So, why is there such a difference in raw sizes between the 10MP 400D RAW image, and the10MP 1DM3 RAW image? Canon lists 9.8MB for the former and 13MB for the later. If the number of bits is 12 vs 14 (with no extra bits) why is the 1DM3 CR2 file 33% larger then the 400D CR2 file?
I just can't figure out any math to explain that difference, unless there are more bits and no compression of any sort. Either that, or there is some significant amount of dark matter somewhere in the file of the 1DM3.
Mark Shaxted wrote:
Oh yeah - raw files actually contain raw data, two jpegs, and a wacky bitmap!
David Kilpatrick writes on DPR's Sony SLR Talk forum:
"Late August is now pretty busy - Pentax European press conference, 23rd; Canon Autumn Collection launch London, 21st; Nikon London launch, 24th. All announced, press contacted, RSVPs done etc.
Nothing so far from Sony and I have emailed to ask if I need to be careful with a rapidly filling diary - no response yet.
If they have a DEALER conference on Sept 6/7, this usually happens two to three days after the press launch, so Sept 4th seems a likely possible press launch date. The normal pattern is: Monday - Japanese arrive, cameras arrive, rehearsal for press day, briefings; Tuesday - press day; Wednesday - postmortem on press day, briefing for dealer launch, cameras distributed; Thursday and Friday - first dealer meetings, cameras in hand, by staff who have already attended the briefings and the press launch. That's a typical week schedule not just for Sony but many companies. They time it so that a Japanese delegation can come over for a week or maybe two (service staff will get separate training or briefing)."
DynoMoHum wrote:
So, why is there such a difference in raw sizes between the 10MP 400D RAW image, and the10MP 1DM3 RAW image? Canon lists 9.8MB for the former and 13MB for the later. If the number of bits is 12 vs 14 (with no extra bits) why is the 1DM3 CR2 file 33% larger then the 400D CR2 file?
I just can't figure out any math to explain that difference, unless there are more bits and no compression of any sort. Either that, or there is some significant amount of dark matter somewhere in the file of the 1DM3.
I don't know. I'm guessing now, so I might be way off...
But their are some noice reduction and anti aliasing going on before the RAW file are created right? (I'm not sure, - but I think this is true).
If so, - I imagine that these algorithms are simply different. They can be for a lot of reasons:
1) 1 D have more computing power (I don't think so, - they'll both use the Digic III)
2) On purpose they make the files a bit worse on the 40D (I don't think so, the competition to other companies are to high)
3) Different cameraes for different customers. For 40D users more noice reduction would be appreciated, where for 1D less would. That could explain a difference
4) The sensor are different. If the sensor in the 1D gives less noice (random data) it will take less space. So that's not likely to be the case (it would give the opposite result). [Actually I think that is the case, - at same resolution each photo-collector is 69% larger, - that I think should lead to less noice. Not much of a difference, - but still.]
There are probably more possibilites and mine might be flawed. I admit that I am very much guessing now!
The 33% difference doesn't really mean too much. I just checked my own 30D cr2s, and they vary between 6.6Mb & 14.2Mb. That's a heck of a dfference. The average is 7.9Mb.
In truth, I've yet to look at a MkIII cr2, so I don't know for sure what else is stored there, nor do I know how large the embedded jpegs are.
The 33% difference doesn't really mean too much. I just checked my own 30D cr2s, and they vary between 6.6Mb & 14.2Mb. That's a heck of a dfference. The average is 7.9Mb.
In truth, I've yet to look at a MkIII cr2, so I don't know for sure what else is stored there, nor do I know how large the embedded jpegs are.
Exactly; ANY compresion algorithm will yield varying degrees of compression depending upon the data being compressed.
Even if it was consistent, we only have a RUMOR of how large a typical raw file from the 40D would be. It's an exercise in mental masturbation, that won't yield anything meaningful. We'll know when the camera is announced.
Oh, and don't get your hopes up on weathersealing either. It's really hard to tell anything from the quasi-legitimate (OK. probably legitimate) shots.
DaDane wrote:
I don't know. I'm guessing now, so I might be way off...
But their are some noice reduction and anti aliasing going on before the RAW file are created right? (I'm not sure, - but I think this is true).
If so, - I imagine that these algorithms are simply different. They can be for a lot of reasons:
1) 1 D have more computing power (I don't think so, - they'll both use the Digic III)
2) On purpose they make the files a bit worse on the 40D (I don't think so, the competition to other companies are to high)
3) Different cameraes for different customers. For 40D users more noice reduction would be appreciated, where for 1D less would. That could explain a difference
4) The sensor are different. If the sensor in the 1D gives less noice (random data) it will take less space. So that's not likely to be the case (it would give the opposite result). [Actually I think that is the case, - at same resolution each photo-collector is 69% larger, - that I think should lead to less noice. Not much of a difference, - but still.]
There are probably more possibilites and mine might be flawed. I admit that I am very much guessing now!...Show more →
Picking up on the above, there are a number of actual and possible contributors for the difference in specified average file sizes between the 400D and 1D Mk III RAW files -
1) 400D uses DiGiC 2, not DiGiC 3
2) 1D Mk III uses 14-bit colour depth, whilst 400D uses 12-bit
3) The embedded jpeg in the 1D Mk II RAW files may be larger than that embedded in 400D files.
The 'leaked' file size for the '40D' RAW files is nearer to that of the 1D Mk III than the 400D, possibly indicating 14-bit but with a smaller embedded jpeg or slightly tweaked encoding algorithm.
I'm guessing the 40D will actually ship with 13 bit colour as it's a consumer model. Can't step on the 1D III's toes. Also it'll only have 1 Digic III processor and the compression won't be as good, so that's why the file size will be the same as the 1D III on average.
IFeito wrote:
I know one Canon guy in Mexico, who actually works in technical services, who says that he gets all his "insider information" from forums such as this one.
imeod wrote:
Well then he is pretty stupid.
Seriously, how many people in Canon do you think know anything a second before we do?
I don't say that there is none. Obviously not. But my guess is that the majority don't know before the press release is out of the door.
DaDane wrote:
Seriously, how many people in Canon do you think know anything a second before we do?
I don't say that there is none. Obviously not. But my guess is that the majority don't know before the press release is out of the door.
I think you have a good point. I worked at Apple Computers, and we did not know anything about any new products coming out that the public didnt know, and that includes all the "geniuses", managers and regional directors.
There must be quite a few people in factories somewhere that know something... There also must be people who work for printing companies or in the printing department at Canon who have seen manuals, etc... Surely if this new camera is going to be ready to ship in September, there must be hundreds or maybe even thousands who've seen or know about something...
I can believe that many people who work for Canon do not know... but there have to be a sizable amount of people somewhere in the world that do know more then a little about the new camera.
