Grain on large format (scanned negative): hp5 plus vs tmax 100

[rpagliari]

I haven't had the opportunity of experimenting with 4x5 HP5+ and Tmax100, as I haven't received my camera yet.

So I'd like to ask if anybody has experience with these two films. The reason is that I like a lot the rendering of HP5+, it's my favorite film with 35mm and 120.

However, tmax is known for being much cleaner when it comes to film grain (at least on a scanned negative), but I've seen comments about the fact that in large format there isn't much gain in using iso 100 vs 400.

As a side note, I'm planning to scan my negatives with a commercial Epson scanner. I have a 550 but might upgrade to the 800 series.

[rdeloe]

There are a bunch of recent threads on scanning large format, including one I started on HP5+ and TMY (TMAX 400). Long story short, they're both great but different. I concluded TMY was the film for me. But if I only had access to HP5+ I'd be happy too.

[Drew Wiley]

Depends somewhat on the specific developer. But in the same nominal speed category, TMY400 is MUCH finer grained than HP5, and TMX100 obviously finer still. I think you'll find that optimally strategizing sheet sizes of HP5 is a somewhat different game than using it in small roll sizes due to the significantly smaller enlargement factor. But this depends on what you use your scans for. TMax films handle high contrast subjects much better, but also need more careful metering.

[Ken Lee]

Given the same developer and film much depends on your subject and whether it has large sections of smooth tone which reveal grain or course tones which obscure grain.


http://www.kennethleegallery.com/images/forum/Road4.jpg

I have made a 40x50 inch print from this 4x5 HP5+ woodland scene, developed normally in D-23 1:1. Because it's comprised mostly of leaves (so-called high-frequency data), it sharpens up nicely and looks quite grainless even at large size.


http://www.kennethleegallery.com/images/forum/Landscape31.jpg

On the other hand, a photo like this (also 4x5 HP5+) is more likely to reveal any grain defects and requires more careful sharpening: avoiding the high values and sharpening only the mid-tones.

To sharpen only specific ranges of the tonal scale, you might find this brief article helpful: Avoid Sharpening Artifacts: Sharpen the Mid-Tones Only (http://www.kennethleegallery.com/html/tech/index.php#Sharpening)

[pepeguitarra]

I have used them both and I cannot see grain in my photos. I doubt I will see grain in a 4x5.

[Pere Casals]

The reason is that I like a lot the rendering of HP5+, it's my favorite film with 35mm and 120.


The HP5+ / TMX grain it's the same in 35mm than in 4x5, simply you enlarge less for the same print size. So your knowledge from rolls is to tell you how grain will be in prints from LF. See how grain shows at x4 and you will know how grain will be in a 16x20" print from a 4x5 negative.



I have a 550 but might upgrade to the 800 series.

of course

[Jim Noel]

Ken,
On my screen these appear to be scans of digital prints,particularly the first one. Is this true?

[rpagliari]

Given the same developer and film much depends on your subject and whether it has large sections of smooth tone which reveal grain or course tones which obscure grain.


http://www.kennethleegallery.com/images/forum/Road4.jpg

I have made a 40x50 inch print from this 4x5 HP5+ woodland scene, developed normally in D-23 1:1. Because it's comprised mostly of leaves (so-called high-frequency data), it sharpens up nicely and looks quite grainless even at large size.


http://www.kennethleegallery.com/images/forum/Landscape31.jpg

On the other hand, a photo like this (also 4x5 HP5+) is more likely to reveal any grain defects and requires more careful sharpening: avoiding the high values and sharpening only the mid-tones.

To sharpen only specific ranges of the tonal scale, you might find this brief article helpful: Avoid Sharpening Artifacts: Sharpen the Mid-Tones Only (http://www.kennethleegallery.com/html/tech/index.php#Sharpening)

Thanks a lot for posting examples. I do mostly architectural/minimalistic photos. The second photo looks absolutely fine with me, so I might try the HP5+ as well; I actually prefer to see some grain in the image, so long as the grain does not dominate the image.

[Ken Lee]

Thanks a lot for posting examples. I do mostly architectural/minimalistic photos. The second photo looks absolutely fine with me, so I might try the HP5+ as well; I actually prefer to see some grain in the image, so long as the grain does not dominate the image.

Theses photographs are displayed at only 1000x800 pixels: a mere fraction of the size of a real print.

Unless we are planning to shoot LF merely for display online, it may not be helpful to judge these factors according to small image files. If we're shooting for the web alone, we might be well served just shooting with smart phones. For online sharing, even many smart phones are overkill.

[rdeloe]

There's also the question of whether or not not you are actually scanning grain. Unless you're using a drum scanner, probably not. Others have explained this well when it came up previously, e.g., https://www.largeformatphotography.info/forum/showthread.php?31574-xtol-dilutions&p=301556&viewfull=1#post301556

In my personal experience, which involves "scanning" with a digital camera at 2,666 ppi, artefacts produced by the digital sensor are what is visible, rather than grain.

[hsandler]

I've scanned a fair bit of 4x5 on an Epson V750 at optimal height at 2400ppi, but I don't have any examples posted online at full resolution to show you. I like HP5 at box speed or FP4 at box speed for even finer grain. Honestly, you would have to print pretty big to see grain in a large format image. As others noted, you are not really scanning grain on a flatbed scanner; it can't resolve grain. You are just seeing aliasing of the grain into larger apparent clumps the size of a scanned pixel. So I shoot these two films largely because they are the lowest cost Ilford films. For medium and smaller format I use Ilford Delta for finer grain.

[Pere Casals]

As others noted, you are not really scanning grain on a flatbed scanner; it can't resolve grain.

You scan grain with a flatbed !!! it resolves true film grain. Of course it resolves better HP5 grain than TMX or Delta 100.

It is true that a drum may deliver sharper grains, and small grains are very aliased with a cheap flatbed.

An epson resolves some 0.01mm, and many grains are way larger. Another thing is bromide crystals in the emulsion, those are seen with electronic microscopes

[mmerig]

You scan grain with a flatbed !!! it resolves true film grain. Of course it resolves better HP5 grain than TMX or Delta 100.

It is true that a drum may deliver sharper grains, and small grains are very aliased with a cheap flatbed.

An epson resolves some 0.01mm, and many grains are way larger. Another thing is bromide crystals in the emulsion, those are seen with electronic microscopes

Kodak defines grain as "A particle of metallic silver or a cloud of dye in a photographic emulsion. Exposed silver halide crystals in raw emulsion that become grains in the photographic process." See Print Grain Index, E-58. C. E. Kenneth Mees, in his book "From dry plates to Ektachrome", discusses grain sizes and what it takes to see them. According to Mees, a grain particle itself is about 1 to 5 microns in diameter (the silver halide crystals).

What we see as grain in a scan is noise and grain clumps. If individual grain particles were "way larger than 0.01 mm", the resolving power of film would be poor.

[Pere Casals]

Kodak defines grain as "A particle of metallic silver or a cloud of dye in a photographic emulsion. Exposed silver halide crystals in raw emulsion that become grains in the photographic process." See Print Grain Index, E-58. C. E. Kenneth Mees, in his book "From dry plates to Ektachrome", discusses grain sizes and what it takes to see them. According to Mees, a grain particle itself is about 1 to 5 microns in diameter (the silver halide crystals).

What we see as grain in a scan is noise and grain clumps. If individual grain particles were "way larger than 0.01 mm", the resolving power of film would be poor.

Please see this document, Page 2 http://www.tmax100.com/photo/pdf/film.pdf This is nice info, a doc worth to keep.

It shows pictorial film grain is 10 to 30 microns this is 0.01mm to 0.03mm. (In microfilm type film, adox cms 20, it is way smaller)

My view is that those grains in the 0.03mm range are well depicted by a cheap (V800, at top performance settings, 0.01mm) flatbed, those grains in the low end (0.01) are mostly blurred by a the cheap flatbed but well depicted by a high end flatbed or drum scanner. A very cheap flatbed (V600) it even blurs the big grains.






______________________________________________________


Aliasing/blurring the big grains it also may happen during Ps edition with resize, rotate, etc. And of course it will happen when digitally printing.

Dealing with grain in the hybrid workflow (small formats) it's not easy if structure counts.

Also note that those scanners that are limited in resolving power by the dpi in the sensor like Nikon LS 9000 or Hassy X5 may deliver aliasing when working at top prformance, these are scanners sporting a very good lens but using a 8000pix sensor. A flatbed like a V850 will never deliver grain aliasing because resolving power limitation come from the optics, the sensor (40,000 pix) outresolves the lens by an x2 factor, the V850 covers always 5.9" so the lens performance is always stressed. The EPSONs may blur the grains, but will never have aliasing.

This effect can be compared with the lowpass optical filter many DSLRs had to prevent aliasing. Today the smaller pixel size makes the LPF less useful (and not placed) because the sensors outresolve the lenses is most of practical situations.


188797


___________________________________

Regarding film resolving power it all depends on contrast, at 1:1000 is when a BW film may go well beyond 100lp/mm, in that situation what is working is very, very small grains that are only exposed at some +3 overexposure, and that are formulated in the emulsion for the curve shoulder shape.

The doc says:
Silver-halide crystals are 0.2-2.0 um
Color Clouds are 10-30 um
BW T grains are 15-25 um

Time ago I made a personal/informal test. I made a tinny cross with a marker in a negative, I observed the fladbed and drums scans, and then I also observed the cross in a microscope at x400. I was very interested for knowing how grain was digitally printed and what happened with scanners. I could identify the same particular grains in the cross and in the scan, so I could compare. It is an excercise I'd recommend to those interesed in grain structure, anyway this is a least concern in LF, but in rolls it can a powerful aesthetic tool for some.

[rpagliari]

I dont have enough technical background for this. However, I do see different grain when scanning different type of negative, and it always increases as asa goes up.

How could that be just digital noise?

[Pere Casals]

I dont have enough technical background for this. However, I do see different grain when scanning different type of negative, and it always increases as asa goes up.

How could that be just digital noise?

You only have digital noise in the crazy high high densities, this would be in the extreme Velvia shadows, this is electronic noise.

Your V550 delivers 1500dpi effective, if you get the V800 you will have 2800dpi effective in the transverse axis and some 2300dpi in the vertical axis, so some 2500dpi effective average, this is a big leap forward concerning grain.

IMHO the V800 is very capable for LF and even MF, for 35mm film you may want an additional 35mm cheap Plustek.

I most practical situations the V800 nearly matches camera/film performance for LF, and for LF grain is usually not a concern.

For roll film you may want to exploit grain structure. To me TMax or Delta flat grains are not interesting for the aesthetics. IMHO those were products intended to minimize grain, not intended to deliver a nice structure.

HP5, FP4, TX, and TXP are a different war. Those films evolved during many decades and are still there because they have an strong aesthetic culture backing them.

After releasing TMax kodak planned to discontinue TX. A riot was assembled, kodak managers didn't know why photographers were buying TX...

TX/TXP delivers more grain in the shadows than in the mids (if scene microcontrast not masking grain), delivering a dramatic mood. HP5 has its peak grain in the mid grays, delivering a particular atmosphere.

Format modulates grain. 35mm shows a lot, 645 vs 6x7cm is not the same. And LF hides most grain.

So, to me, for grain wait until you have the 800 and see. IMHO you'll be happy with MF and LF scans. If wanting a better depiction for 35mm than you'll need an additional cheap Plustek 35mm, if not wanting to spend a lot in a Plustek 120.

Note that V850 bundles a silverfast version that is Multi-Exposure capable, which is good for Velvia extreme shadows, with V800 you can purchase that software feature in the future if you want. This is an important difference I see V800-850. V850 have the lenses coated, but I guess that this is not much noticed.

[rpagliari]

You only have digital noise in the crazy high high densities, this would be in the extreme Velvia shadows, this is electronic noise.

Your V550 delivers 1500dpi effective, if you get the V800 you will have 2800dpi effective in the transverse axis and some 2300dpi in the vertical axis, so some 2500dpi effective average, this is a big leap forward concerning grain.

IMHO the V800 is very capable for LF and even MF, for 35mm film you may want an additional 35mm cheap Plustek.

I most practical situations the V800 nearly matches camera/film performance for LF, and for LF grain is usually not a concern.

For roll film you may want to exploit grain structure. To me TMax or Delta flat grains are not interesting for the aesthetics. IMHO those were products intended to minimize grain, not intended to deliver a nice structure.

HP5, FP4, TX, and TXP are a different war. Those films evolved during many decades and are still there because they have an strong aesthetic culture backing them.

After releasing TMax kodak planned to discontinue TX. A riot was assembled, kodak managers didn't know why photographers were buying TX...

TX/TXP delivers more grain in the shadows than in the mids (if scene microcontrast not masking grain), delivering a dramatic mood. HP5 has its peak grain in the mid grays, delivering a particular atmosphere.

Format modulates grain. 35mm shows a lot, 645 vs 6x7cm is not the same. And LF hides most grain.

So, to me, for grain wait until you have the 800 and see. IMHO you'll be happy with MF and LF scans. If wanting a better depiction for 35mm than you'll need an additional cheap Plustek 35mm, if not wanting to spend a lot in a Plustek 120.

Note that V850 bundles a silverfast version that is Multi-Exposure capable, which is good for Velvia extreme shadows, with V800 you can purchase that software feature in the future if you want. This is an important difference I see V800-850. V850 have the lenses coated, but I guess that this is not much noticed.

Thanks a lot for the info. I'm using a Plustek with 35mm. So, in the end, I'd better upgrade my epson, as I do MF and LF as well..

Thank you,

[mmerig]

Please see this document, Page 2 http://www.tmax100.com/photo/pdf/film.pdf This is nice info, a doc worth to keep.

It shows pictorial film grain is 10 to 30 microns this is 0.01mm to 0.03mm. (In microfilm type film, adox cms 20, it is way smaller)

My view is that those grains in the 0.03mm range are well depicted by a cheap (V800, at top performance settings, 0.01mm) flatbed, those grains in the low end (0.01) are mostly blurred by a the cheap flatbed but well depicted by a high end flatbed or drum scanner. A very cheap flatbed (V600) it even blurs the big grains.



The Tim Vitale article you linked to says the following:

"The term “film grain” is often incorrectly used to describe the “fundamental” particles in a chemical-based photographic image."

and

"Grain is a regular repeating noise pattern larger than the fundamental particles. Some mistake film grain for the image-forming elements in film.

and

"Many Kodak and Fuji “popular” publications, including much of the popular photographic literature (magazines), also make the mistake of referring to fundamental film particles as filmgrain. This further propagates the imprecise usage of the term."

C. E. Kenneth Mees, who worked for Wratten and Wainwright, then Kodak, was involved with decades of research in photographic materials. His definitions differ from Vitale's, and I would hardly call Mee's books "popular" in the sense that Vitale uses. In essence, Mees (in his book From Dry Plates to Ektachrome Film) defines grain as the fundamental particle, or silver halide crystal (in the 1 micron range). The 3-dimensional clumpiness aspect, or structure of the grain particles, he calls granularity, and its appearance as graininess.

This is somewhat semantic, but intuitively, a grain of sand is analogous to a silver halide crystal in emulsion, and it makes more sense to call the individual particles grain.

[Pere Casals]

This is somewhat semantic, but intuitively, a grain of sand is analogous to a silver halide crystal in emulsion, and it makes more sense to call the individual particles grain.

We have silver-halide crystals before development and metallic silver clumps after, we may call grain to both, but they are different things.

My view is that photographers and industry always call film grain to the metallic silver particles we have after development:

188842

"Fine grain" or "coarse grain" always refer to that... Me, I use "silver halide crystals" (or simply crystals) to refrer the "fundamental particles" in the emulsion (a suspension in fact) before development, but when I learn DIY emulsion recipes for my dry plates I find some "grain" word usage with the crystals.

It has to be noted that good technical literature always use crystal for silver halide particles and never grain, for example you will find epitaxial or tabular crystal growth depending on the pAg factor (silver concentration in the addition), a technician-scientist will never say grain for that, IMHO.

It has to be noted that with processing the "fundamental particles" (halide crystals) disapear. The shape of the clump depends on solvent effect, infectious development etc, but the individual crystal shapes are not there anymore, in fact the crystals disolve during development or fixing.

My view is that we should say crystals for silver-halide crystals, as we always are in a technical context when mentioning that then better if we say the real concept, and it's better to leave the grain word for the metallic silver that is mentioned everywhere in general film photography for developed film.


Anyway when scanning we cannot go wrong with semantics, crystals are destroyed in the processing. ...so for sure we speak about the metallic clumps.

[interneg]

[Mees] calls granularity, and its appearance as graininess.


And it is the rendering of this characteristic in low contrast areas (smooth grey sky tones for example) that is a tough test for an optical system's MTF performance. Epsons & most consumer flatbeds fail at this. At 2000ppi & with good MTF performance (aka, high end CCD of various sorts, PMT drum scanners, camera scanning too), you will be able to render this detail proportional to overall resolution. The weird mush from consumer flatbeds that becomes a horrid noisy mess when sharpened is not a reasonable rendering of a film's granularity/ grain character.

[Pere Casals]

The weird mush from consumer flatbeds that becomes a horrid noisy mess when sharpened is not a reasonable rendering of a film's granularity/ grain character.

This an exageration.

Just download samples here:

https://web.archive.org/web/20181002142200/https://petapixel.com/2017/05/01/16000-photo-scanner-vs-500-scanner/
https://petapixel.com/2017/05/01/16000-photo-scanner-vs-500-scanner/

and edit both cases for a best match.

You will find how close result is, and how crazy big you have to enlarge to notice some difference.

The X5 shows more aliasing than the V850 but not better grain:

188845
https://web.archive.org/web/20181002142200im_/https://petapixel.com/assets/uploads/2017/05/sidebyside-800x549.jpg


https://petapixel.com/assets/uploads/2017/05/x1_1250-796x800.jpg


In 35mm film the X5 will show far better grain than a V850, but MF and up the thing is different. In 4x5 the V850 shows better grain. This is because the X5 zooms in, but it only has 8000 pixels in the sensor, while the V has 40000 pixels.

[interneg]

Pere - if you had any real extent of experience with high end scanning & printing rather than spewing links to clickbait, you would know that the claims in that article are at best disingenuous & at worst intentionally dishonest. The real-world differences are not small & become pretty blatantly clear at almost any size of print, especially with current generation inkjet print heads. You have had this explained to you innumerable times by many, many people with years of experience with a variety of high-end scanning equipment, & yet you delude yourself that Epsons are just one unsharp mask from equalling any high end scanner. They are not - and having spent my Thursday afternoon trying to extract decent results from Epson originated files because the original negatives were not available for re-scanning, my charitability towards such claims is currently very limited. For the record, they were from 120 negatives & not big prints - in fact they were well within the resolution you claim the Epson delivers. Large format does not do much better. The totality of the optical system matters, not the nominal high contrast resolution.

Almost any current 24x36 DSLR, a decent macro lens & a light table will do a better job than a consumer flatbed - and with a Nikon D8xx series or similar resolution sensor, up to a 16x20 print you might even avoid having to stitch files.

[Pere Casals]

Almost any current 24x36 DSLR, a decent macro lens & a light table will do a better job than a consumer flatbed - and with a Nikon D8xx series or similar resolution sensor, up to a 16x20 print you might even avoid having to stitch files.

Interneg, in fact a DSLR setup may even outresolve best drum scanner you may find, you only need a high manification macro lens and stitching a high image count.

Beyond that, this is a LF forum, an speaking about sheets you should test what a Hasselblad X1 or X5 is able to do. If one day you check that you will find 1800 effective dpi for 4x5".

Also if you check that with a V850, like I did, you will find that it resolves 2900 dpi in the transverse axis and 2300 in the vertical axis. Of course the X5 is very good for 35mm, but let me reiterate that this is a LF forum.

In practice the V850 performs amazingly for LF. Recently Pali posted some side by side samples, by editing the V700 samples to their best it happened that result was pretty comparable to other high end flatbeds:

https://www.largeformatphotography.info/forum/showthread.php?150020-Scanner-Comparison-2019-Epson-Flatbed-Eversmart-Flatbed-Drum-Scanners&p=1479178&viewfull=1#post1479178

the edited crop in the V700 sample show the same than the crops from very expensive flatbeds:

https://c1.staticflickr.com/5/4857/46755757932_c7010da815_o.jpg

You may try it on your own, just download those samples and sharpen the V700 one with 2.5 radius and 50% setting, not that difficult, I guess.

Something is true, an EPSON flatbed is not a Pro machine, the scan is not digitally optimized like Pro machines do, it always require some Ps edition to reach its best.

EPSON V700/800 flatbeds fall a bit short in 35mm film jobs, but for LF it's an excellent choice, that given the price (new) it's not easy to beat.

[interneg]

Pere - you clearly have no real clue how the optical/ mechanical system in the Flextight works & that the difference between the early Flextight systems & the newer ones is going to be clearly visible on a test target. If you did, you would understand why the optical system is able to effectively deliver lossless resolution to the sensor on the 4x5 setting. There's 3x linear CCD's, not one behind a Bayer array. Furthermore, by your logic, the Precision II should only deliver 1440ppi on 4x5, not the 1800ppi it actually does.
Nor have you made many hundreds if not thousands of scans of LF negatives on both machines & made big prints. You can tell the difference from 3 metres away, even on a 16x20" & closer on smaller prints - the Epson is a pile of crap on anything other than very high contrast edges & fails utterly at handling fine details in low contrast areas, producing only noise when sharpened. I've made 2400 & 3000ppi stitched scans of 4x5 & 9x12cm & the difference is even more ridiculous. Your sharpening test is fundamentally flawed because it does not look at noise generation in smooth tones, but is instead used on high contrast edges. In other words, you are skewing the test with the intention to deceive.

[Pere Casals]

Pere - you clearly have no real clue how the optical/ mechanical system in the Flextight works

Interneg, rather than going to personal attack (you have no clue etc...) you may comment on the samples in post 23.

Later you can go to the collaborative scanner test to download samples, and then explain why the V750 crop with some edition nearly matches the howtek drum result:

https://www.largeformatphotography.info/scan-comparison/

https://farm9.staticflickr.com/8561/28420386682_d481942db8_o.jpg

https://www.largeformatphotography.info/scan-comparison/scanimages/prem-4x5-fullframe-u.jpg

[Ted Baker]

It might be worth remembering that Kodak measure grain using an aperture of 48microns! I have heard that was the size of the hole made into a piece of metal using a sharp point they had around the workshop. 48microns is about 525dpi. So you don't need a very high dpi sensor to "see" or measure grain... That is of course complete different to being able to distinguish particles of around 3-4 microns...

Basically any sensor in 500dpi range should be able detect or show grain such that it can be measured. How it actually looks when scanned and then printed is another thing entirely.

[Pere Casals]

It might be worth remembering that Kodak measure grain using an aperture of 48microns!

Yes... but scanning aperture of the microdensitometer is larger than the grains to see the variable reading as the scanning position moves. Grains may be sized in the 10 to 30 microns, so individual (10um) grains will jut fit inside a 2500 dpi pixel sampling area. 2500dpi is suitable to shape large (30um) grains, while you need more resolving power to shape well the smaller (10um)grains, because at 2500dpi the 10um grain has same size than the pixel.

[interneg]

Later you can go to the collaborative scanner test to download samples, and then explain why the V750 crop with some edition nearly matches the howtek drum result:


That's clearly the softest drum scanner by some margin - possibly in need of servicing etc. Let's see how your claims measure up to one of the the Heidelbergs or the 949 then? They're both in that comparison & how about using the other image crops rather than the high contrast edge one? Those are far more telling.

[MAubrey]

...spewing...clickbait...disingenuous...intentionally dishonest...you delude yourself...my charitability towards such claims is currently very limited
Well, that last bit is certainly clear!

[Pere Casals]

That's clearly the softest drum scanner by some margin - possibly in need of servicing etc. Let's see how your claims measure up to one of the the Heidelbergs or the 949 then? They're both in that comparison & how about using the other image crops rather than the high contrast edge one? Those are far more telling.

Interneg, we may do pixel peeping with the crops as much as you want...

The reality is that your opinion "the Epson is a pile of crap on anything other than very high contrast edges & fails utterly at handling fine details in low contrast areas" is pretty wrong.

The post #23 is not "very high contrast edges", and there the cheap epson is rivaling with high end scanners that costed $ many thousands.

Me, I don't understand why some hate the epson in that way, often throwing the "crap" word when speaking about it... sometimes it is because of commercial interests...

But the sample in the #23 post (for example) speaks on their own.

The V850 has some limitations, as any machine has, but what's for MF and up it performs amazingly, in special a 8x10" V850 scan it's a 100% Pro thing. The larger is the format the better the epson performs, the drawback is that for 35mm it's not the best choice, because it scans 4 strips wide at the same time, so result is not optimal.

Think that if the V850 had a 3rd lens covering a single 35mm strip (24mm scan width) it would deliver an insane resolving power, just multiply. Other manufacturers are lucky that this was not considered by epson.

[Bruce Watson]

I haven't had the opportunity of experimenting with 4x5 HP5+ and Tmax100, as I haven't received my camera yet.

You don't have enough experience to make a decision. You'll need to put some film through your camera and your process before you can possibly know enough to make a decision.

That said, you've got to start somewhere. So pick one and start. You can always change later.

It's not like your first 5x4 film is going to make an award-winning 50x40 print. Even if most of us think that's going to be the case. I know I did. Every shot. :o


So I'd like to ask if anybody has experience with these two films. The reason is that I like a lot the rendering of HP5+, it's my favorite film with 35mm and 120.

Yeah, well, small format isn't large format. This is sorta like trying to compare trumpet with clarinet. Surprisingly little of your years of clarinet will apply to trumpet. Your ability to read music, some ability to play with others and understand a director. But you'll find that LF doesn't play like SF. Just doesn't. Which is again why you need to burn a lot of film learning your new LF camera and your process.


However, tmax is known for being much cleaner when it comes to film grain (at least on a scanned negative), but I've seen comments about the fact that in large format there isn't much gain in using iso 100 vs 400.

LF typically involves less enlargement than smaller formats. So the graininess doesn't get enlarged so much, and is therefore less visible. Doesn't mean the graininess isn't there. Just that it's not so important.


As a side note, I'm planning to scan my negatives with a commercial Epson scanner. I have a 550 but might upgrade to the 800 series.

I see from many of the above replies that the discussion has ranged into "the grain question". Don't hurt me for pointing out some stuff that to me is (now after years of work) completely obvious to me but which many people skip over for whatever reasons. For completeness.

First, film is three dimensional. Second, actual silver grains are really small, as in microscope required. When you use said microscope to look through the film in a single spot, what you'll see in highlight areas (higher density) is a silver grain at one depth in the emulsion tends to "overlap" with other silver grains at other depths in the emulsion. They hardly ever "register" well, so the resulting set of overlapping silver grains is almost always bigger and more oddly shaped than any one of the components. This thing, is called a "grain clump". If you can see it in your image, it's a grain clump, not an actual silver grain. Because no enlarger or scanner can image down to individual silver grains.

Third, your scanner resolution is fixed over your scan (that is, set for the entire image before the scan begins). In the case of a drum scanner, scanning aperture is also set for the entire scan. That is, the scan properties are all deterministic and do not vary across the scan area.

Forth, the system of film grain clumps that make up the image are completely stochastic. This applies to grain clump size, which can cover a huge range of sizes. It also applies to grain clump locations.

Fifth, numbers 3 and 4 together guarantee that you're scanner resolutions and aperture size will never register well with your grain clumps. It's just not going to happen. So you need to disabuse yourself of the whole idea that there's a "correct" scanning resolution or drum scan aperture. Some are better than others for certain purposes, and this will vary from image to image. That's just the laws of physics talking to ya.

Sixth, remember that Callier Effect applies to scanning just like it does to optical enlarging. Light is light. Light (as used in photography anyway) does not penetrate solid silver. So it bounces off in another direction. More density -> more light scatter. More light scatter -> more artifacts. Again, just the laws of physics talking to ya.

Finally, remember the old adage: Expose for the shadows, develop for the highlights. This is true of darkroom printing, and it's just as true for scanning.

Add it all up together and it comes down to just this: use just as much film density as required to do the job, and no more.

It took me a decade to learn that stuff and hook it together to make logical sense in my head. And when I put it to practice I could make beautiful scanned huge prints of the rare (very) exposures that merited being printed large. So I know it works.

Do with it what you will.

[Peter De Smidt]

Bruce gives great advice!

[Pere Casals]

guarantee that you're scanner resolutions and aperture size will never register well with your grain clumps.

I fully agree with the rest, but my view is that this statement can be debated.

Dealing with the film grain structure in the hybrid process it's an interesting matter.

IMHO this is not often considered in LF, because grain is usually the least concern in LF, but with roll film it's a different war ...if wanting to exploit grain structure aesthetics.

[Peter De Smidt]

"can be debated" and "usefully be debated" aren't the same thing.

[Pere Casals]

"can be debated" and "usefully be debated" aren't the same thing.

Peter, it can be usefully debated.

Personally, I've been long time (and I'm) interested in film grain structure, and I've been following how several artisits exploit film grain structure in the aesthetics and trying to learn, and also I've been trying to find how film grain can be optimally processed in the hybrid.

I guess this should be debated in the lounge because it's not much about LF, but mostly about roll film.

Anyway let me point we need some 2500dpi effective to depict big grains, this is to have some 9 effective pixels (3x3) for a "dot" to record "big grain structure" in a film, but we need a top notch scan to depict well the "small" grains in the structure of the same film.

In optical prints many times we only perceive the "big grain", of course depending on the many factors we all know (film/developer/erlargement).

IMHO by recording well big grains we get a substantial share of the film footprint, but not all. Of course YMMV.

[Drew Wiley]

Don't confuse all this grain talk with acutance. Either TMax film will hold far more detail than HP5. But each of these has its own look.

[Pere Casals]

Here we have an interesting reading about the influence of wet mounting and light source collimation in the scanned grain.

It explains why some (Nikon/Imacon) scanners delivered "more grain".

https://web.archive.org/web/20150530210608/https://wetmounting.com/Pages/basics.html
https://wetmounting.com/Pages/basics.html

The wet topic is also covered here in page 20: http://www.tmax100.com/photo/pdf/film.pdf

[mmerig]

We have silver-halide crystals before development and metallic silver clumps after, we may call grain to both, but they are different things.

My view is that photographers and industry always call film grain to the metallic silver particles we have after development:

188842

"Fine grain" or "coarse grain" always refer to that... Me, I use "silver halide crystals" (or simply crystals) to refrer the "fundamental particles" in the emulsion (a suspension in fact) before development, but when I learn DIY emulsion recipes for my dry plates I find some "grain" word usage with the crystals.

It has to be noted that good technical literature always use crystal for silver halide particles and never grain, for example you will find epitaxial or tabular crystal growth depending on the pAg factor (silver concentration in the addition), a technician-scientist will never say grain for that, IMHO.

It has to be noted that with processing the "fundamental particles" (halide crystals) disapear. The shape of the clump depends on solvent effect, infectious development etc, but the individual crystal shapes are not there anymore, in fact the crystals disolve during development or fixing.

My view is that we should say crystals for silver-halide crystals, as we always are in a technical context when mentioning that then better if we say the real concept, and it's better to leave the grain word for the metallic silver that is mentioned everywhere in general film photography for developed film.


Anyway when scanning we cannot go wrong with semantics, crystals are destroyed in the processing. ...so for sure we speak about the metallic clumps.

Whatever they are called, the small particles, either before or after development, are too small for scanners to truly capture. Grain does not really clump together, it's just that there is more than one layer (about 50) of grains in the emulsion, so the exposed grains we see are stacked on top of each other more or less. If anyone was a technician/scientist, it was Mees, and he called both the silver halide crystals and the metallic strands grain. Here is a picture from his Dry Plates to Ektachrome book, showing one layer (Lippmann) of grains before and after development.

189872

I agree with you though, that different names should be used for undeveloped and developed grains. And from a practical standpoint, the individual grains don't have to be resolved to obtain a high-quality, very large print. A scan is a sample, so there will always be some losses compared to the original.

[Pere Casals]

are too small for scanners to truly capture ------------- showing one layer (Lippmann) of grains before and after development.
189872


This is not regular film, lippmann plates have extremly fine grains of about 0.01 to 0.04 micrometres (https://en.wikipedia.org/wiki/Lippmann_plate), while regular film may deliver 30 micrometer grains !!!!!

Grain can be well scanned, depending on grain size and scanning you may place 9 pixels on a single grain.


mmerig, see instead images of HP5+ grains, not Lippmann plates... Not the same 0.01 to 0.04 micrometres than 10 to 30 micrometres. Regular developed film has grains x1000 larger than the ones you mentioned.


Film fine grain takes one pixel in the around 1600dpi to 2500dpi range, so fine grains do require a good scanner, but larger grains are easier to see.

It is also true that films have also a share of extremly small grains that work in the shoulder of the curve, but what we call grain, it can be scanned.


These are hp5 grains (metallic clumps):

189873

189874

https://photo.fleurey.com/blog/pushing-ilford-hp5-to-1600-iso
https://web.archive.org/web/20180217070416/https://photo.fleurey.com/blog/pushing-ilford-hp5-to-1600-iso

[Corran]

What most people see as "grain," especially with higher-speed films, is simply grain aliasing caused by the scanner and grain interaction. What one "prefers" in terms of the look of the film / "grain" is completely up to the individual.

[Pere Casals]

Scanners that may produce aliasing have collimated light and with resoultion mainly limited by the pixel density. This was the case of LS9000.

Those scanners that are well limited by the lens performance do not produce aliasing.

[Corran]

There are a lot of variables but yes the Epson scanners can certainly cause aliasing.

[Pere Casals]

There are a lot of variables but yes the Epson scanners can certainly cause aliasing.

Bryan, not easy, because the sensor (V700 etc) outresolves the optics by a x2.5 factor, this is a lot. Same effect than with LPF in some DSLRs but way more intense, because of the x2.5 factor.

What an EPSON may show is some artifacts from the poor carriage drive, but I've never seen optical aliasing.

[mrred]

Either way, film is film. It's the same stuff on 35mm cuts. Your experience on the 35mm side is exactly the same on LF. The difference is how little it matters on LF.

[mmerig]

This is not regular film, lippmann plates have extremly fine grains of about 0.01 to 0.04 micrometres (https://en.wikipedia.org/wiki/Lippmann_plate), while regular film may deliver 30 micrometer grains !!!!!

Grain can be well scanned, depending on grain size and scanning you may place 9 pixels on a single grain.


mmerig, see instead images of HP5+ grains, not Lippmann plates... Not the same 0.01 to 0.04 micrometres than 10 to 30 micrometres. Regular developed film has grains x1000 larger than the ones you mentioned.


Film fine grain takes one pixel in the around 1600dpi to 2500dpi range, so fine grains do require a good scanner, but larger grains are easier to see.

It is also true that films have also a share of extremly small grains that work in the shoulder of the curve, but what we call grain, it can be scanned.
]

I used the Lippmann plates examples because the grains are in one layer, and it is easier to see the transition from salt to metal during development.

As you sort of mention, the size distribution of the usual film emulsions includes very small grains, matching that of Lippmann plates. The larger 'blobs" that can be called grain are made up of smaller and larger individual grains, in 3-dimensions converted to two. Your last sentence shows some inconsistency in what you are calling grains. Is it the smallest stuff, or just the big? There is also the reverse: in high-density areas of a negative, there are very small spaces between the grain "blobs" that a scanner misses as well.

Some of this grain stuff is moot, because the actual grains cannot be seen without an electron microscope, and cannot be seen with a light-transmitting microscope, and by extension, a super large optical enlargement. The examples of grain size using ordinary optics with light are too crude to demonstrate grain sizes. No one really knew what grain looked like until the electron microscope was invented, around 1940 I think. Mees tried (and failed) to see grain using ordinary optics over a century ago. You are still there.

[Pere Casals]

The larger 'blobs" that can be called grain are made up of smaller and larger individual grains, in 3-dimensions converted to two.


What we have after development/fixing is metallic clumps, and those continuous structures are large enough to be scanned. Just take an optical microscope and you will see what photographers call grain. If you want to see the silver halide crystals that we have before development then you need an electron microscope, but what photographers call grain is what we have after development, not before development.

Grain is what we may see in a very enlarged optical print, and scanners can scan those grains, depending on if the scanner is good enough and the grains are large enough. Film grain is not the substructures inside the grain, but the whole clumps.



Your last sentence shows some inconsistency in what you are calling grains.

Let me explain it better.

While (say) TMX has its "main" emulsion made of tabular grains, it has another layer under the "main" layer made of very small and low ISO grains that are cubic (and ortho IIRC). That layer is only sensitive to insane overexposure and eliminates the TMX shoulder, making TMX linear in the highlights.

That layer is also responsible of the high resolving power of TMX in the ultra high contrast conditions 1000:1 (TOC) reaching 200lp/mm, to scan those grains individually we need well better than 10,000 effective dpi in the scanner performance, but those grains are mostly irrelevant in the grain structure of a film, you won't see those grains in optical prints because the enlarger lens resolves usually less, and nor you will see those grains in scans. That layer may show (if we burn a lot in the print) the textures in the highlights, but those textures never show that small grain.

If that grain resolves 200lp/mm, to show it you need an enlarger lens resolving more, and enlarging to 200/6 = x33 to start a human eye noticing it a bit, perhaps at x100 that grain may be well perceived. So that grain does not count in the grain structure.

[mmerig]

Thanks for trying, but I am not convinced by your explanations -- too many contradictions, and so much depends on the definition of grain and semantics. I'll still go with the definitions and explanations from people that worked in film research.

[Pere Casals]

What contradictions?

developed film has grains (metallic clumps) of different sizes, some are seen in the print grain structure and some are too small to be seen, depending on:

the enlargement,
the particular film,
process N
developer's solvent,
enlarger/scanner illumination (condenser, etc)

you should know that... I guess you know it...

After development crystals are not there, we have metallic clumps, that are way larger than original crystals.

[Chester McCheeserton]

191284191285191286

not sure if this is really useful or me just blowing my horn. but the first image is 4x5 tri-x, and 2nd is 4x5 Fuji acros. Both drum scanned on an aztek....I had both types of film loaded and ended up deciding that the best picture needed to have a splice of each to create more of 5x7 format...wasn't something I did intentionally in shooting the def film types, just what i had... but it was interesting to see the difference in the same exact scene rendered in the different film stocks...not exactly what the OP asked about but likely not entirely dissimilar either....

they looked pretty different at 100% side by side on screen. and you (meaning one of you reading this) could see the difference in the sky in the final 30 x 45 inch print if you looked real close, I'm sure no one else noticed or would care even if they did notice.... the fuji was for sure finer grain and resolved more but not necessarily better....

[Pere Casals]

Both drum scanned on an aztek...

It looks a 2000 dpi scan, perhaps 1800 effective, LF sheets are usually sanned at lower dpi...

If you scan a crop in a drum at max dpi... say 6000 effective...

Grain is more important in MF/35mm, of course...

[interneg]

It looks a 2000 dpi scan, perhaps 1800 effective, LF sheets are usually sanned at lower dpi...

If you scan a crop in a drum at max dpi... say 6000 effective...

Grain is more important in MF/35mm, of course...

Any competent scanner with decent MTF response & a precise enough focusing system will show a detectable difference above about 600 ppi - the standard used for manufacturers testing is a 48 micron aperture - 530ppi for all intents & purposes. Assuming that at below maximum resolution high end scanners do not reach stated resolution is wrong, unless you can provide definitive evidence for the specific scanner in question. Generally, they are very close to the specified resolution up to 4-5000ppi where the very high mechanical precision requirements needed to maintain good resolution & contrast are rapidly affected by the quirks of the design, wear & tear etc. We are talking a handful of microns here.

That consumer grade flatbeds seem to give odd 'grain' response has more to do with their very low MTF performance (not that the optical package may be bad - there is a 6e/4g lens in many of the more expensive Epsons) from low precision construction which people attempt to compensate for by excessive sharpening which results in a disgusting mess of somewhat resolved 'graininess' & sensor noise.

[Pere Casals]

That consumer grade flatbeds seem to give odd 'grain' response has more to do with their very low MTF performance (not that the optical package may be bad - there is a 6e/4g lens in many of the more expensive Epsons) from low precision construction which people attempt to compensate for by excessive sharpening which results in a disgusting mess of somewhat resolved 'graininess' & sensor noise.

Interneg, there is another factor...

Ther V700 scan width is 5.9", in the X axis it resolves around 2800dpi effective, which is 2800x5.9 = 16520 effective pixels in a single pass !!!!! A machine has to do a lot of things well to resolve that.

You may know that the V700 scans four 35mm strips in a single pass, plus the borders...

Now imagine that the HR lens covers 4" only, instead 5.9", the V700 would deliver 2800x5.9/4 = 4130 effective dpi !!!! and 4130 effective dpi covering 4" in a single pass !!!!

The V700 is an impressive machine, but it prorizes making four 35mm strips in a single pass... for LF this is not a problem.

[Chester McCheeserton]

You may know that the V700 scans four 35mm strips in a single pass, plus the borders...


Using Epson scan software? With the 35mm film holders? or are you talking about just laying them on the glass.

The v700 is a great scanner, I have one, and a good photographer with a good negative can get great files to make huge prints with it. (at least if it's 4x5 or above)

But side by side, it pales in comparison to a good drum scanner...and I find it frustratingly slow to scan 35mm on it. Thinking of buying a pakon but the analog hipster kids drove the prices too high.

with 35 or 120 a drum scanner will blow the doors off the epson. And any macro digital camera set up.

[Pere Casals]

Using Epson scan software? With the 35mm film holders? or are you talking about just laying them on the glass.

With holders. The V700 sensor sees 5.9" wide, while a dedicated 35mm roll film scanners see 1", so performance of the V700 (2800 - 2300 dpi, depending on axis) is a miracle. Many 35mm shots don't require more resolving power, because handheld shots are not perfect, but if it is tripod shot, with a good lens in the sweet spot... a better scanner gets slightly better results.

______

MF and up it's a different war, the largest the format the V700 shines the more. For MF it's really competent:

https://www.largeformatphotography.info/forum/showthread.php?151116-Grain-on-large-format-(scanned-negative)-hp5-plus-vs-tmax-100&p=1489423&viewfull=1#post1489423
https://www.largeformatphotography.info/forum/showthread.php?151116-Grain-on-large-format-(scanned-negative)-hp5-plus-vs-tmax-100&p=1489522&viewfull=1#post1489522

A drum is better, of course, but many times the V700 (an other flatbeds) takes most of what the MF negative has.


And for LF it's a great choice, for 4x5 it outresolves the hasselblad X5 and most drum scans that are made at 2000dpi, a drum 4000 dpi scan is not that cheap, if wanting to drum scan 4000 dpi then its cheaper to shot 5x7 and scanning with the V700. And then V700 it makes 8x10.

Also V700 scans require always an optimization in Ps, while Pro gear serves the files digitally optimized with wise algorithms.


V700 drawbacks are BW grain structure in 35mm format, comming a bit short for 35mm film, and if wanting to recover crazy dense shadows in velvia, but the Multi-Exposure feature in Silverfast helps a lot.


______

The best a drum scan has is post processing, many times a pro drum has (or had) a very skilled proffesional near it, someone knowing how to refine a digital image while preserving quality, and having a sound criterion about how to make an interpretation that will shine in a monitor or on paper. This is what, many times, it makes the difference.