Image thickness / density - help me figure out what it is

[kye]

On 10/4/2020 at 12:54 PM, tupp said:

No.  There is no gain of color depth with down-sampling.  The color depth of an image can never be increased unless something artificial is introduced.

On the other hand resolution can be traded for bit depth.  So, properly down-sampling (sum/average binning adjacent pixels) can increase bit depth with no loss of color depth (and with no increase in color depth).

This is interesting.

If colour depth is colours above a noise threshold (as outlined here: https://www.dxomark.com/glossary/color-depth/ ) then that raises a few interesting points:

• humans can hear/see slightly below the noise floor, and noise of the right kind can be used to increase resolution beyond bit-depth (ie, dither)
• that would explain why different cameras have slightly different colour depth rather than in power-of-two increments
• it should be effected by downsampling - both because if you down downsample you reduce noise and also because downsampling can average four values artificially creating the values in-between
• it's the first explanation I've seen that goes anywhere to explaining why colours go to crap when ISO noise creeps in, even if you blur the noise afterwards

I was thinking of inter-pixel treatments as possible ways to cheat a little extra, which might work in some situations.

Interestingly, I wrote a custom DCTL plugin that simulated different bit-depths, and when I applied it to some RAW test footage it only started degrading the footage noticeably when I got it down to under 6-bits, where it started adding banding to the shadow region on the side of the subjects nose, where there are flat areas of very smooth graduations.  It looked remarkably like bad compression, but obviously was only colour degradation, not bitrate.

On 10/4/2020 at 12:54 PM, tupp said:

That is largely true, but I am not sure if "good" lighting is applicable here.  Home movies shot on film with no controlled lighting have the "thickness" that OP seeks, while home movies  shot on video usually don't have that thickness.

True, although home videos are often filmed in natural light, which is the highest quality light of all, especially if you go back a decade or two where film probably wasn't sensitive enough to film much after the sun sets and electronic images were diabolically bad until very very recently!

 

[mrtreve]

I do think this is real. When I see RED images I always think they look so thicc (sometimes not in an entirely pleasing way, but thicc nonetheless).

Perhaps this comes form the trend in flat grades that came about with the widespread adoption of log in digital cameras. Then you couple that with the poor 8 bit codecs that break when you put contrast back in past a certain point, meaning people with Sony A7's etc got used to putting out thin or muddy looking images.

Then I think you can lump in camera inertia, rolling shutter & motion cadence. I think these can add/subtract the feeling of weightiness.. with film being the best at this. I know this isn't exactly related to 'density' but I think they're linked quite a bit.

If I was trying to address this on a cheaper camera I might:

- Shoot raw or externally record to prores so you can push lots of contrast back in.
- Monitor with a contrasty lut and light within this range (also shoot under cloud cover / use haze).
- Use some filtration so that any transitions to pure white are smoothed out (don't be afraid of dense shadows or bright highlights)
- Turn off ibis. Rig the camera to be large and heavy.
- Put the camera out in front of you so the axis of the lens movement somewhat forward from the centre of gravity (imagine how far the lens is from the CG of an Alexa Classic.
- Have a reference frame grab from a dense looking film in Resolve and try to match it
- Watch out for gamma shifts on export and compensate if needed

At the end of the day I think proper cinema cameras are just better at achieving this look, but I sure it's possible to get closer with some effort.

 

 

[BTM_Pix]

Thick Negative sounds like two of the bullet points written on my school report.

[hyalinejim]

Actually from reading that link it turns out that, for good or ill, DXO doesn't take colour accuracy into account in their "portrait" rating.

They say all DSLRs score similarly for colour accuracy.

There's a lot on that page I don't understand. But I have the sense that their portrait rating has a lot more to do with quantitative than qualitative issues.

In short, numbers are compelling but in reality I'd be hesitant to use their rating in deciding which camera to buy to take photos of people in a studio.

[noone]

7 hours ago, hyalinejim said:

Actually from reading that link it turns out that, for good or ill, DXO doesn't take colour accuracy into account in their "portrait" rating.

They say all DSLRs score similarly for colour accuracy.

There's a lot on that page I don't understand. But I have the sense that their portrait rating has a lot more to do with quantitative than qualitative issues.

In short, numbers are compelling but in reality I'd be hesitant to use their rating in deciding which camera to buy to take photos of people in a studio.

I would NEVER use it for SUBJECTIVE choices (like colour choice).

DXO is a great GUIDE but I would never take it as gospel.     If choosing gear, it is just one of a half dozen sites I would look at (including this one when it comes to video).

I have always found Sony to be more ACCURATE colour than most (other than the AWB in some light issue) and when i have had Canon cameras and just shooting Jpegs or video, I have set them to vivid so hardly accurate.

[leslie]

VIVID... noone, you heretic 😀  Mr reid you now have the forums permission to sacrifice noone on the alter of c-log... 

call me an atheist but i'd hope for a thicc bone density xray, other than that i think its all bit naff. Maybe people are trying to carry something over from the days of film.  You remember film surely ? when every home home had one of these (ok maybe not every home 🤔).

My point being digital and film are fish swimming in different oceans as much as people like to emulate film with digital. I hold grave concerns when it come to technical terms of film being applied to digital. it all sounds a  little suss to me.  I'm normally a fan of the things you do here on the forums kye, but i cant help but wonder if your not overthinking it this time.

[hyalinejim]

Colour is a tricky one to talk about for sure. Nevertheless, there's been some interesting discussion in this thread.

Thickness of an image is a term that originated in reference to the density of a film negative. It's still used by some people when talking about digital images, where its meaning is unclear (qualitative rather than quantitative).

Perhaps a more useful avenue is to ask about characteristics of digital colour that are desirable. These could be

- Subjectively pleasing colour

- Objectively accurate to reality colour

- Objectively accurate to a film stock colour 

- The ability to push colour in a grade

On the topic of the first three above, it's interesting to note that Kodak designed their colour film stocks on the basis of customer feedback. They walked around shopping malls in the US and showed people photos with different colour renderings and asked them which they prefer.  More info here:

https://www.canopycanopycanopy.com/contents/color-goes-electric/?fbclid=IwAR0mWacShDadxrKexLSPLJ5tatHtZm4t7yBSUMEGQArQ4eSwSJCU8icX3WQ#six-one

 

[hyalinejim]

1980s Kodak test image from linked article 🙂

[leslie]

thats so VIVID   i need sunglasses to look at it 😉

[mat33]

While lighting, bit depth and resolution all play a role what about the bayer matrix and the colour filters themselves?  For example some of the 'thickest' most filmic images I have seen are from the Digital Bolex which is 12bit and 2K-I've attached a few examples (sadly not mine) from the D16.  I have seen some discussion on the D16 about this:   "Color depth and discrimination in Bayer filters is determined by the purity and saturation of the filter dyes used. The Kodak CCD in the D16 uses a patented set of highly saturated filters designed for scientific and industrial applications where accuracy and fine color discrimination are critical, like color matching in textile fabric dye runs, or in the cameras used on the Mars Rovers. Most video cameras use Bayer filters less saturated to get more sensitivity and just cover the video color gamuts. Deeply saturated primary colors can be hard to reproduce. I normally have to desaturate D16 footage 20-30% to eliminate chroma clipping on saturated colors in REC709. This still looks like a fully saturated video image to gamut limits." So could it be the characteristics of the bayer filters are a big part of the colour thickness?

[Andrew Reid]

In my opinion tonality trumps dynamic range.

You can have very high dynamic range but if the tonality and colour is lacking you get the 'thin' digital low-bit-depth look. A lot of smartphones have it in their HDR modes.

Some of the old CCD sensors produce lovely "thick" files but don't have anywhere near 12 stops dynamic range, not even 11 stops.

And yeah, Digital Bolex was a good example.

You can compare it to the claimed '15 stops' Sony S-LOG and it looks much deeper, more exotic, thicker, more filmic, more organic, despite having noisy shadows and clipped highlights.

[Andrew Reid]

Also...

For me a 'chunky' file is measured in the difference between, and the range, in just a single hue.

Mashed colours with poor tonality scream 'thin' digital file to me.

[mercer]

2 hours ago, mat33 said:

While lighting, bit depth and resolution all play a role what about the bayer matrix and the colour filters themselves?  For example some of the 'thickest' most filmic images I have seen are from the Digital Bolex which is 12bit and 2K-I've attached a few examples (sadly not mine) from the D16.  I have seen some discussion on the D16 about this:   "Color depth and discrimination in Bayer filters is determined by the purity and saturation of the filter dyes used. The Kodak CCD in the D16 uses a patented set of highly saturated filters designed for scientific and industrial applications where accuracy and fine color discrimination are critical, like color matching in textile fabric dye runs, or in the cameras used on the Mars Rovers. Most video cameras use Bayer filters less saturated to get more sensitivity and just cover the video color gamuts. Deeply saturated primary colors can be hard to reproduce. I normally have to desaturate D16 footage 20-30% to eliminate chroma clipping on saturated colors in REC709. This still looks like a fully saturated video image to gamut limits." So could it be the characteristics of the bayer filters are a big part of the colour thickness?

That top shot is gorgeous and a perfect example of a thick image.

I've dismissed upgrading from my 5D3 ML Raw due to the color depth it provides. The FP is the first camera that has me curious about an upgrade.

I'd love to see a comparison between the 5D3 ML Raw vs Sigma FP vs D16.

[kye]

On 10/7/2020 at 8:33 PM, hyalinejim said:

1980s Kodak test image from linked article 🙂

I'm not sure if it's the film stock of the 80s fashion that really dominates that image, but wow...   the United Colours of Beneton the 80s!

5 hours ago, leslie said:

thats so VIVID   i need sunglasses to look at it 😉

Hopefully ray bands - I think they're one of the only 80s approved sunglasses still available.

80s music always sounds best on stereos also made in the 80s.  I think there's something we can all take away from that 🙂 

4 hours ago, mat33 said:

While lighting, bit depth and resolution all play a role what about the bayer matrix and the colour filters themselves?  For example some of the 'thickest' most filmic images I have seen are from the Digital Bolex which is 12bit and 2K-I've attached a few examples (sadly not mine) from the D16.  I have seen some discussion on the D16 about this:   "Color depth and discrimination in Bayer filters is determined by the purity and saturation of the filter dyes used. The Kodak CCD in the D16 uses a patented set of highly saturated filters designed for scientific and industrial applications where accuracy and fine color discrimination are critical, like color matching in textile fabric dye runs, or in the cameras used on the Mars Rovers. Most video cameras use Bayer filters less saturated to get more sensitivity and just cover the video color gamuts. Deeply saturated primary colors can be hard to reproduce. I normally have to desaturate D16 footage 20-30% to eliminate chroma clipping on saturated colors in REC709. This still looks like a fully saturated video image to gamut limits." So could it be the characteristics of the bayer filters are a big part of the colour thickness?

Great images.

Fascinating comments about the D16 having more saturated colours on the bayer filter.  The spectral response of film vs digital is something that Glenn Kennel talks about a lot in Colour Mastering for Digital Cinema.  If the filters are more saturated then I would imagine that they're either further apart in frequency, or they are narrower, which would mean less cross-talk between the RGB channels.

This paper has some interesting comparisons of RGB responses, including 5218 and 5246 film stocks, BetterLight and Megavision digital backs, and Nikon D70 and Canon 20D cameras: http://www.color.org/documents/CaptureColorAnalysisGamuts_ppt.pdf

The digital sensors all have hugely more cross-talk between RGB channels than either of the film stocks, which is interesting.  I'll have to experiment with doing some A/B test images.  In the RGB mixer it's easy to apply a negative amount of the other channels to the output of each channel, which should in theory simulate a narrower filter.  I'll have to read more about this, but it might be time to start posting images here and seeing what people think.

3 hours ago, Andrew Reid said:

In my opinion tonality trumps dynamic range.

You can have very high dynamic range but if the tonality and colour is lacking you get the 'thin' digital low-bit-depth look. A lot of smartphones have it in their HDR modes.

Some of the old CCD sensors produce lovely "thick" files but don't have anywhere near 12 stops dynamic range, not even 11 stops.

And yeah, Digital Bolex was a good example.

You can compare it to the claimed '15 stops' Sony S-LOG and it looks much deeper, more exotic, thicker, more filmic, more organic, despite having noisy shadows and clipped highlights.

I wonder how much having a limited bit-depth is playing into what you're saying.  For example, if we get three cameras, one with high-DR and low-colour separation, the second with high-DR / higher colour separation, and the third with low-DR / high colour saturation.   We take each camera, film something, then take the 10-bit files from each and normalise them.  The first camera will require us to apply a lot of contrast (stretching the bits further apart) and also lots of saturation (stretching the bits apart), the second requires contrast but less saturation, and the third requires no adjustments.  This would mean the first would have the least effective bit-depth once in 709 space, the second would have more effective bit-depth, and the third the most in 709 space.

This is something I can easily test, as I wrote a DCTL to simulate bit depth issues, but it's also something that occurs in real footage, like when I did everything wrong and recorded this low-contrast scene (due to cloud cover) with C-Log in 8-bit..

Once you expand the DR and add saturation, you get something like this:

It's easy to see how the bits get stretched apart - this is the vectorscope of the 709 image:

A pretty good example of not having much variation in single hues due to low-contrast and low-bit-depth.  I just wish it was coming from a camera test and not from one of my real projects!

[kye]

Ok, what do we think about how thick these images are?

vs

vs

vs

[noone]

12 hours ago, kye said:

 

80s music always sounds best on stereos also made in the 80s.  I think there's something we can all take away from that 🙂 

Nah, 80s music sound best when played live by the original bands playing pension fund tours (though mostly with young backing musicians doing a lot of the actual playing).

[Oliver Daniel]

This is definitely a thing, and I feel a lot of this is actually to do with exposure and post colour management. 

The trouble is, a lot of those using cheaper mirrorless cameras shoot stuff in really poor lighting conditions, which obviously stretches the information captured into obscene brittleness.

It matters to pump up those pixels with a high quality of light - ensuring you have juicer, healthier pixels to play with. I’ve even seen some people “light” with their A7S ISO. Great feature for certain needs, but it’s not a light and the image will certainly feel thinner. 

Image processing will have an impact. RAW for example, is a higher data acquisition and expectedly looks “thicker” as a result - but again, poor lighting conditions will kill that data. 

The EVA1 ProRes RAW in 2k120 when “exposed correctly” looks absolutely horrendous, worse than cheap 8bit. A disaster. But 3 stops overexposed and reduced to normal in post, it looks incredible. And very “thick”. 

Also, I think “the emulsion” of your lens matters a great deal to how smooth the image renders on the sensor. For example, an electronic fly by wire 50mm Sony lens feels much more “electronic” then say, a Helios 44-2. It’s as though the Helios is smoothing out the digital characteristics. 

Then lastly, the appearance of very low rolling shutter and using the correct shutter speed for the frame rate and scene provides that optimum quality pleasing motion. 

I’m no imaging scientist, but at a more accessible level, I think we can use this understanding to create work with “thicker” images. It’s all down to your own practical choices, and harder to achieve on a low budget. 

[deezid]

20 hours ago, mat33 said:

While lighting, bit depth and resolution all play a role what about the bayer matrix and the colour filters themselves?  For example some of the 'thickest' most filmic images I have seen are from the Digital Bolex which is 12bit and 2K-I've attached a few examples (sadly not mine) from the D16.  I have seen some discussion on the D16 about this:   "Color depth and discrimination in Bayer filters is determined by the purity and saturation of the filter dyes used. The Kodak CCD in the D16 uses a patented set of highly saturated filters designed for scientific and industrial applications where accuracy and fine color discrimination are critical, like color matching in textile fabric dye runs, or in the cameras used on the Mars Rovers. Most video cameras use Bayer filters less saturated to get more sensitivity and just cover the video color gamuts. Deeply saturated primary colors can be hard to reproduce. I normally have to desaturate D16 footage 20-30% to eliminate chroma clipping on saturated colors in REC709. This still looks like a fully saturated video image to gamut limits." So could it be the characteristics of the bayer filters are a big part of the colour thickness?

 

These are thick looking images!
No noise reduction or other processing weirdness.
Super highly saturated shadow areas which many cameras nowadays desaturate to cover chroma noise
Overall huge tonal variation.
And what I guess - proper looking motion.

Something the A7sIII especially is bad with since it's optimized for low light - CFA prioritizing sensitivity over tonality + tons of temporal noise reduction and sharpening. All ending in a sterile and thin looking image.
Worst thing, doing ETTR to avoid smearing and ghosting at middle gray and below hues start shifting around. 😐

 

[deezid]

19 hours ago, Andrew Reid said:

Also...

For me a 'chunky' file is measured in the difference between, and the range, in just a single hue.

Mashed colours with poor tonality scream 'thin' digital file to me.

That's exactly the problem with most prosumer cameras and their processing nightmare inherited from phones these days.

Only a few cameras available below $6000 which can achieve a thick looking image - especially with internal recording (Pocket 6K Gen5, S1H/S1/S5 V-Log and R5 C-Log2 basically). Above that (below $10000) you can get a used ARRI Alexa, Varicam LT, or FX9, EVA1, Komodo, C200, Ursa Mini 12K and Mavo LF which all will deliver great images.

[noone]

3 hours ago, Oliver Daniel said:

 

It matters to pump up those pixels with a high quality of light - ensuring you have juicer, healthier pixels to play with. I’ve even seen some people “light” with their A7S ISO. Great feature for certain needs, but it’s not a light and the image will certainly feel thinner. 

 

While I think it is crazy to not light something when you CAN, a lot of the time shooting in available light is the goal or even all you can do and in that case having better DR and colour depth and tonality with a camera like an A7s beats a camera that starts off better but falls away a lot quicker.

I disagree that shooting in lower light levels makes an image "thinner" necessarily (though it CAN).