Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

No more FF35 Scarlet?

They think it will take too much business from Epic. I guess two years is how long it takes a company to go from "the underdog that fights the big boys" into an "only bottom line" company. I don't have a problem with it, that's why people start businesses... Now we will also see more proprietary stuff, more regulation on this forum, prices will be "updated" north... etc.

And the funny part is that the only reason this happened is because Red One owners were complaining that they were "left out" etc etc... there was a lot of crying around here last year.

Um, the Scarlet FF35 has always been scheduled further out than the EPIC. It will have the 3rd generation sensor, the MONSTRO.
 
They think it will take too much business from Epic. I guess two years is how long it takes a company to go from "the underdog that fights the big boys" into an "only bottom line" company. I don't have a problem with it, that's why people start businesses... Now we will also see more proprietary stuff, more regulation on this forum, prices will be "updated" north... etc.

And the funny part is that the only reason this happened is because Red One owners were complaining that they were "left out" etc etc... there was a lot of crying around here last year, so... kind of funny.

Remember, the Epic is targeted at customers who have thus far been used to spending $100+k on camera + accessories. To them, the $20k makes no difference - the magnitude of which is further reduced once you account for lenses, modules, etc. Especially, they aren't going to lose out on the speed flexibility and necessary features like HD-SDI. On the contrary, they might even be tempted to go ahead and get an extra B or C camera in the Scarlet S35.

What the Scarlet S35 does is target a new market who couldn't really afford an Epic. This is likely to be the "indie filmmakers".

Now I am sure there are people who would be happy with the Scarlet even though they could have afforded an Epic, but in general, I think Red have got it right.

Moving forward, I would definitely like to see a FF35 Red camera that is more affordable and compact than the Epic FF35. Also, I feel Red should do a 2/3" Monstro version (or maybe a larger sensor still) to refresh the 2/3" M-X Scarlet. Perhaps the Monstro compact line-up could be renamed, Crimson or something. Sensor technology is evolving exponentially, and I have no doubt the difference will be worthy of a new brand name.
 
Anyone seen any recent info on FF35 models of any kind? They are not being discussed at all for the present as that is the next gen Monstro sensor line. The focus is all on getting the M-X based cameras to market as quickly as possible this year and supporting them.
Red has been pretty aggressive in its pricing so far considering there is nothing remotely close to the performance and features they are offering at anywhere near their prices. They do ultimately have to cover their development, production, and support costs if they are to stay in business, just as any company does.
The market for Scarlets will be much bigger than the one for Epic for the same reason that a lot more people buy Sony EX-1 EX-3's than by FW900's and 800's. The people that buy Epics will be in the same market as the higher end Sony, Arri, Panasonic products.
 
7d does not sample 14 bit for video and does not sample the full sensor for video. 90% of the sensor data is thrown out before A/D and encoding. The maximum full sensor raw frame rate for the camera is 8fps for up to 120 frames. Nowhere near the frame or continuous data rates necessary for motion.
Duh,
maybe it's my bad English, but you still don't get it.:banghead:

IF 7D CAN PIPE 2Gb/s FROM SENSOR THROUGH ADC IN "STILL" MODE, THERE'S NO REASON WHY IT COULDN'T DO THAT IN "MOVIE" MODE.

And if you've checked the dpreview link I gave to you or asked anybody who owns 7d and a fast cf card, you'd know by now, that it can write 8fps jpg's to card until it's full.
 
Duh,
maybe it's my bad English, but you still don't get it.:banghead:

IF 7D CAN PIPE 2Gb/s FROM SENSOR THROUGH ADC IN "STILL" MODE, THERE'S NO REASON WHY IT COULDN'T DO THAT IN "MOVIE" MODE.

And if you've checked the dpreview link I gave to you or asked anybody who owns 7d and a fast cf card, you'd know by now, that it can write 8fps jpg's to card until it's full.

But 8fps is the fastest it can deliver sustained throughput of those JPEGs, and JPEGs are not full res. Your idea that there is no difference between delivering a raw image in a single shot and sustaining that much data at 24fps is flawed. There absolutely is a difference.

If Canon were to present us a video mode of 8fps JPEG 2000 it would go over like a fart in church. Then they would be stoned as heretics.
 
Last edited:
Line skipping

Line skipping

Duh,
maybe it's my bad English, but you still don't get it.:banghead:

IF 7D CAN PIPE 2Gb/s FROM SENSOR THROUGH ADC IN "STILL" MODE, THERE'S NO REASON WHY IT COULDN'T DO THAT IN "MOVIE" MODE.

And if you've checked the dpreview link I gave to you or asked anybody who owns 7d and a fast cf card, you'd know by now, that it can write 8fps jpg's to card until it's full.

I guess if you are willing to shoot your movie at 8 fps, you are good to go! :thumbsup:

But, in movie mode, these DSLRs are not downsampling all the lines on the sensor, they are simply skipping more lines than they are recording in order to get down to the 2 megapixels or so that thay are actually recording as 30 fps video. As David said, the vast majority of the information that is available to the sensor to record is thrown out before it even gets gathered by the sensor. Think of of as turning off roughly two out of every three rows of bulbs on a marquee and that's what you are left with to read. And then, that information gets further compressed to h.264. It is actually amazing what pretty pictures can be derived from this system but it does have its downside.
 
Duh,
maybe it's my bad English, but you still don't get it.:banghead:

IF 7D CAN PIPE 2Gb/s FROM SENSOR THROUGH ADC IN "STILL" MODE, THERE'S NO REASON WHY IT COULDN'T DO THAT IN "MOVIE" MODE.

And if you've checked the dpreview link I gave to you or asked anybody who owns 7d and a fast cf card, you'd know by now, that it can write 8fps jpg's to card until it's full.

2Gb/s off the 18Mp sensor gets you 8 frames a second, and the camera can only sustain that for 120 frame burst, not continous. Nowhere near full motion. At 30fps 1080p the max continuous data rate they can do now is 330Mb/sec
 
2Gb/s off the 18Mp sensor gets you 8 frames a second, and the camera can only sustain that for 120 frame burst, not continous. Nowhere near full motion. At 30fps 1080p the max continuous data rate they can do now is 330Mb/sec

There you go! You can make a very high resolution 8 fps, 15 second movie!! :laugh:
 
But 8fps is the fastest it can deliver sustained throughput of those JPEGs, and JPEGs are not full res. Your idea that there is no difference between delivering a raw image in a single shot and sustaining that much data at 24fps is flawed. There absolutely is a difference.
Wrong, those jpg's are full res. 18 Mpx each.

I'm not trying to say that it could transfer and compress 18 Mpx at 24 fps.
I'm trying to say that when it can transfer and compress 18 Mpx at 8 fps, which makes 144 Mpx/s, it could also transfer and compress 6 Mpx at 24 fps, which makes the same 144 Mpx/s.
This is of course not implement in current models, but can easily be in the next ones.

If this is really hard to understand, let me give another possible future feature for Canon's next DSMC: instead of skipping lines & pixel, just use the center of sensor. This is called windowing. Same amount of pixels to ADC, same power to compression, nothings changed, other than all the artifacts from line skipping are gone.
 
Wrong, those jpg's are full res. 18 Mpx each.

I mis-spoke. However my point is that you can not extrapolate this data the way you are. The fact that the camera can handle XMb/sec does not mean that you can divide the data rate by half and then maintain that rate for twice as long. It just isn't that simple.

Are the engineers at Canon able to give you what you are asking for in some future model? Maybe, but I see no point in trying to debate about what they can or cannot do, what they may or may not do. Like you I'll be wkaiting to see their next move with interest.
 
Wrong, those jpg's are full res. 18 Mpx each.

I'm not trying to say that it could transfer and compress 18 Mpx at 24 fps.
I'm trying to say that when it can transfer and compress 18 Mpx at 8 fps, which makes 144 Mpx/s, it could also transfer and compress 6 Mpx at 24 fps, which makes the same 144 Mpx/s.
This is of course not implement in current models, but can easily be in the next ones.

If this is really hard to understand, let me give another possible future feature for Canon's next DSMC: instead of skipping lines & pixel, just use the center of sensor. This is called windowing. Same amount of pixels to ADC, same power to compression, nothings changed, other than all the artifacts from line skipping are gone.

But windowing completely takes away one of the major advantages of this format; namely sensor size, which which is what allows for the ability to do shallow DOF. The windowed sensor is back down to the 1/3 - 1/2" size (I haven't done the math). And it messes with the crop factor on the lens as well.

And, 18 mpx JPEGs are not equal in quality to 18 mpx RAW files, even in a DSLR.
 
However my point is that you can not extrapolate this data the way you are. The fact that the camera can handle XMb/sec does not mean that you can divide the data rate by half and then maintain that rate for twice as long. It just isn't that simple.
Actually it is that simple. There's nothing ipad-GUI-magic here. Just pushing samples through ADC and tranfsering bits from there on.
Maybe, but I see no point in trying to debate about what they can or cannot do, what they may or may not do.
If you look at the thread here, there was this debate, why Canon uses line skipping. I was only trying to make a point, that reason is not just Canon's sensor can't transfer enough samples to ADC or ADC can't convert enough samples to bits or processor can't handle the bitrate or write fast enough to cf.
It was even said that _all_ of these are the reason. But when 7d can do all that in jpg-still-mode, it can't be so.
Only possible explanation that Jannard said, that the problem is sensor's readout time, puzzles me. If it can do the readout fast enough with line skipping, I can't come up any reason why it couldn't do the same thing with windowing.
Of course all is just speculation, like everything about Scarlet, before it's really shipping...
 
But windowing completely takes away one of the major advantages of this format; namely sensor size, which which is what allows for the ability to do shallow DOF. The windowed sensor is back down to the 1/3 - 1/2" size (I haven't done the math). And it messes with the crop factor on the lens as well.
Again, if Canon would introduce a real DSMC, which would have FF, but only about 10 Mpx, the 6 Mpx window would be a lot bigger than 2/3".
I did the math already few posts before.
And, 18 mpx JPEGs are not equal in quality to 18 mpx RAW files, even in a DSLR.
The pipeline in camera is:
1. sensor -> 2. ADC -> 3. processor -> 4. memory card.
Steps from 1 to 3 are identical in both jpg and raw mode.
 
Again, if Canon would introduce a real DSMC, which would have FF, but only about 10 Mpx, the 6 Mpx window would be a lot bigger than 2/3".
I did the math already few posts before.

The pipeline in camera is:
1. sensor -> 2. ADC -> 3. processor -> 4. memory card.
Steps from 1 to 3 are identical in both jpg and raw mode.

I have to admit, your explanations and arguments completely befuddle me. You make claims about what the sensors can do (but which they aren't currently doing). Then, when someone points out to you why the sensors can't do that, you make the claim that they COULD do what you suggest, if only Canon would change them to do it. SO, I don't really understand what claim you are making. And, I'll admit, I long ago forgot what your point is.:banghead:

Could Canon make a chip like the MX... I imagine they could, with time. Have they... no, it wouldn't seem so. Might a 7D do everything you claim if only Canon would wake up and make it so? Well, maybe it could - but there is just no real evidence that says so.

And as far as the steps in your pipeline - how is it you make the claim that they are the same? The "processing" part is where they are different - JPEGs get "processed" in camera.
 
Last edited:
Seems to me that the line skipping is used to reduce the Jello effect while increasing iso at the same time. Half the lines means twice the scanning time per frame resulting more exposure time per frame. Yes?
 
Actually it is that simple. There's nothing ipad-GUI-magic here.

??? OK, that makes sense... You are clearly a very clever engineer who understands this stuff completely.

One of us is saying that the engineering of a digital cinema camera is complex, and the other is saying that it is simple... Oh yes and that the magical iPad GUI has something to do with it somehow... I get it.
 
So based on the performance that Canon 7D has now, 8fps, it will require a sensor that is 3 times faster in it read/write times, with 3 times the data flow that it now can do and 3 times the processing power to move it encode it and store it in some useful format. That is to make it to only 24fps.
Then to make it a practical motion camera they will need to dump the mirror reflex and replace it with an EVF, add pro level audio and video I/O and production features and build it into a larger and different form factor. So now its 3 times the price or more, as has been done by Panasonic with their just announced camcorder version of the GH1.
But Canon, being the conservative mass market company they are will not do this until it has been done by everyone else and the market is well established.
Nobody can do this at the current state of technology at Canon DSLR prices.
 
So based on the performance that Canon 7D has now, 8fps, it will require a sensor that is 3 times faster in it read/write times, with 3 times the data flow that it now can do and 3 times the processing power to move it encode it and store it in some useful format. That is to make it to only 24fps.
Then to make it a practical motion camera they will need to dump the mirror reflex and replace it with an EVF, add pro level audio and video I/O and production features and build it into a larger and different form factor. So now its 3 times the price or more, as has been done by Panasonic with their just announced camcorder version of the GH1.
But Canon, being the conservative mass market company they are will not do this until it has been done by everyone else and the market is well established.
Nobody can do this at the current state of technology at Canon DSLR prices.

But the new Panasonic doesn't do 4K RAW, so it isn't really 3 times faster or whatever the number is. Plus, as you aim for higher targets, there are more bottlenecks that creep in. Now, I don't know how exactly it does 1080p, but so far, all points at it line skipping. If it were to do real 4K RAW at $6k, that would be a "game-changer".

And like we have said several times, the DSLRs line-skip simply because they aren't powerful enough, neither were designed to, do full frame motion at 24 fps. Period. I don't understand why there's discussion about it. If they could do even 3K, the DSLR makers would obviously have implemented it.

It is possible they could dial up the resolution in future generations. I can see the 5D Mk. III doing maybe 2.5K or so. But it is still binning pixels.

And let's not get to the awfully slow read-reset...
 
Back
Top