How is Portra "optimised for scanning"

Ian Grant · Dec 19, 2017

Kodak had issues with surface artefacts also known as micro reticulation, this became an issue when digital negative scanners were used for minilabs, it was the gelatin supercoat rather than the emulsion itself. It was causing interference patters due to scanner resolutions and appeared as excessive graininess, you could print OK optically. Kodak resolved this to a large extent with improved hardeners a number of years ago.

However in more recent years they've made even greater improvements making their films far better for scanning. In simple terms the Gelatin super coat is now much flatter, no longer has a matt look. Kodak make similar statements about improved optimisation for the newer versions of other films (Tmax etc).

Wet mounting for scanning or optical printing was one way to overcome these problems, a technique that goes back to the mid 1920's but of course it's not practical in commercial d&p labs/minilabs.

Ian

Chris Livsey · Dec 20, 2017

A discussion on the Vision3 film in another thread is related as Portra shares Vision3 technology and in a little digging I found this:

https://www.kodak.com/uploadedFiles/Motion/Products/Camera_Films/5203/Resources/5203_ti2657.pdf
The section on Post Production Scanning links, with a little more digging, to this one
https://www.yumpu.com/en/document/v...ommendations-for-extended-dynamic-range-kodak

Which rather changes the OP question about how Portra is optimised for scanning to how should we optimise scanning for Portra!!

Poisson Du Jour · Dec 20, 2017

Ian Grant said:
Wet mounting for scanning or optical printing was one way to overcome these problems, a technique that goes back to the mid 1920's but of course it's not practical in commercial d&p labs/minilabs.

There are definitely commercial labs that carry out wet-scanning. At least on this side of the world!
Some of my production work is produced through wet-scanning in a busy commercial lab (there are two other labs). There are a few of us doing this. True, it is a messy and smelly affair and it requires a block of booked time where no other work is done (that is not wet-scan related). Minilabs? Do they still exist??

Ted Baker · Dec 20, 2017

Chris Livsey said:
The section on Post Production Scanning links, with a little more digging, to this one
https://www.yumpu.com/en/document/v...ommendations-for-extended-dynamic-range-kodak

I guessed as much earlier on

Ted Baker said:
My guess is

A typical 16bit film scanner should cope with this improved range easily, where as the DPX scanner format is/was optimised for a particular generation of cinefilm, to make the best use of the available storage space. (10bits per channel)

Chris Livsey · Dec 20, 2017

But are they 16 bit really? Medium format backs were said to be 16 bit, which they were but it was 14 bits padded out, modern Sony cameras for example go down to 12 bit when anything "pushy" is required, just wondering

Ted Baker · Dec 20, 2017

Chris Livsey said:
But are they 16 bit really? Medium format backs were said to be 16 bit, which they were but it was 14 bits padded out, modern Sony cameras for example go down to 12 bit when anything "pushy" is required, just wondering

That's possible, in any case if they can give good results with a transparency then they should have enough room for these newer negative films

Lachlan Young · Dec 20, 2017

Chris Livsey said:
But are they 16 bit really? Medium format backs were said to be 16 bit, which they were but it was 14 bits padded out, modern Sony cameras for example go down to 12 bit when anything "pushy" is required, just wondering

I understand it can also have a lot to do with the nature of the analogue to digital converter - and if it's linear or logarithmic. For example, apparently the 12-bit log ADC's in Heidelberg drum scanners do a better job of making a 16-bit linear file than some linear ADC's of nominally higher bit depth.

Arbitrarium · Jan 3, 2018

I developed a roll of Portra 160 in the same tank as an Agfa Vista 200 yesterday. The Agfa dried curly as I was expecting but the Portra dried completely flat which really surprised me. So there's that.

Anon Ymous · Jan 3, 2018

Arbitrarium said:
I developed a roll of Portra 160 in the same tank as an Agfa Vista 200 yesterday. The Agfa dried curly as I was expecting but the Portra dried completely flat which really surprised me. So there's that.

All Kodak C41 films I've developed lately dry dead flat. Agfaphoto ones on the other hand...

trendland · Jan 4, 2018

tomfrh said:
Portra marketing blurb boasts that it's optimised for scanning. Optimised in what way? Is is the grain pattern? The colours?

I have real doubts IF in this market situation a manufacturer would improve a film emulsion to optimize scanning properties.
We may imagine this could be cause also disadvantages in concern of analogue workflow. ( the optimized scan characteristics)
So from my understanding you have different emulsions today - some are worst if you want to scan - others
BRING good scan properties.
And the Kodak Portra familie belongs to last category.
Am I right - or am I right ? ....

with regards

trendland · Jan 4, 2018

Lachlan Young said:
I understand it can also have a lot to do with the nature of the analogue to digital converter - and if it's linear or logarithmic. For example, apparently the 12-bit log ADC's in Heidelberg drum scanners do a better job of making a 16-bit linear file than some linear ADC's of nominally higher bit depth.[/QUOTtE]

You are right in concern of wet scanning with Heidelberg machines.
But may be you forget the better resolution (total different scan technique gives much more quality) in concern of much much cheaper todays scan equipment ?
I am not very sure about if the color deep in different (bit) scales is allmost
responsible for better resulting scan quality.
In case of Heidelberg I have a feeling :
It is just the drum...

with regards

Lachlan Young · Jan 4, 2018

trendland said:
You are right in concern of wet scanning with Heidelberg machines.
But may be you forget the better resolution (total different scan technique gives much more quality) in concern of much much cheaper todays scan equipment ?
I am not very sure about if the color deep in different (bit) scales is allmost
responsible for better resulting scan quality.
In case of Heidelberg I have a feeling :
It is just the drum...

My comments were intended to point out one of the more notable differences between the Tango etc & other high end PMT or CCD scanners with a fluid mount capability. There are scanners which will outresolve a Heidelberg, but the Heidelbergs seem to have an edge in terms of what they can usefully dig out of a negative or transparency. Pretty much any high end scanning kit runs rings around your average Epson etc, and for most negative scanning purposes, pretty much all high end scanners are very good indeed - it's really a question of how much resolution you need that's the deciding factor.

trendland · Jan 4, 2018

Lachlan Young said:
My comments were intended to point out one of the more notable differences between the Tango etc & other high end PMT or CCD scanners with a fluid mount capability. There are scanners which will outresolve a Heidelberg, but the Heidelbergs seem to have an edge in terms of what they can usefully dig out of a negative or transparency. Pretty much any high end scanning kit runs rings around your average Epson etc, and for most negative scanning purposes, pretty much all high end scanners are very good indeed - it's really a question of how much resolution you need that's the deciding factor.

I see - that should matter also.
Yes resolution - thats the case of interest.The great point.
Short answer : max. resolution.
By the time - Kodak obviously had also propper scan technique in the 90th.
The machines the used for their Kodak Photo CD.
The proffesional Photo CD had 4000 x 6000 resolution with 135-36 films (1997!!!!!) this was the killer aplication to normal pc.

with regards

Ted Baker · Jan 4, 2018

trendland said:
Am I right - or am I right ?

No, the point being, that all negative film has a latitude that exceeds the capabilities of the print. In black and white you always have the easy option of dodge and burn etc. You could of course do the same with color printing but it's not as straightforward, where as scanning it is easier to make use of the extra latitude.

Bill Burk · Jan 4, 2018

Ian Grant said:
Kodak had issues with surface artefacts also known as micro reticulation, this became an issue when digital negative scanners were used for minilabs, it was the gelatin supercoat rather than the emulsion itself. It was causing interference patters due to scanner resolutions and appeared as excessive graininess...

This makes sense.

Also, to state the obvious... it's not Kodachrome (so it doesn't mess with Digital ICE like Kodachrome does).

Sirius Glass · Jan 4, 2018

trendland said:
I have real doubts IF in this market situation a manufacturer would improve a film emulsion to optimize scanning properties.
We may imagine this could be cause also disadvantages in concern of analogue workflow. ( the optimized scan characteristics)
So from my understanding you have different emulsions today - some are worst if you want to scan - others
BRING good scan properties.
And the Kodak Portra familie belongs to last category.
Am I right - or am I right ? ....

with regards

You can put you doubts aside. Kodak made a concerted effort to improve Portra scanability.

DREW WILEY · Jan 4, 2018

If it's on the film and can be recovered using a scan, it can be recovered and printed fully analog. Making and using supplementary film masks might seem like a nuisance, but so is removing scanning fluid. And either way, you've got to clean dust off first. But the real question is what will be the actual quality of those far reaches of film "latitude"?

NJH · Jan 5, 2018

Bit range is pretty much irrelevant to most ADC systems, the limiting factor will nearly always be your noise floor. Lots of mythology around ADC, sadly including from those in the various industries either using them or making them.

Ted Baker · Jan 5, 2018

NJH said:
Bit range is pretty much irrelevant to most ADC systems, the limiting factor will nearly always be your noise floor. Lots of mythology around ADC, sadly including from those in the various industries either using them or making them.

I'm sorry that's just nonsense, bit range is very relevant particularly in scanning, as the film never represents a linear measurement of the light in a scene. Consequently you need quite high precision in one part of the range and less in others.

As a simple example 18% of the range is roughly half of the range that our eye perceives in terms of brightness.

trendland · Jan 5, 2018

Sirius Glass said:
You can put you doubts aside. Kodak made a concerted effort to improve Portra scanability.

Well - I insist on my opinion.
It is quite clear from history also. Scanning issues wasn't issues before 1995. It was a real niche and I am sure Kodak never mind about (exeptions would have made sence for proffessional films during 1996 - 2000.)
With 2000 the big boom from scanning began (to the same time we noticed digital cameras with pore characteristics).
And after this time improvements made sence (caused from massive scanning trend) we may fixe the year 2005 BUT to this time Kodak was still in the midle of its big crisis.
From were should came the money to new emulsion design for better scan characteristics.
Remember also Kodak cut the money for film research in 2004 if I remember correct. The later improvements (after 2004) came (without exeption) from research made in the years before 2003.
And within this period scanning of film was no issue to a manufacturer producing billions of films - (most for the target group of amateuric photograpers who need films only one time a year : during holidays).
But Ian gave us a hint - so look above :
Kodak reformulate the gelantine supercoat via improved hardeners.
That should be done within the years I reffered (2000 - 2005).
Sirius Glass - so pls. let us speak about a optimization in regard of scanning issues from Kodak.
And not about a reformulation/improvement of the emulsion concerning the Portra family.
(and other films )

That's all I intended to state - because I can't imagine a manufacturer these days will spend money (with extreme costs) and this should stand also to the last 10years - because a small group of costumers had problems from scanning.
And sure I remember Kodak spoke about improvement of Portra to have better scanning characteristics!
But as I mentioned so many times before
you should not believe everything to 100% companies tell you.
Make your own thoughts about - then it often becomes much more clear.
So the statement of Kodak is better to understand as M A R K E T I N G ....
And we should know : They optimized their films (via simple method with the use of improved hardeners)
with regards

PS : Thank you Ian Grand - by the time - for you competent factual information.

trendland · Jan 5, 2018

Ted Baker said:
No,the point being, that all negative film has a latitude that exceeds the capabilities of the print. In black and white you always have the easy option of dodge and burn etc. You could of course do the same with color printing but it's not as straightforward, where as scanning it is easier to make use of the extra latitude.

Well - Ted Baker - that should be a point,
you mentioned, I have to think about first. so thank you therefore.

with regards

PS : .... Glad about to realize I am not allways right .... (hope you recognize the irony )

NJH · Jan 5, 2018

Ted Baker said:
I'm sorry that's just nonsense, bit range is very relevant particularly in scanning, as the film never represents a linear measurement of the light in a scene. Consequently you need quite high precision in one part of the range and less in others.

People were talking about 16 bits, can you tell me what the actual sensor device ADC bit depth of these devices is, their noise floor and actual SNR at the read out (in other words the real number of useful sensor bits)? This is fundamentally not the same as the number of bits output from a system, or for that matter what is useful for doing software based signal manipulation.

Ted Baker · Jan 5, 2018

NJH said:
People were talking about 16 bits, can you tell me what the actual sensor device ADC bit depth of these devices is, their noise floor and actual SNR at the read out (in other words the real number of useful sensor bits)?

On the box it says it 16, but I would not be surprised that it is just marketing and it's 14, which would be enough. But 12 bits for a linear AD would be the absolute minimum, maybe 10bit if the negative was perfect and no tonal corrections are needed. In this context a bit of noise would actually be helpful.

It is very straight forward to test this, create a scan and save with a gamma of 1 (real gamma of 1) and 8bit, this will basically turn your scanner into a scanner with an 8bit A/D then try and make your corrections/post processing.

Sirius Glass · Jan 5, 2018

Ted Baker said:
On the box it says it 16, but I would not be surprised that it is just marketing and it's 14, which would be enough. But 12 bits for a linear AD would be the absolute minimum, maybe 10bit if the negative was perfect and no tonal corrections are needed. In this context a bit of noise would actually be helpful.

It is very straight forward to test this, create a scan and save with a gamma of 1 (real gamma of 1) and 8bit, this will basically turn your scanner into a scanner with an 8bit A/D then try and make your corrections/post processing.

The last time I looked a few years ago, there were few true 16 bit per color cameras, many 14 bit and 12 bit per color cameras because there were problems building 16 bit analog to digital converts that would work in digital cameras.

Les Sarile · Jan 7, 2018

Bill Burk said:
Also, to state the obvious... it's not Kodachrome (so it doesn't mess with Digital ICE like Kodachrome does).

Nikon Coolscan 9000 + Nikonscan ICE has the best implementation of ICE that I have tried especially when used with Kodachrome. Below is a perfect example of how it works without introducing any strange artifacting found in others particularly Canon's FARE.

How is Portra "optimised for scanning"

Privacy & Transparency

Privacy & Transparency