|
|||||||||
|
Thread Tools | Search this Thread |
April 3rd, 2008, 01:15 PM | #1 |
Major Player
Join Date: Jan 2006
Location: Göteborg, Sweden
Posts: 306
|
Is -3 Db gain = 0 Db gain
As XL H1 owner I always shoot @ -3 Db gain as the consensus at the forum was that the -3 Db gain on the H1 is more like 0 Db gain; Same situation withe Xl A1?
__________________
Jonas Nyström, DoP :: HOT SHOT® SWEDEN :: www.hotshot.nu :: RED #1567, RED 18-50mm T3 :: XL A1, Letus Extreme :: XL H1, 20X & 6X lens (for sale) :: www.vimeo.com/nystrom Last edited by Jonas Nystrom; April 3rd, 2008 at 03:47 PM. |
April 3rd, 2008, 05:56 PM | #2 |
Trustee
Join Date: Jun 2006
Location: Burbank
Posts: 1,811
|
I have been very curious about what minus gain is.
Odb gain I guess would be the natural chip, with no enhancement. However, what is -3db gain? Obviously it must be the natural chip with electronic ungain applied. Is that right? Or, as the first post suggests, is -3db just the Canon name for 0db? |
April 3rd, 2008, 06:21 PM | #3 |
Major Player
Join Date: Jan 2007
Location: Portland, OR
Posts: 949
|
It trades noise for dynamic range.
|
April 3rd, 2008, 08:03 PM | #4 |
Trustee
Join Date: Jan 2006
Location: New York, NY
Posts: 1,267
|
0DB gain is what the electronics are set to in the camera based on the manufacturers specs for the units sensors when using the reference amount of light and Iris settings. Just because they use the term 0 don't think the amplifiers getting the signal from the sensors are off. They still could be turned higher or lower.
One model cameras 0DB might look like another models 6DB. You might find one model camera with a lit chart will give you full exposure at F5.6 at 0DB while another model might be at F4 at 0DB. The amount of noise in each picture is part of the equation the manufacture can play with to position the cameras for the market they intend to sell into. The sensors and electronics work together to arrive at these numbers. If the manufacturer wants its signal to noise ratio for the model to be 64DB with a sensitivity of F8 at 200 footcandles then they have to set a normal position for the amplifiers which is called 0DB. If they wanted to make the signal to noise ratio 58DB with the same sensors and amps then the sensitivity would be F11 a stop faster with the same 200 footcandles of light. If they want the 64DB sensitivity adjusting the electronic gain up and down with gain switches gives the cameras more flexibility to work in different lighting situations. From this example when you use the +6DB setting on the cameras signal to noise drops 6DB to 58DB but the sensitivity goes up. The -3DB setting would work the other way making the camera have a signal to noise ratio of 67DB but a corresponding lower sensitivity to light. Depending on the scene you are shooting or the quality of the camera you are using you might not see a visual difference between 0DB and -3DB but you should see a difference in the Iris level which is very useful for controlling the picture. As high end video cameras improved.. the signal to noise ratio went higher and higher to the point that it was hard to shoot with the cameras outdoors without using F stops which were too high to produce clear pictures so besides ND's turning down the gain on the sensors amplifiers was a way of controlling the cameras pictures. Consumer cameras use similar terminology but often have a much lower signal to noise ratio to start with so it is easier to see the electronic gain differences and you might find the manufacturers taste a little to noisy at 0 but acceptable at -3 |
April 4th, 2008, 07:52 AM | #5 |
Wrangler
Join Date: Sep 2001
Location: Northern VA
Posts: 4,489
|
As noted by other, 0 dB gain is the manufacturer's nominal normal (standard) setting for gain. The zero means it is the baseline setting, and other setting are measured as the ratio from that baseline. -3 dB mean the gain is amplifiers is reduce by 3 DB (a factor of about 0.7), and +6 dB gain means it is increased by 6 dB (about 2x) compared to the baseline setting.
The 0 dB setting is typically (or should be) the makers best compromise setting factoring dynamic range (latitude), signal to noise, linearity, sensitivity, etc. (i.e.,image quality) for the camcorder considering it target market segment. When using a camcorder at other than 0 dB gain some aspect of the makers design performance baseline is compromised. Using lower gain (e.g., -3dB) means more light is needed for the same image brightness, and that means the sensor is closer saturation on highlights, but shadows should have less noise. Using higher gain (e.g., +6 dB) means that for the same image brightness the image sensor is has more head room for highlights, but the noise floor is increased. Gain is a lot like push/pull processing of film, and it allows you to use a different exposure while retaining the same nominal image brightness brightness. E.g., for the same image brightness and shutter speed, adding 6 dB of gain allows you to close the iris by a stop. The question for the videographer is whether or not the image artifacts added or changed by using a gain setting other than 0 dB are acceptable. Many folks prefer to use -3 db for the improved shadow noise.
__________________
dpalomaki@dspalomaki.com |
April 4th, 2008, 07:57 AM | #6 |
Great info, Don. Looking at my EX1 on the waveform monitor, the SAW curve has a knee moving into clipping at around 90IRE in the -3dB mode. Switch to 0dB and the knee/clip moves to 100IRE. Definitely indicates a reduction in dynamic range. Good to note that the offsetting effect is lower shadow noise.
|
|
April 6th, 2008, 03:30 PM | #7 |
Regular Crew
Join Date: Aug 2006
Location: Lugano, Switzerland
Posts: 149
|
IMO it seems that the -3dB is adding black instead of removing noise.
I did some tests and i ended up leaving the gain to 0dB, because that added black was slightly noisy. I just make darker blacks in FCP |
April 6th, 2008, 05:15 PM | #8 | |
Trustee
Join Date: Dec 2004
Location: Lipa City Batangas, Philippines
Posts: 1,110
|
Quote:
Richard |
|
April 7th, 2008, 06:37 AM | #9 |
Wrangler
Join Date: Sep 2001
Location: Northern VA
Posts: 4,489
|
"Adding black" can be a result of changes to the effective black, pedestal and set-up settings.
Changing gain from 0 dB to -3 dB, keeping other exposure parameters the same (shutter and aperture) will appear to "add black" as the image become darker and deep shadows get deeper. Also, less noise in the shadows can make them look blacker. However, if exposure is increased by a half stop (shutter or aperture step or more lighting) the net effect is that image should end up the same brightness, and CCD noise (i.e., thermal noise and dark current) should end up about 3 dB lower. The effect on noise that originates in the analog amplifiers that read the CCD is more difficult to predict because it depends on how the gain control is implemented and on the noise characteristics of the amplifiers at various gain settings. Note that using slow shutter speeds can result in an increase in the the CCD dark current noise component. I've not conducted any tests to determine where this becomes a significant factor, but it is probably at shutter speeds below 1/30.
__________________
dpalomaki@dspalomaki.com |
April 7th, 2008, 09:14 AM | #10 | |
Regular Crew
Join Date: Aug 2006
Location: Lugano, Switzerland
Posts: 149
|
Quote:
So do you recommend using -3dB when possible to minimize noise? |
|
April 7th, 2008, 09:29 PM | #11 | |
Trustee
Join Date: Jan 2006
Location: New York, NY
Posts: 1,267
|
Quote:
There may be a way to set the knee clip settings to be optimal in -3DB to regain some dynamic range but then they might not be correct in 0DB. Does the cameras menu let you set the knee or white clip levels or is it an internal factory set item? |
|
April 8th, 2008, 07:02 AM | #12 | ||
Wrangler
Join Date: Sep 2001
Location: Northern VA
Posts: 4,489
|
Quote:
Note that with gain = -3 dB, you are giving up 3 dB of "head room" on the CCD output for the same overall image brightness. Whether or not this is a issue will depend on the image content. Quote:
There are also menu settings for: Gamma, Master Pedestal, Black Stretch/Press, and Setup Level.
__________________
dpalomaki@dspalomaki.com |
||
April 8th, 2008, 07:27 AM | #13 |
Regular Crew
Join Date: Aug 2006
Location: Lugano, Switzerland
Posts: 149
|
I made a test yesterday. A very dark shot, at 0dB and -3dB.
Personally, and i will post the clips tonight in order to let you judge, maybe i'm wrong, i don't see any improvement. Instead when i set to -3dB i see some more black mosquito, but the noise seems to stay. See you later with the clips. |
April 8th, 2008, 08:05 AM | #14 |
Trustee
Join Date: Jan 2006
Location: New York, NY
Posts: 1,267
|
Giovanni,
If the scene is very dark and you can't get a proper exposure with -3DB and 0DB by adjusting the iris then you are probably not going to be testing the difference between the two. Underexposed footage versus even more underexposed footage is not going to reveal the small difference which can be seen between 0db and -3db. You really need a scene where you get a proper exposure with both settings and the difference is the Iris. It is really meant for when you have enough light and want to finesse your Iris setting differently than 0db provides. |
April 8th, 2008, 09:58 AM | #15 |
Wrangler
Join Date: Sep 2001
Location: Northern VA
Posts: 4,489
|
Keep in mind that 3 dB is a relatively small difference; i.e.,about 1.0 vs. 0.7, and while it can be seen in side-by-side comparisons, it will probably not be dramatic.
If the -3 dB gain is compensated by increased exposure time (slower shutter) the dark current component of the noise (that appears as a fixed grain pattern) read off the CCD would increase by 3 dB, due to the longer integration time, and thus net to about zero change in this noise component. In this case it is the thermal noise component that would decrease by about 3 dB. While the above is somewhat of an oversimplification, your overall image noise improvement using -3 dB gain will range from a maximum improvement approaching 3 dB to a minimum that depends on a variety of parameters including the shutter speeds you are using. There is no substitute for adequate light, and best results are obtained when the camera operator knows how his/her gear performs best under various lighting conditions.
__________________
dpalomaki@dspalomaki.com |
| ||||||
|
|