1D X FPS limiting factor?

[PackLight]

Write time is the slower of camera write speed and card speed. I don't know how fast the 1D X write speed is to the card, but you can bet that the published specs were with the fastest card available. Nice and simple...... And wrong! After the image is processed it is stored in high speed memory, which is then dumped to the card. The bottleneck occurs when the read/process/dump to high speed memory occurs faster than you can write to the card.

A point might be, would any of the bodies write to the buffer at 12 fps? Write speed to the card would have nothing to do with the initial frame rate. Other than once you hit the buffer your fps falls to 2 or 3, there about.

 

[Loswr]

Yes sensor size is a factor in determining file size, take the 7D sensors pixel density and make the sensor FF size. See how that affects file size.

Clouds block the sun. Clouds produce rain. An umbrella blocks the sun, therefore, an umbrella produces rain? Correlation isn't causation. The 1DII and 20D were contemporary, same MP count. Same for the 1DIII and 40D. By your logic, where does the relatively tiny 41 MP sensor in the Nokia phone fit into the picture?

FWIW, here's what Canon says:

"With all these benefits, it’s only natural to wonder why all DSLR cameras aren’t full-frame. Ultimately, the issue is money. Research, development, manufacturing and distribution costs are all independent of camera size, so a smaller camera will not cost appreciably less than a larger one for any of these reasons. The end cost difference between small mirrors, mirror boxes, chassis and so forth, and larger ones is not that great. The difference is the sensor."

Not file size, money. Smaller sensors are cheaper.

 

[Loswr]

Speaking of APS-C sensors evolving, does the much newer T3/1100D APS-C sensor produce smaller files than the older FF 5DII?

 

[Don Haines]

Write time is the slower of camera write speed and card speed. I don't know how fast the 1D X write speed is to the card, but you can bet that the published specs were with the fastest card available. Nice and simple...... And wrong! After the image is processed it is stored in high speed memory, which is then dumped to the card. The bottleneck occurs when the read/process/dump to high speed memory occurs faster than you can write to the card.

A point might be, would any of the bodies write to the buffer at 12 fps? Write speed to the card would have nothing to do with the initial frame rate. Other than once you hit the buffer your fps falls to 2 or 3, there about.

Regardless of the card speed, the 1D X is supposed to be able to write 12 frames per second to the buffer.... If the card speed is not enough to keep up, it eventually fills the buffer and then the speed drops to what the card can support. That's why the 60D starts off at 5.3 but when the buffer fills drops down to less than 2. Same problem, just a lot more dramatic difference

 

[Don Haines]

Speaking of APS-C sensors evolving, does the much newer T3/1100D APS-C sensor produce smaller files than the older FF 5DII?

But to further complicate the issue..... Newer cameras MIGHT be running a better compression algorithm....so it is possible that a new 18M camera produces a smaller file than an old 18M camera.... And then there is 14 bit d/a compared to older 12 bit and even 8 bit........

 

[PackLight]

FWIW, here's what Canon says:

"With all these benefits, it’s only natural to wonder why all DSLR cameras aren’t full-frame. Ultimately, the issue is money. Research, development, manufacturing and distribution costs are all independent of camera size, so a smaller camera will not cost appreciably less than a larger one for any of these reasons. The end cost difference between small mirrors, mirror boxes, chassis and so forth, and larger ones is not that great. The difference is the sensor."

Not file size, money. Smaller sensors are cheaper.

Canon's white paper on FF sensors, August 1st, 2006

Has Canon ever told us one thing at one point then something different a few years later?

This was my original comment on the sensors:
"There was an old article I read a while back, put out by Canon explaining why we have crop sensors at all. From memory the whole reason was file size and processor performance at the time. Cheaper sensor cost and other things were not an issue initially."

I am not sure if it was one of the Canon books I have, or somewhere on their website. I would like to find it as a curiosity, it really doesn't matter.

 

[tpatana]

Yes sensor size is a factor in determining file size, take the 7D sensors pixel density and make the sensor FF size. See how that affects file size.

Sensor size does not determine file size. How many times we need to tell you this?

 

[tpatana]

This was my original comment on the sensors:
"There was an old article I read a while back, put out by Canon explaining why we have crop sensors at all. From memory the whole reason was file size and processor performance at the time. Cheaper sensor cost and other things were not an issue initially."

Yes, and we've replied you at least 10 times now that you are mistaken, the crop sensor [size] doesn't affect the file size. Either you just remember incorrect, or read incorrect, misunderstood, or the article you read was incorrect. But sensor size doesn't make the file size. It's the MP count (mostly, plus bit-depth and algorithm and stuff, but not sensor size)

 

[PackLight]

Speaking of APS-C sensors evolving, does the much newer T3/1100D APS-C sensor produce smaller files than the older FF 5DII?

What I think this is a real indication of, is how much effort Canon put in to the Crop bodies compared to the 1D X's smaller files. It's easy to use up file space, tougher to clean it up and make it better and smaller.

 

[Don Haines]

Yes sensor size is a factor in determining file size, take the 7D sensors pixel density and make the sensor FF size. See how that affects file size.

Sensor size does not determine file size. How many times we need to tell you this?

Are you measuring size in area, or size in number of pixels...... One of you is arguing with one definition, one with the other....

Given the same degree of compression and a/d bit depth, an APS C sensor and a FF sensor with the same number of megapixels will produce the same size file.

BUT....
If you are arguing based on area, then to scale up an 18M APS C to FF physical size, you could fit 46M pixels into the area of a FF sensor and of course the file would be bigger

 

[PackLight]

Yes sensor size is a factor in determining file size, take the 7D sensors pixel density and make the sensor FF size. See how that affects file size.

Sensor size does not determine file size. How many times we need to tell you this?

Are you measuring size in area, or size in number of pixels...... One of you is arguing with one definition, one with the other....

Given the same degree of compression and a/d bit depth, an APS C sensor and a FF sensor with the same number of megapixels will produce the same size file.

BUT....
If you are arguing based on pixel density, then to scale up an 18M APS C to FF physical size, you could fit 46M pixels into the area of a FF sensor and of course the file would be bigger

All Crop sensors have less MP than full frame. They do not like me saying file size has anything to do with the physical size of the sensor. Of course two sensors with equal pixel density, the larger produces a larger file. Were not supposed to relate these two in this discussion though. They want to be hard core sticklers and say file size is controlled only by the number of MP, which is somewhat correct except there are a few other factors involved.

 

[tpatana]

They want to be hard core sticklers and say file size is controlled only by the number of MP, which is somewhat correct except there are a few other factors involved.

Argh... it's pretty much correct.

Crop size 18MP and FF size 18MP give exactly same file sizes, assuming algorithm and such is same.

 

[lethalfalcon]

To answer the OP, I'm pretty sure the limit is in the mechanical shutter, however, like Mt. Spokane mentioned, it's likely that the capabilities of the processors and the buffer are matched to the limitations of the shutter mechanism. It makes no sense to over-engineer something if it can never be used. If they could get 14FPS with the shutter moving, they wouldn't have needed it locked, and it would have been better for them if they could have allowed focusing at 14FPS. Saying JPEG only is not dependent on whether the shutter moves, either. Also, it should have been (relatively) easy to slap more buffer RAM on the board if it were needed to handle more data.

Card speed is definitely a limiting factor, though, but more on the continuous shooting side. If you put 1GB of buffer memory into the camera, but it takes 8 seconds to clear it (and that's with Lexar 1000x cards), once you fill that buffer you're going to be waiting if you want to burst again. They might be able to fit 4GB on there, but then you need to wait half a minute to write the whole buffer out. I think the processors can handle a bit more than the 150MB/s of data the card can, especially if you're shooting clean images (ISO 100). The more noise and the more cleaning functions you turn on, the more you'll tax the logic. And they could always go for more chips if they can find the room. Think quad Digic 5+.

As to the later posts about file sizes, consider this: a 7D and a 1DX are really close to the same MP (same written image size of 5184x3456). However, a 1DX takes FAR cleaner images under almost all circumstances. Cleaner images = better compression. The idea is that a cleaner image is likely to have more pixels that have the same value, so the compression algorithm can more efficiently store that. If every pixel is a different color (because it actually is or because of noise changing the output just slightly), compression falls apart. I just took two pure white images (.6" ISO100 no lens, shot at a lit wall) from a 7D and a 1DX. Both show they're completely blown out in LR. Results: 14,498,781 bytes for the 1DX and 14,223,956 for the 7D. I'd guess that 200KB is mostly extra metadata.

 

[rs]

All Crop sensors have less MP than full frame. They do not like me saying file size has anything to do with the physical size of the sensor. Of course two sensors with equal pixel density, the larger produces a larger file. Were not supposed to relate these two in this discussion though. They want to be hard core sticklers and say file size is controlled only by the number of MP, which is somewhat correct except there are a few other factors involved.

Not all crop sensors have less MP than full frame. The original Canon 1Ds was full frame with 11MP, the original Canon 5D had 12MP, whereas the latest generation of APS-C Canon's have 18MP.

However, if what you're on about is scaling up the current generation of crop sensors to FF, along with the appropriate increase in MP, then you'll end up with a 46MP FF sensor (18MP x 1.6h x 1.6v). That would result in more data, but the extra data is directly proportional to the MP increase. For two sensors of equal MP, the sensor size has no effect on the amount of data (other than the larger sensor potentially producing lower noise levels, allowing even the RAW files to be compressed more, as stated by lethalfalcon)

 

[EchoLocation]

All Crop sensors have less MP than full frame. They do not like me saying file size has anything to do with the physical size of the sensor. Of course two sensors with equal pixel density, the larger produces a larger file. Were not supposed to relate these two in this discussion though. They want to be hard core sticklers and say file size is controlled only by the number of MP, which is somewhat correct except there are a few other factors involved.

As has been said above, All crop sensors do not have less MP than full frame sensors.
What do you want from this thread exactly?
There have been numerous discussions on numerous different aspects of files size in relation to sensor size, MP, compression algorithims, and whatnot...
It was pretty clear from the second or third post that no one knows exactly which aspect of the 1DX limits the FPS. There are many different areas of concern that are all near (80%ish) of their capabilities.
There are many different opinions on this but no hard facts, we've heard many opinions(many from good sources) so far, but obviously you haven't heard one that satisfies you.
What exactly are you looking to hear? that the file size and Digic V's are the limiting factor(to corroborate the story you remember)? i'm a little confused on what we are arguing about at this point besides your lack of satisfaction from the answers given.

 

[Don Haines]

All Crop sensors have less MP than full frame.

The T3i and the 1D-X have the same number of pixels...... That's right..... an obsolete rebel has the same number as the just released flagship.....care to restate that statement?

 

[Don Haines]

They do not like me saying file size has anything to do with the physical size of the sensor. Of course two sensors with equal pixel density, the larger produces a larger file. Were not supposed to relate these two in this discussion though. They want to be hard core sticklers and say file size is controlled only by the number of MP, which is somewhat correct except there are a few other factors involved.

The file size produced by the sensor is 5184 pixels X 3456 pixels X 14 bits X 3 colors, all divided by 8 bits per byte for one of Canon's current 18M sensors. That's a 94M file. That is the size of the file an 18M sensor produces... Period! And does not matter if it is in a iD-X or in a point and shoot.

The camera then runs lossless compression algorithms to create the data for the raw file, plus adds in some overhead, such as the Exif data. The size of the .raw file depends on the algorithm used and the nature of the image. Complex algorithms will compress more than simpler algorithms, but at the cost of needing a lot more computing power.... But at the level we are talking about with Dslrs, the difference is minimal.

It is the nature of the image that has the most effect on .raw file size..... Look at the images on your computer and good luck finding two of he same size.... Try taking two images, one of a properly exposed scene with lots of variation (no big areas of the same color) and the other picture with the lens cap on..... That's the effect of image scene and that's about 99.9 percent of what affects file size

 

[rs]

what is limiting the maximum FPS for the 1D X?

is it the bandwith/processing power of the DIGIC5+ CPU´s?

or is it the mechanical shutter construction?

could a dual digic 5+ powered camera deliver 10 FPS for a 21-24MP sensor?

To return to the original question, the 1D X as far as I know is beyond any SLR before in terms of frame rate for a moving mirror at 12 FPS, and also beyond any DSLR's data throughput at 216 MP/s (18MP 12 times each second, it's max in RAW mode).

While both of those are pushing the boundaries, as to which is forming the actual limit, only the Canon engineers will know. However, it's likely the mirror assembly is moving as fast as they could get working reliably at that price point, and the data throughput of the dual digic 5+ at 12 FPS would have been pushed too far with one of the many 20+ MP sensors they have put in every other full frame body released since the 1Ds mk III.

Therefore they're probably both right up at their respective limits, in balance with each other.

To answer your last question, based on the above assumption of an upper limit of 216 MP/s, at 10 FPS a dual digic 5+ camera could work with a sensor of up to 21.6 MP. Any more MP and the frame rate would have to drop further.

And the rumoured 46MP camera with the same processing pipeline would max out at 4.7 FPS - if it's 50+ MP, it'll be 4.3 FPS or less.