[jptreen]

Hard drives will have to improve a lot before being able to handle such a bitrate.

I think they will, but I agree that it's way out there at the moment. Love the IMAX article, by the way. Love IMAX too...

...incidentally, i hear they're planning on bringing out a Digital IMAX camera in the next couple of years...imagine that workflow...

[jmbradd]

To the point about hard drives, with files getting this large photographers better start moving away from platter drives. More sectors written means higher chance of unrecoverable error!

Crazy to think that when I graduated high school my 160gb IDE hard drive was a steal at 80 bucks. I can only imagine my daughter coming home with some "goofy old camera" she found at goodwill called a "645d"


Question, has anyone seen digital medium format used for aerial photography? Seems like it would be better than having a 75 foot roll of film!

[ghelary]

I think they will, but I agree that it's way out there at the moment. Love the IMAX article, by the way. Love IMAX too...

...incidentally, i hear they're planning on bringing out a Digital IMAX camera in the next couple of years...imagine that workflow...

Well, on contrary to pupular wisdom, the writing speed on hard drive have not improve that much in the last year and particulary not on par with hard drive sizes. The current SATA design is capable of handling 3Gps data transfer on theory but commercial hardrives can handle more commonly around 0.5Gbps.

As you can see there is a lot of time to wait before such devices are able to handle the 14GPs described above. And maybe at the end, we would have to use solid state drives, we leads to wonder when we would be able to handle the data amount.

Digital IMAX camera will be more around 2K and 4K than the 18K that I took as an example. Still an insane amount of data, but more likely to be handle in the short term.

[melander]

Well, on contrary to pupular wisdom, the writing speed on hard drive have not improve that much in the last year and particulary not on par with hard drive sizes. The current SATA design is capable of handling 3Gps data transfer on theory but commercial hardrives can handle more commonly around 0.5Gbps.

As you can see there is a lot of time to wait before such devices are able to handle the 14GPs described above. And maybe at the end, we would have to use solid state drives, we leads to wonder when we would be able to handle the data amount.

Digital IMAX camera will be more around 2K and 4K than the 18K that I took as an example. Still an insane amount of data, but more likely to be handle in the short term.

My 80 GB Intel SSD manages to write and read around 1.6 Gbit, (210 MB/s) and at the end of this year comapnies are saying that they will manage 600 GB solid state drives in a consumer market. So there is no lack in writing / reading spead in todays technology.

[falconeye]

Sure, the IMAX boys would know how to use it, but please take in account that 60Mpx @ 8bits (bayer) @ 30fps would make around 14.4 Gbps of data flow without counting any type of header.

My 80 GB Intel SSD manages to write and read around 1.6 Gbit, (210 MB/s) and at the end of this year comapnies are saying that they will manage 600 GB solid state drives in a consumer market.

What are all of you smoking?

The 18K claim in the source given is BS. It was derived from saying IMAX is 9 times the surface of Academy 35mm (which is 1/2 of still photo 35mm). The truth is IMAX film frames are 71mm x 52mm. The source continues in saying "After 1080P comes 2K" ... well 2K is 2048x1080 or 7% more pixels than 1080p. The author is obviously following a hidden agenda...

18K for 71mm is like 9K for 36mm or 56 Mpixel for 36x24mm film. Or a resolution of ~130 lp/mm while we all know that color film never actually resolved better than ~40 lp/mm. Just forget about this source.

Maybe the best statement about IMAX resolution is this: Last Update Was Sunday – IMAX Resolution - TristanPipo.com.
It states a resolution of 6100x4500 or 27MP.

However, the sad truth is that IMAX is now digitally projected using the 2K standard (2048x1080) which, in IMAX 1.36:1 means 1469x1080 or 1.6 MPixels. http://phx.corporate-ir.net/phoenix.zhtml?c=118725&p=irol-newsArticle&ID=1117681&highlight=

And nobody talks about 60p. The best I heard of (James Cameron) is 48p ...


BTW, nobody would use SSD to store a high bitrate movie ... RAID arrays are faster, cheaper and more reliable.

[ghelary]

What are all of you smoking?

The 18K claim in the source given is BS. It was derived from saying IMAX is 9 times the surface of Academy 35mm (which is 1/2 of still photo 35mm). The truth is IMAX film frames are 71mm x 52mm. The source continues in saying "After 1080P comes 2K" ... well 2K is 2048x1080 or 7% more pixels than 1080p. The author is obviously following a hidden agenda...

18K for 71mm is like 9K for 36mm or 56 Mpixel for 36x24mm film. Or a resolution of ~130 lp/mm while we all know that color film never actually resolved better than ~40 lp/mm. Just forget about this source.

Maybe the best statement about IMAX resolution is this: Last Update Was Sunday – IMAX Resolution - TristanPipo.com.
It states a resolution of 6100x4500 or 27MP.

However, the sad truth is that IMAX is now digitally projected using the 2K standard (2048x1080) which, in IMAX 1.36:1 means 1469x1080 or 1.6 MPixels. Press Release

And nobody talks about 60p. The best I heard of (James Cameron) is 48p ...


BTW, nobody would use SSD to store a high bitrate movie ...

I never believed in that 18K thing, I just said that, using this information we're not close to 60Mpx.

Anyway, Digital is just getting getting on par with 35mm film for movies, so the speculations that we are making about 60Mpx movie cameras is just about science fiction.

On a side note, I wouldn't be surprised to see SSD used as a standard format for movie cameras in the near future (like SD or compact flash are for still cameras)

Anyway, the Pentax 645D will certainly not be a 60Mpx camera, and will not shoot at 30fps, or even 10fps. 1fps is more than enough for most usages, MF is not adapted to Sport photography for many other reasons so there's no reason to put a particular effort on this point.

[falconeye]

I never believed in that 18K thing, I just said that, using this information we're not close to 60Mpx.

That's a strange way of saying that. 18K for 1.36:1 translates to 240Mpx ...

SSD and SD are both flash. SSD just uses more chips in parallel and another connector (SATA). Future flash memory for movie cameras will be something in between. One 1.8" SSD and four CF cards have about the same capacity (256 GB) but the SSD drive has about 1.5 times the volume. Capacity density for SD would be 50% better. SSD is at least as large as 6 CF cards or 18 SD cards.

On the other hand, 1.8" hard disk camcorders do exist already and I agree, this category will upgrade to exchangeable SSD disks instead.

---

Last edited by falconeye; 03-03-2010 at 05:12 AM.

[pingflood]

while we all know that color film never actually resolved better than ~40 lp/mm

I know you usually can back your statements up, so can you find a reference for this? To the best of my knowledge good color film resolves a good bit more than that (55-60 lp/mm).

(edit: unless you are talking about system resolution in which case 40-something would be on target.)

[falconeye]

I know you usually can back your statements up, so can you find a reference for this? To the best of my knowledge good color film resolves a good bit more than that (55-60 lp/mm).

(edit: unless you are talking about system resolution in which case 40-something would be on target.)

Well, that was from my memory.

Sources (Leica LFI 3/2001) for best Black&White talk about (Kodak Tech.Pan) 320 lp/mm max. and 100 lp/mm at 50% MTF. Gigabit film and Kodak Tmax 100 are the same for 50% MTF but Gigabit film can be 600 lp/mm with 5% contrast.

As for color, sources are less reliable. Agfa specified all MTF curves and published (Agfa Technisches Datenblatt F-PF-D4 07/2003) figures (for 1.6:1 MTF) of:
50 lp/mm: Agfa Scala 200x, Agfachrome RSX II 100, Agfachrome RSX II 200, Agfacolor Optima 100, Agfacolor Optima 200, Agfacolor Optima 400
55 lp/mm: Agfachrome RSX II 50
60 lp/mm: Agfacolor Portrait 160
The above document publishes complete MTF curves and can be expected to be accurate information.

This online document (http://cool.conservation-us.org/coolaic/sg/emg/library/pdf/vitale/2007-04-vi...resolution.pdf) quotes the following for 30% MTF:

Kodachrome 25/64: 50 lp/mm
Ektachrome 50: 40 lp/mm
Ektachrome 160: 35 lp/mm
Fujichrome EI 100: 45 lp/mm
Fuji Velvia 50 RVP: 80 lp/mm

The same source states that the effective resolution including lens and film transport is lowered by 30-80%, in particular turning 50 lp/mm film resolution into 33 lp/mm when scanned.

I think 40 lp/mm is a fair summary of achieved resolutions during the film era, knowing that the combined use of exceptional film (Velvia, Tmax 100), outstanding lenses (Leica, Zeiss, some Pentax) and extreme care in the setup (camera calibration etc.) could yield 2x better results. At the same time I know that most people, when scanning their film archive, achieve much worse results ...

And I have no reason to believe that the IMAX film capturing workflow was on the exceptional side of things. Which is why they used such a huge frame size to start with

[kenyee]

Any hints on whether the 645D will have an ISO3200 mode like the phaseone?
Medium format joins the high-ISO wars? (Phase One P40+ 40mp digital back review) - 1001 Noisy Cameras

[ghelary]

Any hints on whether the 645D will have an ISO3200 mode like the phaseone?
Medium format joins the high-ISO wars? (Phase One P40+ 40mp digital back review) - 1001 Noisy Cameras

I would doubt it as it is a Phase One patent.

[falconeye]

If the sensor doesn't have a variable analog gain then there is no advantage in high ISO settings. Just underexpose and push process.

If the sensor does have a variable analog gain, then all vendors using the sensor will offer the corresponding ISO stepping.


As for "This is because of the new Sensor Plus technology that turns four pixels in one-giant pixel": This is a trivial patent which only the US could grant. But never mind, it is useless anyway. Just underexpose by 1 stop, pushprocess 1 stop, downsample to 50% size and you obtain the exact same result from a standard LR workflow.

[thibs]

If the sensor doesn't have a variable analog gain then there is no advantage in high ISO settings. Just underexpose and push process.

If the sensor does have a variable analog gain, then all vendors using the sensor will offer the corresponding ISO stepping.


As for "This is because of the new Sensor Plus technology that turns four pixels in one-giant pixel": This is a trivial patent which only the US could grant. But never mind, it is useless anyway. Just underexpose by 1 stop, pushprocess 1 stop, downsample to 50% size and you obtain the exact same result from a standard LR workflow.

Yep, I thought so too but I wouldn't dare to say so since I'm not confident enough.

Thank you Falk.

[cfraz]

This is a trivial patent which only the US could grant. But never mind, it is useless anyway..

This would be really funny if it didn't hurt so much. Unfortunately many companies use trivial patents to an effect which is directly opposed to the purpose of the patent system. Their abuse of the patent system stifles useful innovation and the advancement of science and technology for the public's good.

[falconeye]

Their abuse of the patent system

While I agree with your sentiments, the US patent system isn't abused.

The problem is that the officers in the US patent office don't do their job. They leave it to lawyers and expert panels to do so. This renders the entire system absurd as going to court was always an option... If you know that Einstein once worked as a patent officer you know what I mean

The time has arrived that a patent officer be sent to jail for not protecting the interest of the general public which is (or should be) partially paying him (tax payers). Interestingly, the EU patent office issues far less trivial patents.