Taking into account that with increasing viewing distance detail become less obvious

- is the a rule of thumb by which a file of say 2 or 3 MB can be blow up without noticable deterioration in viewing quality??

- is the a better option to Photoshop enlargement tool?

2 or 3 mb is not really useful, you need the pixel size

Do you know what file format you are shooting in, or are they scans?

There are way too many factors involved to give a viable reply to the question. As already mentioned, the pixel count is important as well as the sharpness of the original photo. The resolution is another factor. I always shoot at the highest resolution a camera can provide, but many people fill up their camera with hundreds of shots shot at medium or low resolution. The quality of the camera's sensors is another consideration. The picture quality of an 8 mp Nikon or Canon SLR will usually be better than a 10 mp photo taken by a smaller point and shoot camera.

The older shots are done with a Canon G, 6million pixel, now I use more Panasonic TZ10, with 14 mm pixel.alll highest resolution.

Still don't understand correlation old 35 mm film and digital. E.g. I used a lot Ilford XP2 , wonderful b&w film. Did enlargements of up to 1m x1,5 m on photo linen . Not too much detail required.

Assuming an equally large sensor 24x36mm , how many pixel will equal the same detail? Grain is a given if used artistically.

In other words what pixel value will give the same quality output, be it Kodakchorme 64 or whatever.

When enlarging a negative beyond x you get grain but also no info. If you extrapolate with photoshop pixels new artificial pixel will be created, which still should be better in theory than no info but grain.

This brings me back to my question

- by how much you can enlarge x pixel, before it becomes in-viewable at a distance?

- are there better tools for extrapolation than Photoshop?

In this context the NY Times article from David Pogue is interesting

http://goo.gl/n0wep

He argues 10mm pixel will yield a print of 15x 20 inches. If that is a basis for an non-extrapolated print adding a factor of 2 ought be doable, or not?

The older shots are done with a Canon G, 6million pixel, now I use more Panasonic TZ10, with 14 mm pixel.alll highest resolution.

Still don't understand correlation old 35 mm film and digital. E.g. I used a lot Ilford XP2 , wonderful b&w film. Did enlargements of up to 1m x1,5 m on photo linen . Not too much detail required.

Assuming an equally large sensor 24x36mm , how many pixel will equal the same detail? Grain is a given if used artistically.

In other words what pixel value will give the same quality output, be it Kodakchorme 64 or whatever.

When enlarging a negative beyond x you get grain but also no info. If you extrapolate with photoshop pixels new artificial pixel will be created, which still should be better in theory than no info but grain.

This brings me back to my question

- by how much you can enlarge x pixel, before it becomes in-viewable at a distance?

- are there better tools for extrapolation than Photoshop?

In this context the NY Times article from David Pogue is interesting

http://goo.gl/n0wep

He argues 10mm pixel will yield a print of 15x 20 inches. If that is a basis for an non-extrapolated print adding a factor of 2 ought be doable, or not?

I'm not really sure what you're asking but maybe this will help.

I still shoot film and also use a 21mp digital camera.

The digital camera produces a 16 bit 300 dpi TIFF file of some 120Mb

At the same time a scanned 35mm slide, using the above parameters will produce a 178Mb file

With regards to upsizing PS Bicubic smoother is regarded as the safest to avoid interpolation artifacts i.e. jagged edges

Conversely Genuine Fractuals is highly regarded as is Topaz for upsizing although I have no experience of their use

Turning to detail comparisons i.e. resolution and resolving power, the general consensus is a digital image of 60Mp NATIVE is required to equate to a 35mm Velvia 50 iso slide.

There's a very good evaluation shown on the Luminous Landscape site (Here) comparing 2 x 40mp high resolution digital MF camera's and an 18mp FF Leica M. The Leica was sharpest!

Vulkan, that was exactly what I was looking for but that the dimensions would be so dramatically different, like shown in the Velvia=60MB I would never have thought. Now I understand why I could do mega blow ups from my Agfa Scala slides. I dont think that any extrapolation program could artificially create so much from nothing, i.e. no existing pixels.

I guess that settles my dream to have mega blow up from a 6MB file.

But can you give a guesstimate how big one can go with 6MB assuming you want poster quality?

Vulkan, that was exactly what I was looking for but that the dimensions would be so dramatically different, like shown in the Velvia=60MB I would never have thought. Now I understand why I could do mega blow ups from my Agfa Scala slides. I dont think that any extrapolation program could artificially create so much from nothing, i.e. no existing pixels.

I guess that settles my dream to have mega blow up from a 6MB file.

But can you give a guesstimate how big one can go with 6MB assuming you want poster quality?

I assume you're referring to the cameras native resolution of 6mb.

Also it depends on the dpi output elected - the higher the dpi the smaller the print (trade in for quality)

You can see all the changes in PS by going to Image/Image size and UNCHECK the RESAMPLE box. You can then type in whatever resolution you desire and the document size will change but the pixel dimensions will not. Then you elect quality v size.

Typically 6Mp will produce as 3072 x 2048 pixels. This will print as:-

180 dpi = 17.067 x 11.378 inches

300 dpi = 10.24 x 6.827 inches.............

.........and numerous other configurations.

The maths are to divide the pixel size by the desired resolution to establish print size i.e. 3072 divided by 300 = 10.24 (etc)

Albeit that the industry standard output is 300 dpi it's worth noting that "most" desktop printers optimise at 260 dpi.

In conclusion I would say the answer is that a 6mb image will print at 12 x 8 in general terms and "poster size" would fail.