vincent thomas
Well-known member
More footage and informations will be actually welcome, im keeping myself an eye on this camera even if Canon disapointed me lately
Canon 1DX is the "SAME" camera as the 1DC except FIRMWARE...
- Thread starter david_p
- Start date
electricjib
Active member
Epic and Scarlet are the same weight. I'm not sure who has done it, but apparently a few individuals have done research and found that the inside of an Epic and Scarlet are Identical and your paying for firmware. Someone I spoke with said they knew someone that had gotten two Scarlets to run like Epics through creating new software in separate parts of the camera. Which would mean the insides are identical.
I won't say It's true, but it wouldn't shock me....
I won't say It's true, but it wouldn't shock me....
John Andersson
Well-known member
If this camera delivers as much or greater real resolution as the C300 and gets 25p in 4K then I very much like it.
Samuel H
Major Contributor
The ASICs on the Scarlet are running at slower MHz. It could be because RED wants to skim different market tranches, by selling an artificially crippled product, but it could also be because the ASICs can't actually run any faster.
This processor binning is very common, every chip maker does it: all transistors are imperfect, the worse ones will work only at increased voltage and/or reduced MHz, and the whole processor has to work at the voltage and MHz that allow its slowest transistors to work fine.
If RED's ASIC is big and complex, which I guess it must be, it would not be uncommon to have a yield of 30% or less, that is, 70% or more of the chips the foundry makes are unable to run at the desired specs. By defining a lower spec, you get a second tranch of chips that you can sell: they would lead to lockups and data corruption if you tried to make them run at the higher spec, but they'll work perfectly at reduced MHz and/or increased voltage. These are chips you would otherwise have thrown to the dustbin, so they're basically free.
Often, all of this is even planned in the design of the processor: nvidia and ATI design their chips in a modular fashion, so if a big lithography error appears in one module (one that will make the transistor unable to work, at any speed, or that would only work at unacceptably slow MHz), instead of throwing away the chip, the module can be shut down permanently and this "salvage part" can be sold.
Right now, nvidia sells several graphics cards that use the same GK104 processor. The top of the line is the GTX 680 (8 working modules, running at 1006 MHz), the second one is the GTX 670 (7 working modules, running at 915 MHz).
In the old days, before this modular design, the difference in MHz was bigger; it was not uncommon to have, say, a top part that ran at 450MHz, and a salvage part that ran at 325MHz. I guess something like this is happening on the Scarlet.
This processor binning is very common, every chip maker does it: all transistors are imperfect, the worse ones will work only at increased voltage and/or reduced MHz, and the whole processor has to work at the voltage and MHz that allow its slowest transistors to work fine.
If RED's ASIC is big and complex, which I guess it must be, it would not be uncommon to have a yield of 30% or less, that is, 70% or more of the chips the foundry makes are unable to run at the desired specs. By defining a lower spec, you get a second tranch of chips that you can sell: they would lead to lockups and data corruption if you tried to make them run at the higher spec, but they'll work perfectly at reduced MHz and/or increased voltage. These are chips you would otherwise have thrown to the dustbin, so they're basically free.
Often, all of this is even planned in the design of the processor: nvidia and ATI design their chips in a modular fashion, so if a big lithography error appears in one module (one that will make the transistor unable to work, at any speed, or that would only work at unacceptably slow MHz), instead of throwing away the chip, the module can be shut down permanently and this "salvage part" can be sold.
Right now, nvidia sells several graphics cards that use the same GK104 processor. The top of the line is the GTX 680 (8 working modules, running at 1006 MHz), the second one is the GTX 670 (7 working modules, running at 915 MHz).
In the old days, before this modular design, the difference in MHz was bigger; it was not uncommon to have, say, a top part that ran at 450MHz, and a salvage part that ran at 325MHz. I guess something like this is happening on the Scarlet.
Last edited:
regnaleb79
Member
Samuel, you know too much, you're scaring me.
vincent thomas
Well-known member
I will like so much to have better highlight roll off...
__________________________________________________________
Vincent Thomas
Senior Lighter & MattePainter / Cinematographer
http://fr.linkedin.com/in/vincentthomas
(Available Now)
__________________________________________________________
Vincent Thomas
Senior Lighter & MattePainter / Cinematographer
http://fr.linkedin.com/in/vincentthomas
(Available Now)
Most reasonable explanation is that the orders for Epics were starting to drop off. How many Epic cameras can they sell worldwide until that pricepoint is saturated? So they started selling the same parts at a lower price point. Just lock it down in firmware.
"Artificially crippled" is pretty much moot anyway. Every single company does that. Anyone noticed how the Sennheiser hd558 is the same as the higher price hd598? They just add a bit of plastic to reduce the sound quality and then sell at a cheaper price.
Boohoo.
William_Robinette
Shallow Member
Do you really, actually believe this? Wow...
G
Guest
Guest
Perhaps if Canon had one or two different connections on the side of the Camera for the 1D-C than it would have been something more understandable, but just a firmware change in resolution recording, and a badge, hmmm, how many thousands is that worth? Of course the value of large sensor 4k is subjective in this market at this time, and while it may be feasible in a much cheaper model, there are still many early adopters that are willing to pay premiums. And yes, the coders and technicians need to eat too.
http://www.dpreview.com/products/compare/side-by-side?products=canon_eos1dc&products=canon_eos1dx
http://www.dpreview.com/products/compare/side-by-side?products=canon_eos1dc&products=canon_eos1dx
squig
Veteran
Your forgetting the $7000 headphone minijack :thumbsup:
Well, if Canon were to drop the 1D-C and just sell the 1DX as a 4K camera at $6,800, they would surely kick butt again for a while! 
Too bad they are so corporate. All you need is a bit of ego and have the CEO/engineers get sick of all the BMCC talk. I'll show those s**b*'s what a camera is ...
We can hope, huh? Of course I have no real need for 4K right now, but ...
My 1DX shows up tomorrow. Guess it is time to start hacking it.
Nest,
Michael
Too bad they are so corporate. All you need is a bit of ego and have the CEO/engineers get sick of all the BMCC talk. I'll show those s**b*'s what a camera is ...
We can hope, huh? Of course I have no real need for 4K right now, but ...
My 1DX shows up tomorrow. Guess it is time to start hacking it.
Nest,
Michael
That you can hack/overclock Scarlet is not all that shocking, but it doesn't mean Epic/Scarlet differ only in firmware. I guarantee those hacked Scarlets have heat and/or stability problems.
Maybe. Or maybe it's like the CPU-world where overclocking rarily gives any heat/stability problems.
legrevedotcom
Veteran
Why? If it's identical hardware, then there shouldn't be any issues once the firmware has been updated.
In the case of the Canon article they actually go as far as stating that a modded 1D X would have no heat or stability issues with a firmware hack.
I say go go and make it happen. I wonder also now wether this is also the case in the matter of FS100 vs FS700....
Greyfus Wolf
Active member
Hopefully high end cameras like the RED can be made smaller and to consume less power in the next few years
With stuff like Globalfoundries recently announcing the 14nm process to launch in 2014, it shouldn't take too long I hope.
Samuel H
Major Contributor
As I explained above, "identical hardware" doesn't mean it's able to run at the same speed. You're overclocking the chip; if it was clocked down because RED wanted to skim the market, you'll be fine; if it was clocked down because the lithography was not good enough to run at the higher spec, it will get hot, and you'll have data corruption and lockups. BTW, if you think lockups are the nastier sibling, you're wrong: data corruption without lockups is an awful thing to have, you won't notice something went wrong until you're at the PC, finding corrupted frames here and there, and maybe also corrupted files that can't be recovered. I've been an overclocker since I found out my Pentium 120MHz could run at 133MHz, but I would advise against doing this in a computer where data corruption can cost lots of money.
I'd contribute...
It's a different sensor and thus a different way of aquiring an image, so that's never going to be possible.
The C300 acquires the image differently but it’s still just a good old fashioned Bayer filter array. It’s a 4k (quad HD) sensor that just does a 2*2 pixel bin to give 1080 instead of debayering to 4k. So there is nothing about the 4k aps-h crop on the 1dx/1dc that means it couldn’t theoretically do the exact same kind of binning.
So debayered 4k may have very similar actual resolution to 2*2 binned 1080.
Of course to have 2*2 binned 4k you would need an 8k sensor.