Pages: 1 2 3 [4]   Go Down

Author Topic: dynamic range  (Read 58412 times)

TechTalk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3612
dynamic range
« Reply #60 on: October 08, 2007, 05:42:25 am »

Quote
Since 40C is quote hot (except perhaps by the standards of Queenslanders like Ray), it seems to be used to give a cautious, pessimistic value by using the upper end of normal expected operating range. CCD's do not generate a great amount of heat AFAIK, especially the Full Frame type of CCD's in MF backs which are only active for a brief time during each exposure, not for the far longer periods involved in producing video output for viewfinders.

Bear in mind that Kodak publishes its sensors spec's for use by expert customers who buy the sensors directly, expecting them to be read by the technical staff at companies and scientific labs which incorporate the sensors into other devices such as MF backs. So there is very little room for fudge and spin in the way the spec's are presented.
[a href=\"index.php?act=findpost&pid=144468\"][{POST_SNAPBACK}][/a]
40°C is not that hot for a sensor. A range of 30° to 40°C would not be an unusal operating temperature. There is the heat generated by the sensor itself (which varies with integration/exposure time) plus the absorbed heat from ambient and surrounding components. If you place your hand by a fan or heat sink or other radiating body, you can feel a portion of the heat being dissapated.
« Last Edit: October 08, 2007, 05:58:48 am by TechTalk »
Logged
Respice, adspice, prospice - Look to the past, the present, the future

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
dynamic range
« Reply #61 on: October 09, 2007, 12:47:41 am »

Quote
Since 40C is quote hot (except perhaps by the standards of Queenslanders like Ray.....[a href=\"index.php?act=findpost&pid=144468\"][{POST_SNAPBACK}][/a]

40C is quite cool in the Simpson desert   .
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
dynamic range
« Reply #62 on: October 09, 2007, 12:52:16 am »

Quote
It's useful only as a point of reference under a specific set of conditions. Any test will only give you a reference or specification that represents a particular set of conditions.

More useful is knowing that the dynamic range of a sensor and camera system changes and is not fixed. It might also be useful, when considering dynamic range of a system, the rate at which dark current doubles.

If you don't find it useful information, O.K.. Just trying to help.
[a href=\"index.php?act=findpost&pid=144566\"][{POST_SNAPBACK}][/a]
 

As BJL says, it's useful for camera companies when selecting sensors to use for their own designs of cameras. It's not particularly useful for the photographer.

Knowing that the dynamic range increases substantially when the camera is cold, if this is the case, could be useful. For example, you could try placing your camera in the car freezer for a time prior to taking that important shot of a high dynamic range scene.  
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
dynamic range
« Reply #63 on: October 09, 2007, 04:37:51 am »

To continue with my reply (the slot machine internet connection having run out before I'd finished my previous response, not realising the hotel has a free wireless internet service - I happen to be in Khao Yai national park at the moment) I'd say that the whole dynamic range issue is not being adequately covered.

There are vague statements around, indicating that noise increases with temperature, but we have no visually empirical evidence to indicate just how significant such variations are.

I don't recall seeing any comments from members of Michael's group tours to the Antartic along the lines of, " Wow! Those sub-zero temperatures really expanded dynamic range. The deepest shadows in the deepest crevices are as clean as mid-tones."

However, I have come across comments on this site to the effect that Canon DSLRs do not appear to have any significant DR advantages in very cold climates. But I've not seen any rigorous testing of such issues. Maybe they do for all I know.

Maybe I'll just have to do the tests myself. Stick my 5D in the freezer then take a few shots, then a few more after the camera has warmed, ensuring the lighting is constant of course, which probably means taking the shots indoors with artificial lighting.
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
dynamic range
« Reply #64 on: October 09, 2007, 11:52:03 am »

Quote
40°C is not that hot for a sensor. A range of 30° to 40°C would not be an unusal operating temperature. There is the heat generated by the sensor itself (which varies with integration/exposure time) ...
[a href=\"index.php?act=findpost&pid=144567\"][{POST_SNAPBACK}][/a]

Have there not been tests, including by our host Michael Reichmann, showing that noise does not get noticeably worse with extended usage of an initially cool camera, suggesting that sensor heating is not much of an issue?

As I indicated previously, a medium Full Frame type CCD is only active for about 1 second per exposure, not all the time, so I very much doubt that it generates much heat. Likewise other digitally related components other than the rear LCD operate only during exposures, and non-digital stuff like AF and AE are not big heat producers. So the LCD seems the main likely culprit for heating a sensor to 40C (104ºF).

With smaller formats at least, is fairly well known that LCD's consume far more power than the rest of a camera's components, as indicated by the greatly reduced battery life when the LCD is used a lot compared to when it is used little. SO LCD's seem to be the main heat source. Could LCD heating really heat a sensor to 40C or beyond?
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
dynamic range
« Reply #65 on: October 10, 2007, 01:24:44 pm »

Quote
Bullcrap. There's also noise to consider, the LEAST significant of which is quantization error. With the vast majority of cameras out there (MFDB, DSLR, and digicams) the shot noise, read noise, dark current noise, and other forms of noise that end up in the RAW come into play before you start having quantization issues. Going from 12 bits to 14 bits is meaningless if the median RAW value is 128 (on a 14-bit scale) when shooting a dark frame, you have only 7 stops between the clipping point and the noise floor.
[{POST_SNAPBACK}][/a]

The above assertions are made in the absence of any data that back them up. First of all, dynamic range is defined as the full well capacity/read noise, both expressed in electrons. When one determines read noise with a dark frame exposure  at maximal shutter speed, there is no shot noise and thermal noise is negligible, so they are irrelevant to the discussion.

Read noise is strongly influenced by the ISO setting, with higher read noise at low ISO. Why is this the case? Look at [a href=\"http://www.clarkvision.com/imagedetail/evaluation-1d2/index.html]Figure 4[/url] in Roger Clark's analysis of noise for the Canon EOS 1D MII. Read noise at ISO 50 is 31 electrons but decreases asymptotically to about 4 electrons at high ISO. According to Roger's analysis, total read noise has two components: ADC noise introduced by the analog to digital converter and sensor read noise, which is introduced by the sensor. At low ISO, the ADC noise is predominant. At ISO 50, the gain of the camera is 26 electrons per 12 bit data number, so an error in the least significant bit (LSB) could represent up to 13 electrons. At ISO 1300, the gain is one electron per 12 bit data number, and the LSB error is negligible. A more rigorous mathematical analysis is given here.

With a 14 bit ADC, the gain at ISO 50 would be about 8 electrons, down from 31. A LSB error has less effect. The signal to noise ratio of an ideal ADC is given by the formula SNR = 6.02N+1.76 dB, where N is the bit depth of the device. The S/N of an ideal 14 bit ADC is 86 dB and that of a 12 bit device is 74 dB. Expressed in f/stops (log base 2 rather than log base 10), these ratios correspond to 14 and 9.2 f/stops respectively.
Logged

Nick-T

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 462
dynamic range
« Reply #66 on: October 10, 2007, 02:06:51 pm »

Quote
The above assertions are made in the absence of any data that back them up. First of all, dynamic range is defined as the full well capacity/read noise, both expressed in electrons. When one determines read noise with a dark frame exposure  at maximal shutter speed, there is no shot noise and thermal noise is negligible, so they are irrelevant to the discussion.

Read noise is strongly influenced by the ISO setting, with higher read noise at low ISO. Why is this the case? Look at Figure 4 in Roger Clark's analysis of noise for the Canon EOS 1D MII. Read noise at ISO 50 is 31 electrons but decreases asymptotically to about 4 electrons at high ISO. According to Roger's analysis, total read noise has two components: ADC noise introduced by the analog to digital converter and sensor read noise, which is introduced by the sensor. At low ISO, the ADC noise is predominant. At ISO 50, the gain of the camera is 26 electrons per 12 bit data number, so an error in the least significant bit (LSB) could represent up to 13 electrons. At ISO 1300, the gain is one electron per 12 bit data number, and the LSB error is negligible. A more rigorous mathematical analysis is given here.

With a 14 bit ADC, the gain at ISO 50 would be about 8 electrons, down from 31. A LSB error has less effect. The signal to noise ratio of an ideal ADC is given by the formula SNR = 6.02N+1.76 dB, where N is the bit depth of the device. The S/N of an ideal 14 bit ADC is 86 dB and that of a 12 bit device is 74 dB. Expressed in f/stops (log base 2 rather than log base 10), these ratios correspond to 14 and 9.2 f/stops respectively.
[a href=\"index.php?act=findpost&pid=145127\"][{POST_SNAPBACK}][/a]
Yeah I was just about to say that..
Logged
[url=http://www.hasselbladdigitalforum.c

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
dynamic range
« Reply #67 on: October 10, 2007, 03:41:04 pm »

The overall reason for my skepticism is that
1) Using a 14-bit ADC instead of 12-bit in the 1DMkII would have added very little to the cost of that camera,
2) There is plenty of room for a 14-bit ADC: after all Sony is putting thousands of 12-bit ADC's on its new EXMOR sensor.
3) Despite rather widespread 'internet arrogance' ("we know better how to design cameras than even the most successful camera making companies"), the engineers at Canon were probably quite aware of the idea of using a 14-bit ADC instead of 12-bit in the 1DMkII, and if that option really could have increased DR by up to three stops, and significantly reduced shadow noise at low ISO speeds, they probably would have done it!

Quote
According to Roger's analysis, total read noise has two components: ADC noise introduced by the analog to digital converter and sensor read noise, which is introduced by the sensor.
[a href=\"index.php?act=findpost&pid=145127\"][{POST_SNAPBACK}][/a]
As far as I can tell, Clark ignores the possibility that noise from the pre-amplifier contributes to an overall maximum attainable electrical S/N ratio. Pre-amp output signal with S/N ratio of about 4000 would itself limit noise in AD output to a minimum of about one level.

Clarks units of electrons are potentially misleading: he actually has data in terms of A/D converter output levels. So what the really has is something like
ISO 100: 1.28 "levels" of noise in A/D output, which he combines with gain of 13.02e per level in 12-bit output, to get 1.28x13.02e = 16.61e as his stated noise value.
ISO 3200: 9.59 "levels" of noise in A/D output, combined with gain of 0.41e per level to get his 3.93e of noise.
At other ISO levels, the actual noise measurements in A/D output ranges from 1.28 to 9.59 levels.

So Clark's data are consistent with a noise floor in the input to the ADU of not much less than 1.28 levels with 12-bit, or S/N not much better than 3190:1, independent of the ADU's performance.

In fact, such a limit seems the most likely reason why Canon did not bother to use a more precise ADC.
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
dynamic range
« Reply #68 on: October 10, 2007, 05:19:57 pm »

Quote
Clarks units of electrons are potentially misleading: he actually has data in terms of A/D converter output levels. So what the really has is something like
ISO 100: 1.28 "levels" of noise in A/D output, which he combines with gain of 13.02e per level in 12-bit output, to get 1.28x13.02e = 16.61e as his stated noise value.
ISO 3200: 9.59 "levels" of noise in A/D output, combined with gain of 0.41e per level to get his 3.93e of noise.
At other ISO levels, the actual noise measurements in A/D output ranges from 1.28 to 9.59 levels.
[{POST_SNAPBACK}][/a]
Well, obviously if you are determining noise from bias frames, the only thing you have to work with are the ADU's. One can determine the gain by standard methods and convert from ADU's to electrons. How is this misleading? Interested readers should refer to the link below for more information.

[a href=\"http://www.photomet.com/library/library_encyclopedia/library_enc_gain.php]Gain[/url]

Quote
So Clark's data are consistent with a noise floor in the input to the ADU of not much less than 1.28 levels with 12-bit, or S/N not much better than 3190:1, independent of the ADU's performance.
[a href=\"index.php?act=findpost&pid=145155\"][{POST_SNAPBACK}][/a]

According to the quoted formula for the S/N of a ADC, an ideal 12 bit ADC has a S/N of 74 dB, which is 5012:1 RMS. Of course no device is ideal and a S:N of 70 dB would  give  3162:1. Regardless of the S:N of the sensor, the ADC limitation would apply. To realize the full potential of a sensor with a higher S:N than this would require an ADC with a higher bit depth. I think you have it backwards.
« Last Edit: October 10, 2007, 05:25:10 pm by bjanes »
Logged

John Sheehy

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 838
dynamic range
« Reply #69 on: October 10, 2007, 08:55:48 pm »

Quote
As far as I can tell, Clark ignores the possibility that noise from the pre-amplifier contributes to an overall maximum attainable electrical S/N ratio. Pre-amp output signal with S/N ratio of about 4000 would itself limit noise in AD output to a minimum of about one level.
[a href=\"index.php?act=findpost&pid=145155\"][{POST_SNAPBACK}][/a]

That's part of the foundation for his conclusion that "current DSLRs are shot noise limited".  He blames most of the blackframe noise at low ISOs on the ADC, and then discounts it in his analysis.

The evidence doesn't fit his model, IMO.  First of all, Canon DSLRs have a number of transistors at each photosite, and if I remember correctly, I have read Canon statements that the amplification at the photosites is optimized for each main ISO.  When you look at the blackframe noises from Canon DSLRs that have 1/3 stop ISOs, you find that the noise in electrons is almost exactly the same for each ISO in a group of three, such as 100, 125, and 160, and then jumps to a new value with 200, 250, 320, etc.  The gain in the 5D and 1D* cameras is analog, as there are no ISO-related spikes or gaps in the histograms, so if the differences between ISOs were solely due to analog gain, there would be no highly variable groups of similar (electron unit) noise in threes.  That wouldn't make any sense.  Also, he doesn't seem to see much significance in the fact that the noise in electrons varies tremendously between the low ISOs on Canons, but does not vary much at all with other brands, like Nikon.  The Nikons fit the model he is presenting; the Canons do not.

I think it's just really hard to read a well with lots of photons, and do it well with current technology.  Readout is much easier with smaller full well values, and Canon is able to much better optimize the high ISOs.
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
dynamic range
« Reply #70 on: October 10, 2007, 10:17:17 pm »

Quote
According to the quoted formula for the S/N of a ADC, an ideal 12 bit ADC has a S/N of 74 dB, which is 5012:1 RMS ... To realize the full potential of a sensor with a higher S:N than this would require an ADC with a higher bit depth. I think you have it backwards.
[a href=\"index.php?act=findpost&pid=145177\"][{POST_SNAPBACK}][/a]

Why backwards? And why the hypothetical about a sensor with a higher S/N ratio than this when there is no evidence of the S/N ratio being better? The 1DMkIII and 1DsMkIII might instead have improved to the point that 14-bit conversion is useful.

Your 5012:1 RMS figure for a good 12-bit A/D converter simply reinforces the idea that the measured signal to noise levels of no better than 3200:1 are significantly lower than can be explained by only the limitations of a good 12-bit A/D convertor plus the less than 4e RMS of noise from the electron wells themselves.
One explanation is that Canon put a low quality A/D convertor into a $4,500 camera, wasting a stop or two of DR to save a few pennies. Another is that Clark has overlooked another noise source like pre-amplification, so that the 1DMkII's A/D convertor is not the limit on overall S/N and DR.
Logged

John_Black

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 264
    • http://www.pebbleplace.com
dynamic range
« Reply #71 on: October 11, 2007, 12:09:09 am »

Quote
That's part of the foundation for his conclusion that "current DSLRs are shot noise limited".  He blames most of the blackframe noise at low ISOs on the ADC, and then discounts it in his analysis.

The evidence doesn't fit his model, IMO.  First of all, Canon DSLRs have a number of transistors at each photosite, and if I remember correctly, I have read Canon statements that the amplification at the photosites is optimized for each main ISO.  When you look at the blackframe noises from Canon DSLRs that have 1/3 stop ISOs, you find that the noise in electrons is almost exactly the same for each ISO in a group of three, such as 100, 125, and 160, and then jumps to a new value with 200, 250, 320, etc.  The gain in the 5D and 1D* cameras is analog, as there are no ISO-related spikes or gaps in the histograms, so if the differences between ISOs were solely due to analog gain, there would be no highly variable groups of similar (electron unit) noise in threes.  That wouldn't make any sense.  Also, he doesn't seem to see much significance in the fact that the noise in electrons varies tremendously between the low ISOs on Canons, but does not vary much at all with other brands, like Nikon.  The Nikons fit the model he is presenting; the Canons do not.

I think it's just really hard to read a well with lots of photons, and do it well with current technology.  Readout is much easier with smaller full well values, and Canon is able to much better optimize the high ISOs.
[a href=\"index.php?act=findpost&pid=145226\"][{POST_SNAPBACK}][/a]

ISO 50 on the 1Ds2 and 1D2 is a software trick; their base ISO is 100.  50 and 3200 are available via custom function.
Logged

John Sheehy

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 838
dynamic range
« Reply #72 on: October 11, 2007, 05:14:21 pm »

Quote
ISO 50 on the 1Ds2 and 1D2 is a software trick; their base ISO is 100.  50 and 3200 are available via custom function.
[a href=\"index.php?act=findpost&pid=145253\"][{POST_SNAPBACK}][/a]

Actually, the base ISO of the 5D and 1Dmk2 is about 70.  ISO 100 on these cameras does not cover the full range of sensor charges.  In that sense, even though 50 is 50 only as far as metering is concerned (and not as far as highlight headroom is concerned), it still has something to offer that 100 doesn't (lower shot noise with higher absolute exposure), so calling it a trick is not totally accurate.
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
dynamic range
« Reply #73 on: October 12, 2007, 04:33:11 pm »

Quote
One explanation is that Canon put a low quality A/D convertor into a $4,500 camera, wasting a stop or two of DR to save a few pennies. Another is that Clark has overlooked another noise source like pre-amplification, so that the 1DMkII's A/D convertor is not the limit on overall S/N and DR.
[{POST_SNAPBACK}][/a]

I doubt that Clark overlooked another source of noise, since he measured total noise and its two main components under the shooting conditions, shot noise, and read noise. Shot and read noise accounted for the total noise, so it is not likely that any noise source was overlooked.

Canon is a leader is sensor technology and I'm sure that they have very good engineers, who exercised sound professional judgement when they used a 12 bit ADC in the EOS 1D MII.

There are two threads on the usenet where Roger discusses some of these factors with some knowledgeable other people:

rec.photo.digital: megapixels and noise
rec.photo.digital.slr-systems: read noise: get it 30x down

Here are quotes from Roger:

[discussing why 16 bit ADCs are not used in current 35mm style DSLRs]
"Of course there are other penalties too, such as slower
processing due to the extra bits, and the higher memory
requirements, none of which actually contribute to
useful image improvement.  Indeed, that is why 12-bit
A/D's were used until the most recent cameras arrived,
even though the sensors could provide slightly better
dynamic range (by a couple dB) than a 12-bit A/D.  The
advantage of going to a 14-bit A/D just was not worth
the cost, at that time.  Now, with more parallel
channels from sensors, and 8 Gb vs.  1 Gb CF cards, the
advantages of a 14-bit A/D have a lower cost, and are
viable.

14-bits gives an 84dB range, but unfortunately, A/D converters
above 12 bits that have to work as fast as those in digital
cameras (several tens of megahertz) are not that accurate.
For example, see:
[a href=\"http://www.analog.com/IST/SelectionTable/?selection_table_id=124]http://www.analog.com/IST/SelectionTable/?...on_table_id=124[/url]
for an extensive list of current A/D products (Analog Devices
is a major manufacturer).  Look at lower power devices that
run in the ~30 to 100 megahertz range.
12 bit A/Ds:  ~68 - 71 dB SNR (12-bit range = 72.2 dB)
14 bit A/Ds:  ~70 - 75 dB SNR (14-bit range = 84.3 dB)
16 bit A/Ds:  ~70 - 80 dB SNR (16-bit range = 96.3 dB)"


The newer cameras with 14 bit ADCs apparently have more parallel readout channels, which places less of a burden on the ADC.

Bill
Logged

TechTalk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3612
dynamic range
« Reply #74 on: October 12, 2007, 08:03:22 pm »

Quote
As BJL says, it's useful for camera companies when selecting sensors to use for their own designs of cameras. It's not particularly useful for the photographer.

Knowing that the dynamic range increases substantially when the camera is cold, if this is the case, could be useful. For example, you could try placing your camera in the car freezer for a time prior to taking that important shot of a high dynamic range scene. 
[{POST_SNAPBACK}][/a]


Quote
There are vague statements around, indicating that noise increases with temperature, but we have no visually empirical evidence to indicate just how significant such variations are.

I don't recall seeing any comments from members of Michael's group tours to the Antartic along the lines of, " Wow! Those sub-zero temperatures really expanded dynamic range. The deepest shadows in the deepest crevices are as clean as mid-tones."

However, I have come across comments on this site to the effect that Canon DSLRs do not appear to have any significant DR advantages in very cold climates. But I've not seen any rigorous testing of such issues. Maybe they do for all I know.

Maybe I'll just have to do the tests myself. Stick my 5D in the freezer then take a few shots, then a few more after the camera has warmed, ensuring the lighting is constant of course, which probably means taking the shots indoors with artificial lighting.
[a href=\"index.php?act=findpost&pid=144787\"][{POST_SNAPBACK}][/a]


Quote
Have there not been tests, including by our host Michael Reichmann, showing that noise does not get noticeably worse with extended usage of an initially cool camera, suggesting that sensor heating is not much of an issue?

As I indicated previously, a medium Full Frame type CCD is only active for about 1 second per exposure, not all the time, so I very much doubt that it generates much heat. Likewise other digitally related components other than the rear LCD operate only during exposures, and non-digital stuff like AF and AE are not big heat producers. So the LCD seems the main likely culprit for heating a sensor to 40C (104ºF).

With smaller formats at least, is fairly well known that LCD's consume far more power than the rest of a camera's components, as indicated by the greatly reduced battery life when the LCD is used a lot compared to when it is used little. SO LCD's seem to be the main heat source. Could LCD heating really heat a sensor to 40C or beyond?
[a href=\"index.php?act=findpost&pid=144834\"][{POST_SNAPBACK}][/a]
As long you're assuming, speculating, and as far as I knowing in regards to how much warmer a CCD becomes than the ambient temperature–maybe you could tell me how much higher, if any, you believe a CCD's operating temperature is above ambient. I've talked to a couple of engineers about this subject (from two different companies) and was told that 10-20°C would not be unusual, but that CCD temperature above ambient changes due to a host of variables.

As for a simple visual test, I did a google search (ccd temperature effects) and found this as the very first hit... [a href=\"http://www.dpreview.com/news/0005/00050104ccdtemperature.asp]Camera CCD Temp Test Link[/url]  You can dismiss the comparison as of no value because it was done with a compact digital camera if you like, but I'm not going to invest a lot of additional time in trying to convince anyone that heat and dynamic range have a strong link. If you don't want to believe it then I can't change your mind.

For a more scholarly discussion, this article on CCD dynamic range by the Dept. Head of the Imaging Devices Group at Philips (now Dalsa) gives a thorough discussion of CCD dynamic range specifications and considerations...CCD Dynamic Range Specifications Link  If you don't want to read the text, you can skip to "figure 2" in the article which has a chart of temp. to dynamic range.

For those that don't want to believe that CCD temp. matters, you can dismiss this article for using a 6MP sensor that is out of date as an example for the article's discussion. You may also want to note, however, that the CCD shown in the article is LESS sensitive to rising temperature effects than current 39MP sensors.

On a practical note, if your hands get cold when shooting with a digital back. You can warm them next to the fan or heat-sink. Just don't ask where the heat is coming from.
« Last Edit: October 12, 2007, 08:09:34 pm by TechTalk »
Logged
Respice, adspice, prospice - Look to the past, the present, the future

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
dynamic range
« Reply #75 on: October 13, 2007, 04:19:32 am »

Quote
As for a simple visual test, I did a google search (ccd temperature effects) and found this as the very first hit... Camera CCD Temp Test Link You can dismiss the comparison as of no value because it was done with a compact digital camera if you like, but I'm not going to invest a lot of additional time in trying to convince anyone that heat and dynamic range have a strong link. If you don't want to believe it then I can't change your mind.

TechTalk,
I wouldn't dismiss the dpreview test on the grounds that it used a P&S camera, but I'm a bit dubious about the relevance of a test carried out in the year 2000. The photosites were CCD and the technology was in its infancy.

This is not a matter of belief (why introduce religion?), but a matter of empirical evidence. I agree there is a principle that noise will increase as temperature rises, whether or not that rise is due to an increase in ambient temperature or due to heavy, continuous use of the camera.

There is also a principle that small sensors (photosites) are noisier than larger ones, but I doubt that a Canon 40D pixel is noisier than a Canon D30 pixel (that's the 3mp D30, Canon's first DSLR).

As far as I can see, technology is all about understanding the priciples and limitations of nature, and then overcoming them.

I look forward to the day when the laws of diffraction will be circumvented.
Logged

Diapositivo

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 57
    • http://
dynamic range
« Reply #76 on: October 13, 2007, 07:16:58 am »

Quote
Have there not been tests, including by our host Michael Reichmann, showing that noise does not get noticeably worse with extended usage of an initially cool camera, suggesting that sensor heating is not much of an issue?

As I indicated previously, a medium Full Frame type CCD is only active for about 1 second per exposure, not all the time, so I very much doubt that it generates much heat. Likewise other digitally related components other than the rear LCD operate only during exposures, and non-digital stuff like AF and AE are not big heat producers. So the LCD seems the main likely culprit for heating a sensor to 40C (104ºF).

With smaller formats at least, is fairly well known that LCD's consume far more power than the rest of a camera's components, as indicated by the greatly reduced battery life when the LCD is used a lot compared to when it is used little. SO LCD's seem to be the main heat source. Could LCD heating really heat a sensor to 40C or beyond?
[a href=\"index.php?act=findpost&pid=144834\"][{POST_SNAPBACK}][/a]

I would say by instinct that considerable heat should be generated in the camera by the JPEG processing engine. That crunches numbers and it must do it fast, so it run at high clock speed and is probably the warmer piece of electronic inside the camera.

Cheers
Fabrizio
Logged
Pages: 1 2 3 [4]   Go Up