Subscriber Discussion

The Real Inventor Of HD-CVI And HD-TVI?

U
Undisclosed #1
Apr 13, 2015
IPVMU Certified

SingMai Electronics, from Thailand, makes this bold claim on their website:

They call it aCVI though...

Lest one think that this is merely an overzealous OEM, passing off the Dahuan Dynamite as its own, consider the fact that SingMai Technical Director Daniel Oligvie authored this technical paper on the technology in December of 2012, which appears just days after the first press release by Dahua on HD-CVI.

Not only that but this paper is hands down the most detailed and comprehensive document I have ever seen regarding Analog HD. Chock-full of original diagrams, tables and differential equations, it's a welcome addition to the dearth of technical HD-CVI literature.

What do you think?

Did SingMai invent HD-XVI and sell it to Dahua? Or?

Regardless, take a look at Daniel's paper...

U
Undisclosed #1
Apr 14, 2015
IPVMU Certified

Found another document in ee times. If the "company" mentioned is Dahua then it sounds like SingMai was retained to design a single cable analog HD system.

Must have been a work for hire :(

A company told us of an increasing need to be able to transmit HD video over long cable runs; e.g., for security camera installations. In many cases, the cable installation is pre-existing and uses low cost RG-59 cable, but this cable is limited to low frequency use (<200MHz); at 1GHz its attenuation is 28dB/100m. Existing methods to transmit HD video include separate analog RGB/YPbPr, which requires three coaxial cables to transmit, or HD-SDI, a serial digital transmission method which runs at a bit rate of 1.485Gbps and can only achieve small distances with such cable.

Considering this problem, I remembered MUSE (and also the European EUREKA project), which encoded HD video into a single signal. These systems use time-multiplexing, but instead I thought of the Sony Effio approach and we came up with aCVi (Advanced Composite Video Interface), which uses a modified form of the well-known NTSC analog composite video standard to create the signal for transmission [3]. Distances of greater than 300m are achievable and in excess of 500m at 720p/60Hz with some small signal degradation. As with most analog transmission methods, the signal degradation is 'graceful' with no sudden cut-off of the signal that is encountered with digital methods.

Sounds like he's a shoe-in for 'expert technical witness' at any possible Dahua - Hikvision patent litigation.

aw
alon wallach
Apr 15, 2015

To Undisclosed A :

I also found it hard to believe at first, but look at his information package and reference design for an aCVI encoder and decoder, and they are also selling a dedicated test equipment, an HDCVI picture signal generator for 1890$ , something no other vendor offers.

Maybe the truth is that Dahua took the concept and reference design from Singmai, or even from Daniel Ogilivie personally, and made it a commercial success with a big investment, leaving Singmai in the dark with very little or no royalties.

Oligvie might feel offended by not receiving credit on his invention....

Please refer to the HDCVI Crosstalk discussion for more technical details.

And, I would be honored to know your identity or email to you directly : btawire@netvision.net.il

Alon

U
Undisclosed #1
Apr 15, 2015
IPVMU Certified

Oligvie might feel offended by not receiving credit on his invention....

That would be as they say the "Story of his Life"

At assembly the next day although most got mentions for their efforts I didn't get one word. I caught that disease from my Dad, although in my case it was my stroppiness that prevent getting any recognition in life. - Daniel Oligvie's Bio

:(

Maybe the truth is that Dahua took the concept and reference design from Singmai, or even from Daniel Ogilivie personally, and made it a commercial success with a big investment, leaving Singmai in the dark with very little or no royalties.

SingMai was started by Daniel. The way in which it is described it sounds like a 'work for hire', and he didn't think anything of it at the time, except for 'work is work'.

Are you at the show? Once I get my lanyard I will be easily identifiable. ;)

aw
alon wallach
Apr 15, 2015

Good job finding his Bio. He went to Thailand to find some peace of mind with his Thai wife......

He has had bitter experiences as a businessman, but is obviously a first class engineer, although without any formal education.

Put this personally history vis a vis Dahua and Chinese way of doing business, and this is probably the result.

I contacted him to see f he can tell us more. Let's see what he replies.

Alon

LM
Luke Maslen
Apr 16, 2015
IPVMU Certified

Hi Alon, that was a great idea finding and inviting Daniel Ogilvie into this discussion. Thanks very much!

aw
alon wallach
Apr 16, 2015

The man wants to give his version, but he cannot see what we all wrote about it.

"Hello Alon,

I am not a member so I cannot see the discussion.

Kind regards,

Daniel Ogilvie"

Somebody from IPVM : Please allow this man to the discussion without payment. you can send him a temporary user name and password to daniel@danploy.com

U
Undisclosed #1
Apr 16, 2015
IPVMU Certified

Sorry, I had problems with my invites, he should have it now though...

aw
alon wallach
Apr 16, 2015

Thanks.

Looks like Undisclosed A can pull some strings.....

JH
John Honovich
Apr 16, 2015
IPVM

Alon, fyi, we have an invite feature.

DO
Daniel Ogilvie
Apr 16, 2015

The story is a long one and boring to everyone, I am sure.

We are essentially a design house, offering IP cores (proven design blocks) and some modules (mostly to support those IP cores). Our background is high end video, including broadcast and industrial inspection. We were asked to design an ISP for a security camera by a small company in Hong Kong.

New to the security business, I became aware that the HD interface they were using was HD-SDI and given the requirements (long distance, cheap cable) it seemed wholy unsuitable. So I made a proposal to this company for an analogue transmission method which I called HD-CVI. As it was based on NTSC it did not seem to be to be patentable (it isn't) so I proposed it as an open standard and wrote that article. We produced demonstrators for it and showed at at a Broadcast exhibition in 2010 (who also have requirements for long distance transmission).

Lots of other stuff happened, but to to target the Dahua/HikVision/TechPoint relationship only. We had demonstrated aCVi to HikVision but they showed little interest from their management. We were then made aware of Dahua's release of HD-CVI and we were told it was essentially a copy of our proposal. We have never had any (or very little) contact with Dahua. Their HD-CVI is their own, but the principals behind it are the same as our proposal. We renamed HD-CVI to aCVi to avoid any confusion.

Dahua then filed for patents which, as a (very) long experiened video engineer, are nonsense, and also we could prove prior art. But for us to challenge the patents in China was a no-win situation. We did work briefly with another company to support them challenging Dahua's patents but they did not pursue it.

TechPoint bought the aCVi IP cores from us, but we have had no contact with them since and they have not used us as consultants.

We decided not to continue selling aCVi into the security market - or at least make no active efforts to do so. We always designed aCVi to broadcast specifications (SMPTE), we only have a discrete solution (no IC) and therefore we concentrate on other long distance transmission markets which have more margin.

(By the way, not that it matters, but I do have formal engineering qualifications. :-))

I am happy to answer any other questions anyone may have - but I do not want to be drawn into a 'who is better' war. The answer to that is aCVi - obviously!

(And by the way - watch out for a new 'aCVi' in about 5 month's time. We won't be publishing any technical information this time!).

(11)
LM
Luke Maslen
Apr 16, 2015
IPVMU Certified

Hi Daniel,

Thank you for joining this discussion and for your fascinating story. While you may not have benefitted as much as one might hope, I am most grateful for what your technology has made possible. It has provided high quality, reliable and affordable CCTV solutions to many who otherwise could not have afforded it. I look forward to developments with your new aCVi.

My previous job for a decade was in broadcast video and film but I don't recall hearing anything about HD-CVI or aCVi. Can you tell me where the technology appeared, e.g. in some specific products? I'm puzzled as to why I had not heard about the technology before it showed up in CCTV cameras. Thank you for your help.

aw
alon wallach
Apr 16, 2015

Wow... Multiple inventors story again.

My personal belief is that this thing was coming, and most probably was thought and designed in parralel by more than one team.

Daniel : Thank you for disclosing your side of the story.

My interest is less on the patents and market side, and more on the technical side.

I bet you can easilly outline the major technical differences between HDCVI, aCVI, AHD and Hikvision. Are they all having the more or less same pre-emphasized Luma channel with used spectrum of DC to 20-30 Mhz, and then a subcarrier for the Chroma well outside this band ?

And, how is the handling ( manual or automatic ) of cable slope done in all 4 systems to reconstruct the correct de-emphasized composite video ?

Alon

LM
Luke Maslen
Apr 16, 2015
IPVMU Certified

Hi A, well done on discovering this information and joining the dots. This has been a very interesting insight into the world of analog HD cameras.

(1)
DO
Daniel Ogilvie
Apr 16, 2015

I don't want to trivialise this - after all it is a brilliant invention! - but basically, the idea was, NTSC/PAL can be transmitted long distances with maybe a little receiver high frequency compensation. There was no HD equivalent to this, the only analogue HD was YPbPr needing 3 cables. So it was obvious to create a HD version of NTSC. Sony already had this idea with 960H - we just took it further. As security is a closed system we did not have the problem of making a universal standard.

However NTSC had a couple of problems, crosstalk because the subcarrier was within the luma bandwidth (because of legacy reasons), and the receiver high frequency compensation (usually an analogue network). So for aCVi we separated the luma and chroma (moving the chroma subcarrier above the luma bandwidth) and applying transmitter pre-emphasis (similar to the J17 curve for radio transmission of NTSC). Nothing new, just a new application for old ideas.

I believe the first implementation for Dahua still used receiver high frequency peaking. Instead, for aCVi, we used transmitter pre-emphasis that was controlled by a low frequency bi-directional data link in the VBI of the video (similar to closed caption but bidirectional). The receiver measured the high frequency loss and sent data to transmitter to control the pre-emphasis. The pre-emphasis curve was matched to cable type so at the receiver we only needed to apply flat gain, no analogue frequency networks.

Dual developments? Perhaps. And at the end a Chinese company, whose market is in China, chose to use exactly the same name as we gave the technique - a Roman name - two years before. A remarkable conicidence.

(1)
(3)
aw
alon wallach
Apr 16, 2015

Daniel.

Dahua is selling all over the world, not only in China.

It does not matter if they sell most of their stuff in China. If they applied for a US patent over the HDCVI and you have proof that you started working on it before they did and even had correspondence about this work, then you can kill their patent on the prior art clause, and this alone is worth a lot of money to them.

If you feel robbed, you can most probably arrange a US attorney to represent you and to demand Dahua USA to pay you or else, and he will do it on a contingency basis. If you look carefully, unlike their Chinese parent company, their website does NOT contain HDCVI, but it is surely distributed by them all over the country. Its not an accidental ommission. They know the heat is coming..... And don't underestimate the HDCVI sales in the US. Its picking up and they are in the lead.

On the technical side, I read your explanation about the pre emphasize calibration in your design and Dahua. Both methods might be good enough to set the high frequency level at one particular frequency, with hope that it will be correct for all frequencies, but I doubt if this will hold for all types of cables with different DC losses, and more importantly with different slopes. I see this as the main issue for all analog HDs. There is another issue of crosstalk of the color signal between two cameras due to color carrier leaking from one cable to a parralel one ( especially with CAT5 ). In old fashion multi channel cable TV the NTSC worked well because each channel was AM modulated on a different frequency. In your design however all cameras are just giving baseband composite signal, and the color subcarrier of all cameras will therefore be on the same RF frequency, be it 40, 45 or 58 Mhz. Crosstalk can also appear due to ground issues in DVR side. Remember that this signal is the one overgoing the highest attenuation, and therefore prune to noise.

Can you relate to these technical issues.

Thanks

Alon

(1)
U
Undisclosed #1
Apr 16, 2015
IPVMU Certified

If they applied for a US patent over the HDCVI and you have proof that you started working on it before they did and even had correspondence about this work, then you can kill their patent on the prior art clause, and this alone is worth a lot of money to them. If you feel robbed, you can most probably arrange a US attorney to represent you and to demand Dahua USA to pay you or else, and he will do it on a contingency basis.

Alon, I'm sure that Daniel appreciates your enthusiasm and validation. I am also pretty sure that Daniel understands the IP issues surrounding aCVI better than us, and has likely spent a night or three contemplating these issues.

Specifically, if Daniel's gut is right and aCVI is not patentable, then he would not be able to patent it either, right? And even if he could prove prior art and have their patent overturned on that basis (no small feat itself), if he could not then protect it, he would just be a spoiler for Dahua.

Furthermore any "demand to pay him or else" fails because paying Daniel does not prevent anyone else from overturning the patent on prior art.

Finally the fact that Daniel has not filed, AFAIK, for this patent, and at the same time openly published detailed specifications will complicate a new claim at this later date.

Just saying that 'get lawyer on contingency and kill their patent' is something he has no doubt considered already.

aw
alon wallach
Apr 16, 2015

To Undisclosed A .

Specifically, if Daniel's gut is right and aCVI is not patentable, then he would not be able to patent it either, right? And even if he could prove prior art and have their patent overturned on that basis (no small feat itself), if he could not then protect it, he would just be a spoiler for Dahua.

Neither of us has any proof of working on the issue at earlier times than Dahua's filing date, but Daniel probably has, so he is the only possible spoiler for the Dahua patents, and they claim of a series of patents.

From my experience, many patents are accepted by the examiner, simply because there is no other patent in the field, and the patent office database and search engine is not really focusing on Google or asking people in the trade if this or that patent sounds familiar...

There is a very famous Austrian patent examiner that was undoubtedly a very bright guy, but was not a good examiner. He later won a Noble prize for the photoelectric effect although he never connected a wire in his life or screwed a screw. They are smart, but not necessarily knowledgable...... It is very difficult to master so many fields. I can easilly see the examiner approving a patent of an extended NTSC encoder ( especially that many of the examiners in the USPTO are with Chinese names ), merely for an inclusion of a different modulation or for an automatic de emphasize circuit, and it will pass.

I have registered about five patents by myself and seen it first hand. 95% of the patents are repetitions and bullshit, but the industry is worshipping the patent concept and gives it great respect, while the USPTO makes money on submissions, renewals etc.

A possible spoiler in this case can get rich !!!!!

Alon

U
Undisclosed #1
Apr 16, 2015
IPVMU Certified

Neither of us has any proof of working on the issue at earlier times than Dahua's filing date, but Daniel probably has, so he is the only possible spoiler for the Dahua patents, and they claim of a series of patents.

Alon, this what Daniel said above:

So I made a proposal to this company for an analogue transmission method which I called HD-CVI. As it was based on NTSC it did not seem to be to be patentable (it isn't) so I proposed it as an open standard and wrote that article. We produced demonstrators for it and showed at at a Broadcast exhibition in 2010.

To begin with there are two spoilers in Daniel's mind, his invention as well as NTSC itself. His invention is open to be used by anyone, since it was publicly demonstrated before Dahua's filing.

I'm not saying that Daniel would be wrong to pursue something, I'm just saying it's not clear to me what the basis would be at this point.

What Daniel mentioned above though does solve one mystery: Why Techpoint/Hik is unconcerned about Dahua's patent rumblings...

DO
Daniel Ogilvie
Apr 16, 2015

To fight Dahua on their home turf was never going to turn out well for us - the fact they made no attempt to disguise the origins of their 'invention' shows their certainty on that outcome.

The core of the patent filing is the separation of the chroma and the luma (I only saw the original application before filing so I am unsure exactly what was allowed). This is clearly nonsense as there are multiple example of non-overlapping subcarriers on a single cable interface. Also any patent officer worth his salt should have been able to Google HD-CVI and find us. Personally I doubt they will file in the US, because of the likelihood of it being refused and it is not their principal market.

Personally I don't want to be embroiled in this so I put it down to experience and move on.

To answer the technical point above - originally we designed our pre-emphasis filter to match the response of RG-59 cable. Later we added other curves for other cables, but yes, ideally the filter should match the cable type. It is still better than adding high frequency compensation at the receiver because of better SNR and we can have variable compensation whereas the discrete analogue compensation can only be done in steps.

aCVi was really only intended for coaxial cable. Then we got requests for UTP cable. If multiple signals were to be sent along the same twisted pair cable then we could change the subcarrier for each signal to reduce crosstalk. For our present markets we do not have that issue.

For UTP cable we drive and receive the signal as true differential, for coaxial we receive the signal as pseudo-differential so we can reduce low frequency crosstalk or hum.

As I mentioned above, aCVi achieves 30MHz luma bandwidth at 300m of RG59 cable. I see the nonsense over distance claims and quietly tut to myself. Distance without a specification is worthless (although clearly not!) so when we claim a distance we state a SNR and pulse response and bandwidth/flatness. That is why we have an aCVi pattern generator, to measure the performance (and soon a waveform monitor too). And that is also why we left the security market (in China at least).

(2)
U
Undisclosed #1
Apr 16, 2015
IPVMU Certified

Thank you, Daniel, that was clear.

A few questions:

Are you at liberty to say who the initial Hong Kong company who you worked with is? If not, are they a known security company?

Did you do the detail design of the upstream VBI camera control commands and grammar?

Are you familiar with AHD, a variant from Nextchip in Korea, is it the same basic thing?

When you sell an IP (Intellectual Property) core, like you did with Intersil/Techpoint, what rights does that typically convey?

Is there any hope that these analog QAM technologies can be extended into the 4K range and beyond?

Thanks!

Background: (please forgive us in some of these discussions for our ignorance, there was very little in the way of documentation.)

HD-SDI Vs HD-CVI Vs HD-TVI Vs HDcctv?

Hikvision HDTVI VS Dahua HDCVI

Testing Analog HD

The Company Behind HDTVI

Dahua Attacks Hikvision

What Is Analog HD (AHD) - Competitor To Dahua And Hikvision?

DO
Daniel Ogilvie
Apr 17, 2015

The Hong Kong company are not a security company: they are a design house too. They were designing a complete security camera for a China company, but I do not know who.

We use the control interface for the cable length compensation. It is a simple instruction/data format which the customer can use himself - we have not prescribed a fixed instruction set. We are likely to make our own aCVi camera at some point (we already have an aCVi camera module) and we will then make use of that control bus ourselves (e.g. pan/tilt/zoom or image sensor control functions).

I am not familiar with AHD although I am aware of it. I have seen the letters from Dahua to/from Nextchip. I heard from customers at SecuTech last year it had the worst performance of the 3 competitors.

TechPoint are not Intersil. Intersil bought Techwell and the two founders of Techwell then formed TechPoint. We sell our IP cores as Verilog source code for a one off payment that gives a license to the user for unlimited time/unit use. They only bought the digital part of the transmitter/receiver - the analogue part is integral to its performance. Our analogue is discrete and I suspect their problems are making an ASIC of that analogue.

We offer 1080p/50-60Hz for aCVi but at a reduced bandwidth (50MHz) for the luma (compared with the SMPTE specification). This is to make the analogue components affordable. So in theory it can be extended but the cable length would be restricted. To be honest, it would be better to just make 960H or 1280H work properly.

I am happy you ask the questions. Analog HD can never replace HD-SDI or 3G. But given the poor quality of the sensors in security it is adequate and it does solve the problem of driving long cable lengths (and the poor quality cable used). Whilst I understand the cost issues it seems to me it would be better to spend a little more money and design the whole system properly. A lot of the 'real' images I see are unfit for purpose in my view. I have been to many Chinese 'manufacturers' and they have no test equipment at all. Most just buy in a sensor module and stick it a box. Four of these companies can be in the same building, each shipping 10k cameras/month.

HikVision had aCVi one year before Dahua had HD-CVI but their management failed to see the opportunity (their engineers did). I think this is because HIK see the future as IP cameras and analogue seemed a backward step. When they started to play catchup they went to TechPoint because of a past relationship with TechWell (and maybe they don't like Thai companies!).

But if the image is too poor to recognise the intruder/thief whatever, well it seems a dummy camera would do a better job. But that is my view from the luxury of a market that actually has a profit margin.

If you (the West) see opportunity in analog HD, my own view would be to find a company that could design an interface for you (:-)) that circumvents Dahua's patents (which are unlikely to be valid anywhere except China anyway) and do it yourself. If you know a US camera manufacturer interested (yourself?) then please let me know.

(1)
(4)
UE
Undisclosed End User #2
Apr 17, 2015

A bunch of Executives and Engineers, around the world, are now salivating. I can even see 100% pure IP camera companies selling HD CVI cameras down the road as part of their lines.

Avatar
Kyle Folger
May 01, 2015
IPVMU Certified

I'm glad I simply did a search for CVI encoders, which I normally do to see if there is anyone talking about them and what the advantage might be compared to simply using the CVI recorders as ONVIF encoders. Daniel, thank you for your awesome and insightful posts. They have been very informative. As to this:

The story is a long one and boring to everyone, I am sure.

It couldn't have been too boring, I read the whole thread.

(2)
Avatar
David Johnson
Apr 18, 2015

Well Mr. Ogilvie this may be the most salient comment of this entire discussion

"But if the image is to poor to recognize the intruder/thief whatever, well it seems a dummy camera would do a better job."

I've been involved in the industry since vidicon tube days and never cease to be amazed at how many well intentioned and sometimes shockingly expensive design solutions have fallen woefully short of this mark.

Every aspect of image capture/delivery/storage are equally important in acheiving what the customer ultimately finds useful and are all throttled by the budget available.

I have to say that when taking the end result of the above equation into full context, I have been mightily impressed by the recent few solutions our firm has managed to deliver with an HD-TVI version of your creation and offer my kudos to you for an inspired bit of practical thinking.

Thanks as well to IPVM for rooting this out - no doubt some of the above revelations regarding cable choice offer clear warning signs when considering suitabilty of existing cable plant.

UM
Undisclosed Manufacturer #3
Apr 19, 2015

For those of you interested to see a different analog ( super ) HD solution ( actually just a camera, as it works with the existing legacy analog hardware ) without cable limitations and very modest storage requirements over a standard old fashion analog DVR , while still enabling to recognize the intruder / thief, please take a look at the following patent http://www.google.com.ar/patents/US20080002045 . There is also a low resolution slow motion demo video of the concept at http://www.wireless-avionics.com/images/129804-Hawk-I.wmv . It is based on the understanding that continuous high resolution is usually NOT mandatory, and real time coverage can also be compromised with a fraction of a second break every several seconds. The result is a 100% standard D1 resolution video stream, with 5 Mpixels still image ( 2.5 times better than all analog HD systems with optimal cable and calibration ) every several seconds, that can be easilly reconstructed to a perfect Mosaic from the D1 recording, or just viewed slowly frame by frame to see 16 subpictures, precisely making this 5 Mpix mosaic. It can also work on lower resolution still picture of 2.5 and 1.2 Mpixels, and then the "break" is decreased to just 4 video frames or 1/4 of a second.

Do you think such camera and concept can find its place in the market after all the analog HDs that popped like mushrooms in the last year or so ?

RH
Richard Hemzal
Jul 22, 2019

Daniel can you please comment technical solutions which was used later in analog HD systems to overcome 1080p? It looks they shifted luma and chroma into higher bandwidth when increased resolution into 3 or 4Mpix. Later they reduced frame rate to reach 5,6,8 Mpix.
Your insightful comment really welcomed.

(1)
UM
Undisclosed Manufacturer #4
Aug 09, 2019

+1

This is an excellent thread and would love to see it updated.

New discussion

Ask questions and get answers to your physical security questions from IPVM team members and fellow subscribers.

Newest discussions