PTGui panorama Maximum resolution

[Liviu Popica]

Hello, what is the maximum resolution output , i have a phase one camera and i want to create gigapanoramas but i can't find any info about the maximum resolution in pixels supported, and the depth output for 16 bit images cause i want to print it on 10-15 meters long by 2 meters wide, and after that to edit them in photoshop. thank you!

[PTGui Support]

Hi Liviu,

There's no real limit. 15x2 meters at 300 dpi should not be any problem,
if you have enough disk space and patience.

See 6.11 for more information:
https://ptgui.com/support.html#6_11

Use a computer with fast NVME disk storage and the maximum amount of RAM.

Editing in Photoshop might be a different story though, I don't know how
well it handles multi gigapixel images these days.

Kind regards,

Joost Nieuwenhuijse
www.ptgui.com

> --
> You received this message because you are subscribed to the Google
> Groups "PTGui Support" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to ptgui+un...@googlegroups.com
> <mailto:ptgui+un...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/24d449ac-3a0a-438e-aae2-e8d93a512631n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/24d449ac-3a0a-438e-aae2-e8d93a512631n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Erik Krause]

Am 11.05.2023 um 14:51 schrieb 'PTGui Support' via PTGui Support:

> Editing in Photoshop might be a different story though, I don't know how
> well it handles multi gigapixel images these days.

If I calculate correctly, a 15m wide image at 300dpi would be roughly
180000 pixels wide, which is well below the 300000 pixels limit of even
older photoshop versions. And 300dpi is a pretty high resolution for
such a wide image.

The usual resolution for a 1m sized poster is 100dpi, larger images are
often printed at 50dpi or even 20dpi. So in theory, you wouldn't even
need a gigapixel. Half a gigapixel would still print at 100dpi.

The actual resolution you need depends on the distance from which the
print is viewed. The ratio is given here f.e.:
https://www.numinoos.com/quick-printing-guide


--
Erik Krause
http://www.erik-krause.de

[Liviu Popica]

Thank you for your response, i intend to print it on a 600dpi or more, the maximum resolution the paper  is 2400dpi and also the printer is a fine art printer, and the result will be very big! i want to show the full capabilities of the phase one IQ4 combined with the size of multiple frames from a panorama! 

[Mike Cowlishaw]

Just curious .. if this is to be a panorama composed of multiple images, what's the advantage [other than needing less time and fewer images]  of using the Phase One IQ4 over a lower-resolution camera?

For example, just last week I took an 102-image panorama of Kenilworth Castle whose default stitch with PTGui was 108797x22404  (I chose to reduce it to 65500x13488  to stay within JPEG capabilities, which resulted in a 340MB file).  The camera sensor was 25Mpel using 50Mpel mode and lens was a 200mm prime.  Capture time was a little under 15 minutes.  (I plan to post the resulting image as soon as that gets to the top of my to-do list.)

[Erik Krause]

Am 12.05.2023 um 11:05 schrieb Liviu Popica:
> Thank you for your response, i intend to print it on a 600dpi or more, the
> maximum resolution the paper is 2400dpi and also the printer is a fine art
> printer, and the result will be very big! i want to show the full
> capabilities of the phase one IQ4 combined with the size of multiple frames
> from a panorama!

As Mike stated: The resolution of a stitched panorama is pretty
independent of the used camera. It all depends on the lens used.

And for anything beyond 300dpi you will need a magnifier to see the
details. The unaided human eye can't resolve finer.

However, there is some confusion about the term dpi while printing:
Often it is referred to the print dots of an inkjet printer, in which
case it has little to do with the pixel resolution. Since an inkjet
printer can only print a dot or no dot, it needs a specific amount of
dots to print one pixel. The more different colors it can print, the
more dots you need for a pixel. Hence the paper resolution is the number
of print dots, not the number of pixels, same as the advertised printer
dpi.

[Mike Cowlishaw]

(Forget to mention .. this was a µ4/3 camera, so the lens was equivalent to 400mm on full-frame.)

[Liviu Popica]

indeed, 300dpi is enought but it is not about the details 300dpi, is about the colors, the details from the darkest black point to the whitest. there i need above 300dpi, a 300dpi paper capabilities is like a 8 bit jpeg image, a 256 levels of grading from the darkest point to the whitest ( red/green/blue) or just like in print CMY and K, but then we use a 12 colors printer the quality of the print changes!

[Liviu Popica]

When we use a Phase One IQ4 we got another range of quality, for example a 16bit RAW from a Phase One has 4x the dinamic range/colors increased over the 14 bit raw cameras (full frame/1.5x.1.6x or 2x crop factor)

This means that i have to do fewer images over HDR, so in many situations no braketing needed, as you said less time and also fewer images for the final shot! When the enviroment changes very rapidly, the sunset/sunrise, the clouds, are not moving as fast. when the perfect conditions are met time is key, we may got the chance to get the same conditiones next day or maybe next month or next year! light changes very quickly when shooting landscapes. When we can control the light off course that you can work and have more time to finish a panorama, but to landscapes is a total different world! 

the digital image is composed by bytes data, a jpeg has 8 bits of data, this means that every photodiodes from the sensor camera knows to capture light in 256 different steps (from the darkest to the brightest) the same case in displays LED, knows to light every Red/Green/Blue (RGB) led in maximul 256 different steps, when all the RGB led works toghether they compose a color, for example Red is 238/green is 67 and blue at 145 is created a pink color or #ee4391 known as his unique #code. 

So from the 8bite image to a 10bit image or how they call now HDR the TV-s there is a 4x the quality of colors

8bit/channel or 24bit color depth has - 256 levels from darkest point to the whitest (16.7 millions of colors: 256 levels of red x 256 levels of green x 256 levels of blue)

9bit/channel has - 512 levels from darkest point to the whitest

10bit/channel has 1024 levels from darkest point to the whitest ( 1.07 billions of colors : 1024 levels of red x1024 levels of blue x1024 levels of green )

11bit /channel 2048 levels

12bit /channel 4096 levels

13bit/channel  8192 levels

14bit/channel has16384 levels for every red, green and blues photodiodes from the sensor can capture the darkest point to the whitest, here are most of the consumer DSLR and Mirrorless cameras

15bit/channel has 32768 levels

16bit/channel has 65536 levels  for every red, green and blues  photodiodes, this are the Phase One sensor capabilities and also the high quality cinema cameras ( red digital, arri alexa, etc)  (281474 billions of colors that can recreate) 

Some scientist said that the human eye has about 24bit/channel capabilities in levels of capture from the darkest point to the whitest and the eye can distinguish about 10 million colors. but in using this camera is not only about the colors, is also about the HDR capabilities and the sensor size that is 53,4. by 40mm. The image has a more natural feeling. Also the lens has better quality in color rendering, this means less distorsions, less chromatic aberations. 

[Erik Krause]

Am 15.05.2023 um 16:52 schrieb Liviu Popica:

> indeed, 300dpi is enought but it is not about the details 300dpi, is about
> the colors, the details from the darkest black point to the whitest.

Exactly. And since any image pixel can already have 16 million colors,
you need more print dots to reproduce them, since a print dot is always
either on or off. That's why paper is classified as 2400dpi, meaning it
can resolve 2400 print dots as single dots, instead of smearing them
into one larger blob. And that's why ink jet printers have a way higher
dpi than your image should have.

Hence, the use of "dpi" to specify the pixel resolution of an image is
actually wrong. It should be named "ppi" - pixels per inch.

If you print a 600ppi image with 1200dpi there are only 2x2 = 4 print
dots per pixel. This means that per color channel, only 2^4 = 16
different shades can be printed. And that's why any printer has a native
pixel resolution, that is far below the nominal dpi value. In fact, the
printer driver (or the RIP) reduces your image resolution to the native
pixel resolution of the printer if it's higher. So there is no use to go
beyond the native pixel resolution of the printer.

Also see f.e.
https://creativepro.com/pick-right-resolution-printing-photos/

[florianjosef]

Markus Bredenfeld produced a 120 m x 3 m long panorama in an Underpass for cyclists and Pedestrians in eastern Tirol City train station Lienz

https://images.app.goo.gl/MhZTTcqDH4pW2YcS9

https://images.app.goo.gl/gncfAhzZK8AFGzqq8

https://bahnhof.oebb.at/de/tirol/lienz
Look at the Bildergalerie

It defintly works with PTGUI

Calculating resolution is also dependent on the viewer's distance to the surface.

[Mike Cowlishaw]

Some scientist said that the human eye has about 24bit/channel capabilities in levels of capture from the darkest point to the whitest and the eye can distinguish about 10 million colors. but in using this camera is not only about the colors, is also about the HDR capabilities and the sensor size that is 53,4. by 40mm. The image has a more natural feeling. Also the lens has better quality in color rendering, this means less distorsions, less chromatic aberations. 

<sigh>  "Some scientist" gets blamed for a lot of strange data :-).

The eye/brain system can indeed cater for a very wide range of lighting conditions -- but for any one lighting condition (e.g., when viewing a print) it has vastly less adaptability.  That's why for a standard display screen at a standard viewing distance (about arm's length) you only need about 9 bits per pel (pixel) for colour images,  or 4 bits for monochrome.   It's convenient to use more for calculating and processing images but it's not necessary for their display.  Indeed, for decades computer monitors used a colour palette of 256  colours (8 bits).

For the explanation of this and more detail, here's a paper that I wrote almost 40 years ago and which was published in 1985: 'FUNDAMENTAL REQUIREMENTS FOR PICTURE PRESENTATION' (  https://speleotrove.com/misc/cowlishaw1985-fundamental.pdf).  This research was what kicked off the JPEG Standards work, although I got moved to another interesting project (the computerisation of the OED) and so did not directly participate in  the JPEG work except as an advisor.

Mike


 

[Velson Horie]

Just because we cannot see the detail does not mean it does not exist.  That is what microscopes and telescopes are for.

Film X-radiographs have a large range of densities, far more than one can see with standard back lighting. When I was examining them, I changed the intensity of the light used for transmission viewing from mild to glaring, depending on features at the local density. Sometimes it was necessary to use glancing reflected light to see the varying thickness of the emulsion to extract more data.

The idea that one can get a decent idea of reality by looking at an 8 bit computer screen is the triumph of the virtual over the real. A step forward would be to view a 16 bit image on a 8 bit screen by sliding the 8 bit window of viewing up and down the 16 bit scale.

[Erik Krause]

Am 22.05.23 um 13:13 schrieb Velson Horie:

> Just because we cannot see the detail does not mean it does not exist.
> That is what microscopes and telescopes are for.

Do you usually view your printed images with a microscope?

Of course, you can provide magnifiers to your audience, like Joergen
Geerds did many years ago on an exhibition of his prints:
http://luminous-newyork.com/

But if you don't do, more than 300ppi is a waste of resources, and that
would be for a viewing distance of 30cm.

--
Erik Krause

[Velson Horie]

I never print the images I make. I tried some of the best art printers I could and they topped out at 600 dpi. It looked as though the images had been printed in mud.

Images are made available on line using iiif at the highest resolution that shows the original detail, so the users can compare them at the resolution of the individual engraved line on the original C19th impressions from metal plates.  The resolution of the lines on these impressions varies from the equivalent of ca 1200 to 3200 dpi. I am still working on that estimation and how to make it.

This is comparable to what the RijksMuseum did for the Black Watch.

So yes, I do view the original printed impressions with an analogue lens, and expect the resulting images to be viewed at the same level of detail, digitally.

Velson

[Erik Krause]

Am 22.05.2023 um 17:06 schrieb Velson Horie:

> Images are made available on line using iiif at the highest resolution that
> shows the original detail, so the users can compare them at the resolution
> of the individual engraved line on the original C19th impressions from
> metal plates.

Interesting! iiif means what is described on https://iiif.io/ I suppose,
but what do you mean by "original C19th impressions from metal plates"?

However, this has little to do with the original question, which was
about printed landscape panoramas.

[Velson Horie]

I was answering Mike Cowlishaw's point. And it is about  "PTGui panorama Maximum resolution", but my panoramas are landscapes painted by Turner, translated onto metal printing plates, printed on paper then scanned and stitched - using PTGui.  Some of them are 23 GB.

It all depends on the context.

[Mike Cowlishaw]

Hi Velson, I completely agree with you.   The source of an image is where the data and the details are.

My 'point' is that -- however enhanced or altered -- ultimately the eye is limited in its capability and when the data is finally presented to the eye it only requires a limited amount of data per pel.  And as Erik points out, the original question was about the presentation of an image to the user, and that's essentially the same whether using a print, a computer screen, or any other device.

Possibly interesting is that I came to this conclusion after working with X-rays.  Back in 1982 the group I worked in (in the IBM UK Scientific Centre) together with (I think) UCL were the first to digitise X-ray films and then show that using digital enhancements the users of the X-rays could make better and more accurate diagnoses than using their traditional methods.   We then built a state-of-the-art 1000x1000 monitor to view the digital images -- which is when we found that we had over-engineered it and as it was a monochrome monitor it really only needed 4 bits per pel whereas we designed and built a monitor with a 10-bit (per pel) display buffer.  Memory was very expensive then and was most of the cost of the monitor, so  we spent 2.5x more than necessary.  A lesson learned :-).

As you say, the original data is ultimately what matters.  There's a huge amount of detail in my recent panoramas (see https://speleotrove.com/gallery/Kenilworth_Castle.html) but the eye cannot process more than a tiny amount of it.  For example, to see the full detail available in the largest images there one has to magnify (enlarge) by a factor of 28.

Mike 

[Panoramabuch Panoramabuch]

in fact, i targeted the panorama width exactly to the PS max size of 300k to allow a complete edit and color correction. the resolution was calculated to 60dpi which was pretty sufficient for a viewing distance of 3 meters due to the bicycle lane between the panorama glass wall and the pedestrian walkway. technically the print was made sectionally for each of the 98 glass panes and upscaled without upsizing to 100 dpi as the minimum printer's resolution. besides the usual bottlenecks (mainly HD speed for swapping and some waiting times due to sheer image size) the project's size wasn't any problem, neither for PTGui nor for PS.

here are some more pictures: https://bredenfeld.art/portfolio/landschaft-vier-sonnen-wv397/