GH1 colour space sRGB or Adobe RGB

[Otis Grapsas]

The video itself never has a wide color space. It's all an interpretation of ITU709.

If the image looks different, it's because the color space is wrong in the AdobeRGB mode. You will see less saturation and you might like it, but find another way to reduce saturation, your colors are all over the place if the camera indeed stores adobeRGB on a video file. Using adobergb in video will make a camera store a larger color space in the same data, this will give you less precision and inferior gradients. You final video will be interpreted in ITU709/SRGB color so it makes sense to capture only what you can deliver.

The stills in video mode could be adobergb. In order to see these correctly you need Safari or Firefox. Internet explorer shows adobergb images in srgb so they appear a lot less saturated and their color is massively distorted.

Adobergb is not very useful for the web. In a best case scenario, when there is no color that exceeds srgb and when the viewing software can correctly interprete adobergb, the image will be identical to an srgb image. If one of the two is not true you are seeing something different than everyone else. Different colors. Bad reference. Bad monitoring, and it's not even your fault. Adobergb is quite rare, not even 10% of the monitors out there have good coverage, it's not safe to assume it will look right. Srgb is a lot safer. A wider color space is not a better colorspace. It's a compromise we take when we do need the color space to be wide for an application.

 

[j1clark@ucsd.edu]

Interesting. I'll have to shoot my own tests with both color spaces and see how they compare. But if I'm not mistaken, once the image is encoded to AVCHD, technically it is no longer sRGB or AdobeRGB. It's in that 4:2:0 AVCHD colorspace, which is maybe REC709?
It is intersting that there's a difference in the recorded video image, but sRGB and AdobeRGb are irrelevant when talking about editing, because AVCHD, Pro Res, etc. are not in those colorspaces. So, the only relevance is how it affects the image as it gets recorded. But that IS significant. I thought it was only for stills and had no effect on the video. I'm definitely going to do a test!

From the manual, there an indication that for 'motion pictures' the color space is set to sRGB.

So, I'm wondering how people are 'seeing' a difference, or even how they 'know' their media is in either color space.

When I look at the 'video color space' metadata for a recent capture, the metadata field is blank, at least when displayed in CS5 Premiere... The EXIF info for colorspace is also blank.

 

[Otis Grapsas]

From the manual, there an indication that for 'motion pictures' the color space is set to sRGB.

So, I'm wondering how people are 'seeing' a difference, or even how they 'know' their media is in either color space.

When I look at the 'video color space' metadata for a recent capture, the metadata field is blank, at least when displayed in CS5 Premiere... The EXIF info for colorspace is also blank.

You can only tell a difference in color space if the color space is not interpreted correctly. If the image is shown in the wrong color space. Or if a color is out of the gamut of the smaller color space and you have a great eye (rare).

There is the possibility that when a camera is in a different color space in still mode, it uses entirely different gamma curves in its video mode. That would be a bug. It could also be that the video mode indeed stores a different color space. That would also be a bug. If the manufacturer wanted us to have an option in saturation or gamma, they could just add documented options in the image settings.

I looked at this thread:

http://www.dvxuser.com/V6/showthread.php?t=203804

I would say the "after" image uses a larger aperture (check the DOF difference in otherwise equivalent exposures, the eos ad is equally sharp but everything outside the window is out of focus) and a lot lower gain (check the noise level). A color space change, and the minor saturation and contrast differences shown cannot create this difference in noise.

 

[Svart]

From the manual, there an indication that for 'motion pictures' the color space is set to sRGB.

So, I'm wondering how people are 'seeing' a difference, or even how they 'know' their media is in either color space.

When I look at the 'video color space' metadata for a recent capture, the metadata field is blank, at least when displayed in CS5 Premiere... The EXIF info for colorspace is also blank.

That's what I said in the other thread. I'd swear up and down that I didn't see any difference at all between the two settings when I tried them. I figured that it didn't make a difference in video mode.

 

[Tesselator]

If you want to look at different color spaces over the web it's very worth installing Safari. It free for windows and OS X. I haven't tried it on windows yet but in OS X it displays every color space. And if PS is your reference you'll be happy because the outputs match more closely than any other browser. I've tested with 12 different color spaces across 2 different color models. IE gott only one right, the new FF got 2 (and a half ) right, Opera got two right, and Safari nailed all of them. I didn't try Chome. So even if you non't use Safari as an everyday browser for the web it's useful to have for checking stuff. I assume it behaves the same on Windows.

Here it is: http://www.apple.com/safari/download/


.

 

[ChipG]

If you want to look at different color spaces over the web it's very worth installing Safari. It free for windows and OS X. I haven't tried it on windows yet but in OS X it displays every color space. And if PS is your reference you'll be happy because the outputs match more closely than any other browser. I've tested with 12 different color spaces across 2 different color models. IE gott only one right, the new FF got 2 (and a half ) right, Opera got two right, and Safari nailed all of them. I didn't try Chome. So even if you non't use Safari as an everyday browser for the web it's useful to have for checking stuff. I assume it behaves the same on Windows.

Here it is: http://www.apple.com/safari/download/


.

Apple just released Safari 5 about two hours ago, it is 30% faster than Safari 4 and they are saying 3% faster than Chrome, 50% faster than Firefox (I think that number is for an older version of Firefox tho).

I'm using it right now, it is a better / faster multimedia viewing experience.

 

[Otis Grapsas]

http://www.eoshd.com/blogs/1-EOSHD.com?tag=deep+field

Is there any hard proof for the adobe RGB+spot metering hypothesis?

 

[Adventsam]

Spot metering, where did that come from?

 

[Otis Grapsas]

The site automatically replaces some links to articles, so that link didn't get through.

But spot meter is everywhere on the net. It's one of those rumours that are accepted as being true when a cult is formed, just because nobody questions the wisdom of those who create the cult

It's impossible that metering method has an effect on camera processing, but there are so many reference to those "golden" settings. You know, this guy only uses that setting, and that guy said that setting is better, and I have seen that camera on the set of that show/film and so on. Only a few bother to test though. I hope they are not excommunicated because of the tests

 

[Vitaliy Kiselev]

This is people, Otis.
They like to mimic and follow the lead.
So, "golden settings" will be always present.

 

[Lpowell]

On color space:

Using the wider AdobeRGB color space (top diagram) will not necessarily produce more accurate images than using sRGB (bottom diagram). Check out the green and cyan areas in the top diagram that extend beyond the AdobeRGB triangular color space. Unless you have a calibrated wide-gamut monitor, you likely won't see any discernible difference between these extremely saturated colors and the HSL limits of the two color spaces. Consumer photo printers have similarly limited color gamuts as well.

When in AdobeRGB mode, the GH1 is using the same amount of bits to encode a broader color gamut than when it's in sRGB mode. For that reason, AdobeRGB can actually degrade the camera's ability to discriminate shades within the subtle color gradients that do fall within the sRGB color space. When using heavy data compression techniques on captured image data, reducing the contrast range and chroma gamut often produces more detailed results. While AdobeRGB is technically superior to sRGB in processing uncompressed raw data streams recorded with greater than 24-bit RGB precision, that level of fidelity is far beyond the capabilities of the GH1.

 

[Blaine]

Apple just released Safari 5 about two hours ago, it is 30% faster than Safari 4 and they are saying 3% faster than Chrome, 50% faster than Firefox (I think that number is for an older version of Firefox tho).

I'm using it right now, it is a better / faster multimedia viewing experience.

I just uploaded Safari 5. Nice. I had to jump through some hoops though since I was running a 3rd party toolbar. It caused Safari 5 to crash. After deleting the "Conduit" folders and voila, everything's good.

 

[alignment1]

Using the wider AdobeRGB color space (top diagram) will not necessarily produce more accurate images than using sRGB (bottom diagram). Check out the green and cyan areas in the top diagram that extend beyond the AdobeRGB triangular color space. Unless you have a calibrated wide-gamut monitor, you likely won't see any discernible difference between these extremely saturated colors and the HSL limits of the two color spaces. Consumer photo printers have similarly limited color gamuts as well.

When in AdobeRGB mode, the GH1 is using the same amount of bits to encode a broader color gamut than when it's in sRGB mode. For that reason, AdobeRGB can actually degrade the camera's ability to discriminate shades within the subtle color gradients that do fall within the sRGB color space. When using heavy data compression techniques on captured image data, reducing the contrast range and chroma gamut often produces more detailed results. While AdobeRGB is technically superior to sRGB in processing uncompressed raw data streams recorded with greater than 24-bit RGB precision, that level of fidelity is far beyond the capabilities of the GH1.

I haven't actually thrown in a timeline yet and checked on an HD montior but in cam on the lcd...I see definite difference. Without question, using AdobeRGB, there is more detail in highlights/shadows.
I've done a number of sided by side tests- not moving the cam and shooting foe 3-5 sec. in both color spaces. When I playback in cam- AdobeRGB shows detail in a reflection from a picture frame - which is otherwise blown out white in sRGB.
In another instance - There is a shaded hallway where the wall(beige color)meets the ceiling(white color)...in sRGB it's gray mush, in AdobeRGB I can clearly see the color difference and the fine line which connects the 2 surfaces...

I'm awfully curious to get these in the computer- I'll post results, when I get some time.

 

[ChipG]

I just uploaded Safari 5. Nice. I had to jump through some hoops though since I was running a 3rd party toolbar. It caused Safari 5 to crash. After deleting the "Conduit" folders and voila, everything's good.

Remove all 3rd party tool bars and it is really fast!

 

[Blaine]

Indeed. Using it right now. :thumbsup:

 

[Tesselator]

Using the wider AdobeRGB color space (top diagram) will not necessarily produce more accurate images than using sRGB (bottom diagram). Check out the green and cyan areas in the top diagram that extend beyond the AdobeRGB triangular color space. Unless you have a calibrated wide-gamut monitor, you likely won't see any discernible difference between these extremely saturated colors and the HSL limits of the two color spaces. Consumer photo printers have similarly limited color gamuts as well.

I've read this argument a lot on the net at various forums and while wide gamut monitors will improve some things this rumor itself is untrue. I have dual 24" LG cheap-o TN type monitors (model no. L246WH) and the differences between sRGB and ProPhoto can be astounding. I say "can be" because it depends on the photo and the original colors captured. If there is a large percentage of pixels that fall outside the sRGB "triangle" then the image will be severally degraded by applying sRGB to it. It will retain all of (more of) the original colors when using ProPhoto - even as displayed on my lowest possible quality monitors.

When in AdobeRGB mode, the GH1 is using the same amount of bits to encode a broader color gamut than when it's in sRGB mode. For that reason, AdobeRGB can actually degrade the camera's ability to discriminate shades within the subtle color gradients that do fall within the sRGB color space.

Yes, this is partially correct. But I would rather have broader ranges than more precise narrower ones. For most of the still photos I work with broader ranges are more desirable than minutely more precise narrower ranges. Thus why so many applications and artists suggest editing in the widest color space available and ONLY reducing to the common (least common denominator) sRGB for file sharing and exchange (like web posting and etc.)

When using heavy data compression techniques on captured image data, reducing the contrast range and chroma gamut often produces more detailed results. While AdobeRGB is technically superior to sRGB in processing uncompressed raw data streams recorded with greater than 24-bit RGB precision, that level of fidelity is far beyond the capabilities of the GH1.

This shouldn't be the case. Both color spaces were designed around 8 bit ("24 bit") precisions. It's not until we look at much wider spaces like ProPhoto and WideRGB that we see color spaces intended for higher precision transformations. <shrug>

 

[photoSmart42]

The professional printer shop that makes prints for me recommended I use sRGB instead of AdobeRGB because it was easier to match print colors with sRGB, so since then I changed to sRGB for my still. I can't say I noticed a difference myself, but this back and forth between the two format is interesting to read.

 

[Tesselator]

Yeah, notice that they said "easier" not better!

And yup! In 8 out of 10 photos you won't see ANY difference AT ALL no matter what the output device is. It's that other 20% where the gotchas get ya.

 

[alignment1]

The professional printer shop that makes prints for me recommended I use sRGB instead of AdobeRGB because it was easier to match print colors with sRGB, so since then I changed to sRGB for my still. I can't say I noticed a difference myself, but this back and forth between the two format is interesting to read.

Although these pesky little cameras have a 'cross-mojination to the 2nd degree with optional flux capacitor':laugh:

it certainly collides the worlds of still photography and motion video--

however, I think the thread's purpose was intended for video inquiries--and how these color spaces affect the camera in 'movie mode'.

correct me if I'm wrong here someone, I just don't want to cross mojinate off the beaten path:cheesy:

 

[Tesselator]

Although these pesky little cameras have a 'cross-mojination to the 2nd degree with optional flux capacitor':laugh:

it certainly collides the worlds of still photography and motion video--

however, I think the thread's purpose was intended for video inquiries--and how these color spaces affect the camera in 'movie mode'.

correct me if I'm wrong here someone, I just don't want to cross mojinate off the beaten path:cheesy:

All mojination aside, how color spaces affect still images will also be how they affect video. It's not like it's suddenly a whole different thing just because we're talking about video - with or without flux-capacitation.

If indeed it's applied at all... (See message 19, 21 & 22)