standard equipment on most computers or laptops.
* Fiber optic cables must generally be purchased pre-made, making cables/splicing requires expensive/specialized equipment.
* Cost is generally much greater than copper.
Advantages of Ravenna
* Designed with Audio in mind, supports standard audio formats already.
* Cost probably lower as commercial constraints were considered in design.
* Other hardware and software vendors will (hopefully) support the standard, allowing interoperability with other hardware.
Disadvantages
* Would need modification to support high bandwidth formats such as Sigma-DPCM at ~25Msps
* Does not support fiber optics, limiting maximum cable run (still very acceptable at 3~~??km)
* FPGA code not obviously/immediately available, open specification might require additional development.
* Though referred to as an open standard, no mention of what license is used is available AFAIK.
These two technologies are very similar in both technical specifications and implementation. They differ in two key ways; their licenses are different, and they each were developed with different requirements. By using the WhiteRabbit network, which is admittedly overkill for most audio formats, exponentially greater bandwidths will be available if needed. This allows for future applications such as SDR, RADAR, GPR, Radio Astronomy, Psycho-acoustic and Hypersonic Acoustic studies. Synchronization of samples of under 1 nanosecond over networks spanning up to 10km is enough overhead to allow almost any complex studio or laboratory application.
Let's start a discussion. What are your opinions on these two similar technologies and the licenses they would require?
-Alexander Countey
We are in a very early stage of development. Technical specifications and major hardware choices are still being finalized.
I want to discuss a new approach to dealing with Digital Audio in a studio or laboratory situation, in addition showing two possible standards using this approach to transmit data over Ethernet/fiber with accurate synchronization over large areas. These are both "open hardware" projects, but they differ in the licenses used.
The Merging HORUS Converter, still in development, supports the recently announced Ravenna Open Standard, which utilizes Gigabit Copper Ethernet to transmit data, allowing many new interesting ways to wire and route digital audio via standard equipment. This has been demonstrated with compatible tech from other companies, including a pair of Genelec Studio Monitors featuring Ethernet. The Ethernet switch used in demo's was not specialized for Ravenna, but rather a standard gigabit network switch.
The WhiteRabbit network has been released by CERN in Geneva, on the Open Hardware Repository (OHWR.org) website. It promises high fidelity of synchronization of many nodes (1000+), and support for large runs of cable (10km+), over either copper or fiber Ethernet. This has been released under the CERN OHL Open Hardware License. It was developed not for audio, but for data from the large amount of sensors required at CERN for various scientific projects and experiments, many of which involve the Large Hadron Collider, the worlds largest particle accelerator. With the large facility required for the LHC, the WhiteRabbit network is a vital part of interpreting data properly from the experiments with great timing fidelity.
Cited from WhiteRabbit Project:
"White Rabbit is a fully deterministic Ethernet-based network for general purpose data transfer and synchronization. The aim is to be able to synchronize ~1000 nodes with sub-ns accuracy over fiber and copper lengths of up to 10 km. The key technologies used are physical layer syntonization (clock recovery) and PTP (IEEE 1588)." http://www.ohwr.org/projects/white-rabbit
Advantages of WhiteRabbit
* VHDL Code for programming an FPGA available at no cost.
* Generous overhead for current requirements, allows further expansion beyond current goals
* CERN OHL is a robust open hardware license.
* Supports any novel formats of data such as 5, 6 or 8 bit at 25~Mhz, 12-bit at 100Mhz, etc.
* Supports Fiber or Copper Ethernet.
* Generous overhead allows expansion of channels/resolution for uses beyond audio frequencies, allows utility for ultrasonic/RF.
* Would be the highest specification for a audio system as far as synchronization to bit accuracy at beyond DSD rates.
Disadvantages
* Might Require expensive dedicated hardware for switching for maximum specifications
* cost might be prohibitive if requirements are too demanding otherwise.
*If using Fiber Optics*
* Fiber optics are not standard equipment on most computers or laptops.
* Fiber optic cables must generally be purchased pre-made, making cables/splicing requires expensive/specialized equipment.
* Cost is generally much greater than copper.
Advantages of Ravenna
* Designed with Audio in mind, supports standard audio formats already.
* Cost probably lower as commercial constraints were considered in design.
* Other hardware and software vendors will (hopefully) support the standard, allowing interoperability with other hardware.
Disadvantages
* Would need modification to support high bandwidth formats such as Sigma-DPCM at ~25Msps
* Does not support fiber optics, limiting maximum cable run (still very acceptable at 3~~??km)
* FPGA code not obviously/immediately available, open specification might require additional development.
* Though referred to as an open standard, no mention of what license is used is available AFAIK.
These two technologies are very similar in both technical specifications and implementation. They differ in two key ways; their licenses are different, and they each were developed with different requirements. By using the WhiteRabbit network, which is admittedly overkill for most audio formats, exponentially greater bandwidths will be available if needed. This allows for future applications such as SDR, RADAR, GPR, Radio Astronomy, Psycho-acoustic and Hypersonic Acoustic studies. Synchronization of samples of under 1 nanosecond over networks spanning up to 10km is enough overhead to allow almost any complex studio or laboratory application.
Let's start a discussion. What are your opinions on these two similar technologies and the licenses they would require?
-Alexander Countey
Alexcount
Aug 16, 2012, 2:11:00 PM8/16/12
to open...@googlegroups.com
Video would actually be within the capability of the WhiteRabbit system, even though the complexity of the whole increases, I think the base project will be much the same. I had not thought of adding it, but it would make it even more general purpose, and would be a good exercise in taking other novel formats of data. Ideally a system for video would have far greater bandwidth, as video formats such as uncompressed 16-bit 4K video takes amazingly high data rates, and would increase cost prohibitively. Compressed video, at 2k and lower resolutions would certainly be possible, and quite handy.
I have planned to make this adaptable enough to support almost any format, with the hopes if the project is broad enough, it will spark interest in a wider range of people. Adding video capability could be a very useful addition for studios, broadcast, scientific researchers, schools, etc. And a high quality video link attached to high speed data streams could be very handy as synchronization between them would yield more high quality data.
One of my grander goals of this project, is to see how many sensors and ADC's can be packed into a confined space, to create a "TriCorder" type device, capable of "seeing" the world with greater depth than a person can "see". The combination of many tools and instruments into one could revolutionize how scientific research in the field is conducted, and if mass produced could lower the cost of these instruments through economy of scale. Imagine a device to visualize, reocrd and analyze not only sound frequencies above and below what we can see, but electromagnetic waves from rf to gamma radiation, and everything in between including IR and UV. With the right software, the device could anylyze color spectra, species of bats, and ideally analyze any chemical. Such a device was science fiction, but seems closer to reality every day, and can already be realized in a more primitive form.
-Alexander Countey
On Thursday, August 16, 2012 1:34:35 AM UTC-4, Marco ter Bekke wrote:
I have planned to make this adaptable enough to support almost any format, with the hopes if the project is broad enough, it will spark interest in a wider range of people. Adding video capability could be a very useful addition for studios, broadcast, scientific researchers, schools, etc. And a high quality video link attached to high speed data streams could be very handy as synchronization between them would yield more high quality data.
One of my grander goals of this project, is to see how many sensors and ADC's can be packed into a confined space, to create a "TriCorder" type device, capable of "seeing" the world with greater depth than a person can "see". The combination of many tools and instruments into one could revolutionize how scientific research in the field is conducted, and if mass produced could lower the cost of these instruments through economy of scale. Imagine a device to visualize, reocrd and analyze not only sound frequencies above and below what we can see, but electromagnetic waves from rf to gamma radiation, and everything in between including IR and UV. With the right software, the device could anylyze color spectra, species of bats, and ideally analyze any chemical. Such a device was science fiction, but seems closer to reality every day, and can already be realized in a more primitive form.
-Alexander Countey
On Thursday, August 16, 2012 1:34:35 AM UTC-4, Marco ter Bekke wrote:
Hi Alexander,
Let`s start stating that this is a rather ambitious project, with many hurdles to overcome in both hardware, software etc. In my eyes also an excellent place to aim for a final solution for digital audio. What makes it special is the goal to also make it work for scientific purposes. In that case, white rabbit is the only way to go. What might be even more logical to me, is the combination of audio and video. I guess a lot of broadcast companies might love a solution for that. What do you think?
Alexcount
Aug 16, 2012, 2:14:51 PM8/16/12
to open...@googlegroups.com
Also, Video resolution could be exchanged for frame rate and keep the same data rate, this would allow 8k resolution and beyond at low frame rates.
-Alexander
-Alexander
Reply all
Reply to author
Forward
0 new messages