4:4:4 12-bit Uncompressed DVX100

[Wayne Morellini]

I understand where your coming from too, if a camera has re-programmable (even tapping in a new rom) stages (no just a processor but several processor, DSP's and DSP compression stages, which need to be bypassed if they are not reprogammable). For a tech centre to propperly test a sensor for fault they also would need a way to check the whole frame uncompressed. Then you can get that information, even if you get some abstraction of it (say 8 bit before comrpession but with better routines, then you are doing much better than the straight camera). In this modern day and age I would not be suprised if a number of cameras use reprogrammable stages all the way, and nearly all of them use at least one reprogammable stage. Maybe this camera is one that has it wired down propperly, but that doesn't mean all cameras are like that. It also maybe cheaper to tap and the signals (realisign that there will be light loss from splitting the signal and maybe some extra niose) but for the end user the process is fair bit more excpensive. Of curiosity how much is the conversion now?

The best thing is that camera hacking is uncommon, so some manufacturers may not have taken enough precautions against it. Just a thought for anybody out there that wants to do it. I'll just leave it there.

Thanks

Wayne.

[Juan P. Pertierra]

I haven't had experience with all cameras, but I do know how all the prosumer Sony cameras are setup (PDX10, PD150, PD170, etc) and even all the consumer models because i have access to all their service manuals with full schematics. I'm also extremely familiar with the DVX100/A.

There is simply NO WAY of doing what you say with just repogramming software in any of these cameras, period. The problem is that it is NOT cost effective for the company to use all programmable stages when they can purchase set-hardware IC's that are extremely cheap and do the job. You CAN'T reprogram these IC's.

And BTW, what is more cost effective for servicing: writing tons of design software and modifying the entire circuit to allocate for service features such as outputting uncompressed frames

OR

simply probing the chip like i'm doing and getting raw frames anyway.

There's just zero cost overhead with the last method. They have no reason to make it more complicated. It's a lot cheaper for them to just replace the CCD without even testing it.

And i'm not really sure what you're talking about light loss from 'splitting a signal' and 'extra noise'. Digital signals don't suffer any degradation when being probed directly, that's the whole point. These are basic concepts.

Juan

[Wayne Morellini]

Sorry if I have annoyed you Juan, I said I would leave it at that last time. I think it is cheaper to design with reprogrammable stages that are at least, poked different values for different cameras' different feature sets, if they don't they are probably either making them unhackable or still stuck in the stone age. Either way I'm surprised. And I wasn't talking about digital signal degregation.

What's the pricing on the mod now?


Thanks

Wayne.

[Régine Weinberg]

Dear Wayne: as Juan told you there is no way other as to take literally the raw data out.

Sony Germany and my guy there I do know since 25 years, here is nothing hidden in a Prosumer or Consumer or Pro cam. They do not have a modul approach even if you don't believe it.

The only cam, Pro, I do know is the Ikegami edid Cam, since 5 years out, writing direct to disk and came from an Avid Idea as ENG Cam. Never a big bang as ENG People and the rest are tape wise.

There was a small Company in Berlin, they build the same, but as they had no big name and no marketing staff they sold only a few and are now Ikegami dealer. Maybe Ikegami is the only one even in the Pro scene, with a modul aproach anyway.

You wont believe, look in a consumer cam as a XL xx if a controller for a 5 axis welding machine would be build this way no service and no time between failure !!

[Wayne Morellini]

I have accepted Juan's explanation, though I have doubts, and suprised at the Camera industries backwardness (which must cost them to make different IC's (=small chip runs) for each camera rather than have a general reprogrammable (which canbe setup through on chip registers) parts that can be shared accorss the millions of cameras sold, rather than hundreds of thousands or less for individual models).

No wonder they make so little money on something that should cost half the price (allways wondered where they wasted that money). But I lack the time to examine servicing manuals, in detail, or to get and detail look at all the proper datasheets, that is what is really needed, so I'll leave it at that.

It's is no wonder the projects are coming in at such cheap prices compared to the camera companies. Inevitably small consumer electronics companies are going to send them virtually broke. Look at those sub $200 Solid State MPEG4 cameras and realise that, in effect, the physical performance difference (please note what I exactly mean here) between them and a proper camera is narrowing fast.

Sony might have to rename themselves, PLAYSTATION INC, just to survive . So I can't help camera companies ignorance, serves them right for doing us over for years.

[Juan M. M. Fiebelkorn]

Wayne:

In fact the ENG market is a very lucrative one.Lots of cash flow just to get the latest fastest thing available....
Also just think about the cost of a pro Video Zoom lens.How much of those 50,000 bucks do you think goes for the manufacturing cost?

What is not so lucrative on the camera side is the consumer market, that is why they get their money from tape, memory sticks and the like.

Most of the camera's cost goes to: Lens system , housing and tape mechanism.
IC's are in fact very cheap for big companies which have their own foundrys in most cases.Also take note that many many different models share the same IC's.
For example DV compressor is the same for your handycam and for an ENG style camera..

[Juan P. Pertierra]

It's not a matter of backwardness or not. It's just a matter of circuit design. There are very few circuits out there that use ONLY programmable logic! Most use a combination of cheap IC's and programmable IC's (either ASICS or microprocessors/controllers). You can't just 'jumper' and bypass these set hardware units.

Let me be more specific. The firewire link layer on the DVX100 is a set chip. Right there, your idea is dead. Because the DV stream functionality of firewire is built into the set circuitry in this chip, you'd have to have a competely programmable circuit connected to that IC in order for you to program your own customized firewire protocol for the raw data. but guess what, that chip is actually(among others) hooked up to the MPEG compressor layer IC, which is ALSO set circuitry. You get the picture?

This is not just the DVX, I know for a fact all Sony consumer and prosumer cameras are setup in a similar way, and because it is such a standard way of setting up these kind of circuits, i'm sure almost every other camera is setup like this. This is not backwardness, or mediocrity, this is THE WAY this kind of circuits are designed. Why would you ever sit down to design an A/D, or a firewire layer, or an MPEG layer IC, when you can just purchase one? manufacturers only use programmable logic for what they NEED programmable logic. Everything else they would be wasting money. And they are right in doing so.

Juan

[Wayne Morellini]

So they do use shared IC's, as I guessed, but it seemed to be implied that they were each with different ones, which would lead to short runs and high prices, and you know how much third party IC's are used in CE. So how come cameras have different controls and features if they use the same IC. I would guess it is that features are selected for the model, would that be right, by on chip registers, would that not be the case in a stratified market like this in many different electronics products. So that is programmable. It occured to me a while ago that what I meant by progammability, you didn't mean. I let it be known that the register stuff was included, and shared memory, that another processor can get at before processing.

So propper saving is using a lrage volume IC that can be configured in different ways, usingf short run IC's would be backwards. Now the moral question, not stoping people from modifying a camera (through design or otherwise) and giving the most features for a price is the moral thing to do, otherwise is backwards. So that is the angles I was responding to, and what I mean by backwards.

Now on camera pricing, how muchg more do you think it really costs to manufacture most of these cameras compared to a modern $249 VHSC camcorder, all those flexible prices, like the GS120 that is around $500 at the moment, exactly, we are hoodwinked. Where ever the profit is being made by the IC manufacturer on higher performance parts, or the camera maker, or even retail end, there is significant profit. The real cost difference between the low performance parts and the higher perforamnce parts might only be upto double, the case, lense, sensor, block (as panasonic obviously has done) can be mass produced cheaply. This stratification is market driven economies at work. It is not until you get to the high end, or brand new catergories (like even the JVC HD1/10 was when it was first released) that you are likely to see truely high cost parts to get the maxium perforamnce. Is it a conspiracy? Well, I leave that up to others than me.

Now Jaun, are you saying that the firewire chip, is not a chip that can be setup to accept different format of data, like the firewire standard provides for? Do you concure, that through reconfiguring setup, and/or command, registers, and accessing by the controll processor to the data through shared memory,l that it might be possible to gather and pass raw output to the firewire that is designed to operate between 12.5MB/s (as used in old cameras) to 50MB/s?

It is a matter of design, and with the examples of real world camer reprogrammability I have mentioned before, I think it is highly unlikely that it is impossible on all cameras out there. But this does require significant time and effort as you suggest.


I am glad to have the chance to debate the two Jauns at the same time, but lets just leave it there and move on. I have a new camera you might want to try, I'll wait to I'm certain they are on the market, and can't be made incompatible. I am certainly impressed with the results from the DVX100 compared to normal, and compared to what we have been getting with the Micron 1.3MP based Cinema camera projects. I imagine that cheaper chips could even have enough get up and go to match the Micron based camera. Would you agree? So what do you think, would a $500 GS120 3 chip, or the 400, be able to match the Micron, or a IBIS5? None of these are the camera I was talking about, but are an interesting discussion, especially if they have pixel shift and the mod canbe done cheap enough to warrant it.


Thanks, and my Regards,

Wayne.

[Juan P. Pertierra]

So how come cameras have different controls and features if they use the same IC.

Because the one or two(at the most) programmable IC's are embedded in several layers of NOT programmable logic. Even if we suppose for a second that reprogramming an FPGA already on a printed circuit board doesn't take any modification to the board itself, that will not do you any good if there are four pre-set hardware IC layers between your FPGA code and the firewire port. Get it now?

And this applies for microcontrollers/microprocessors/ASICS, etc as well.

Now Jaun, are you saying that the firewire chip, is not a chip that can be setup to accept different format of data, like the firewire standard provides for?

In the printed, and soldered circuit board in which it is in, in these cameras, yes that's correct. There's nothing 'software' or 'slight' modification that you can do to reconfigure the firewire interface.

In fact, the firewire link layer is hardwired to the MPEG encoder IC which is also not-programmable hardware.

You have to build a new circuit, period.

And if you can point to one camera that doesn't work in this manner, i'd like to hear it, make sure you have actually seen the schematics. Hand waving arguments about how you think cameras work are fine, but what i'm talking here are facts that I can point out in schematics, not opinions.

Cheers,
Juan

[Wayne Morellini]

<<<-- Originally posted by Wayne Morellini : I am glad to have the chance to debate the two Jauns at the same time, but lets just leave it there and move on.-->>>

<<<-- Originally posted by Juan P. Pertierra : Because the one or two(at the most) programmable IC's are embedded in several layers of NOT programmable logic. Even if we suppose ... that will not do you any good if there are four pre-set hardware IC layers between your FPGA code and the firewire port.-->>>

Thanks, that's something.

Wayne.

[Ron Severdia]

I've been waiting months and months to see this project materialize and nothing yet. . .

Is this going to happen? Where's the website that was once "in progress"?

[Milosz Krzyzaniak]

What is the level of progress of this project? I've not been here for a lot of months and aren't able to read all the posts - could someone summarize?

[Flax Johnson]

<<<-- Originally posted by Ron Severdia : I've been waiting months and months to see this project materialize and nothing yet. . .

Is this going to happen? Where's the website that was once "in progress"? -->>>

http://www.reel-stream.com/

[Wayne Morellini]

<<<-- Originally posted by Flax Johnson : <<<--
http://www.reel-stream.com/ -->>>

Yes, it has already happened.

[Joel Corkin]

Juan,

Do you have an estimate of when you might start to sell this product? I'm interested in knowing a timeframe if that is at all possible to give.