8mm film strips from Epson v550 to avi file from slides

[mammo1789]

Hi Guys
I started of scanning some 8mm reels of film on an Epson flatbed v550 it has plastic cover for 35mm films and photos so I made one myself with black tapes from sides so it fit 8mm widens. Epson and other software (Vuescan I tried it but it is too complicated for me) I scan them as films negatives what is strange that scans are again negatives (35mm films for example are converted just fine and seen and scanned in normal color not inverted) after been scanned (I use Photoshop to just invert the slides and save them again as source 48bit 6400 4800 or 3200 dpi depends.
My question is what after that (how to make individual frames or video avi file to use fred script after that and so on) I saw some “tutorial” on internet (vimeo) of how to stitch the slides but the links are not working. From software side I have Photoshop, Premiere, after effects, ARCsoft stitch it (it does stitch the individual slides in one big jpeg but I don’t know what to do whit it also). I downloaded the software CinetoVID Pro but doesn’t want to find my slides. So the point is to get either individual full frames so I can use them in Virtualdub as video or to get Avi video file.

I converted them to jpeg the cinetovideo says " Error occurred. Could not scan strips. Is the scan program (e.g. the ScanController) specified ? " I don't know what that means.
I can extract individual frames in Photoshop, with 6400dpi one is more less 1365x1066 which i want to be 1280x720 video file or 1024 768

[VSz]

I'm not quite sure what you are trying to achieve, but if understand correctly you want to turn your 8mm film into a digitalized video?

When i did my super 8mm films, i didn't take the scanning route, instead what i did was purchase a cheap 8mm projector and screen from ebay, then set my 4K camcorder up and recorded the projected video. Then i cleaned up the video in editing software.

Maybe that method would be easier for you too?

[mammo1789]

Hi
I already did the point and shot route long time ago with dv interlaced camera. I scanned the slides properly and i think I can make it to faster change the films ( still manually thou ) my point is to use some software ( I can make 8 separate pictures in Photoshop and get 8 single frames per group and use them in virtualdub from picture sequence to make video but that is very slow ( considering that the scanning process itself is far slower than point and shot method with a camera) then use them to extract the slides (which has a group of 8 frames (32 total in one slide or 4 separate) for 8 mm films with 10 ( 2 half frames in slot that the 35mm holder can hold). maybe avysinth I don't know, I know that it is easier with camera but the quality is much much higher with scanner ( if done properly) thanks anyway

[zing269]

8 mm film isn't a negative so if you scanned them as negatives the scanning software converted them to a "positive". Scan them like a slide and you won't need to convert the image. But you don't really want to do it this way if the cinetovideo software is a dedicated film to video program.

A quick google didn't turn up any software called cinetovideo so this is just a guess but cinetovideo probably wants to do the scan itself and can't find the scanning software.

[mammo1789]

"8 mm film isn't a negative so if you scanned them as negatives the scanning software converted them to a "positive" " Yes i found that out, but other software that depend on cinetovideopro are some other software that he mentions but cannot be found
https://vimeo.com/30294450links:
and now this
https://vimeo.com/22015773

[zing269]

The tools mentioned in that vimeo post and the ones list at the wkurz.com site are available in the tools section of this site. I looked at that site and wiki and took a brief look at the help files and it looks like you can skip the actual scanning part and process your already scanned film if you set up the project to take into account the existing folder paths and image file names. There's a forum dedicated to CineToVidPro at http://hosting.aktionspotenzial.de/CineToVidWiki/index.php/Spezial:AWCforum/ that may prove useful.

[mammo1789]

Thanks zing269 yesterday the page was not opening for me probably was my isp or av, today I opened it I also saw the pictures of his setup but he scans horizontal and slides the film under the sandwich of scanner glass and another glass. Is this ok ? I am afraid to slide the film over the surface of the scanner ( does v550 has some couting for protection of scratches like gorilla glass on mobiles or not ?) my films are all very curly and raf and I think if I slide them it will scratch the glass of the scanner and I dont want to do that or to use some tablet protection glass on top of the scanner glass then slide the film over the protective glass and it is thin and. I basically use the same 35mm plastic holder for the scanner and put tapes around it the same as he is doing and in the end it has the same strips, The scratch of scanner glass is also what concerns me, also some of my films dont have that black boundaries some of them are white or even reddish color so I will try to use photoshop first on the whole sequence then using the extracting software and see how it goes. Till now I wasn't able to figur it out how to open the strips and scan them in video. Is there any other software that can do that maybe paid. To do simple extract individual frames rotate crop and line them then make mov or avi in motion

[zing269]

The scanner was never meant to be used with film so I doubt that the glass has any scratch protection properties. I've only scanned photo negatives, never messed with film, so I'm not aware of any software designed to do the conversion. It does seem that his software does have settings to deal with curl, he calls it pincushion, and ways to deal with less than perfect film so I would just try the frame extraction with your current scans before using Photoshop. Good luck with your project.

[smrpix]

The best labs in the world will give you a student rate of about 20 - 25 cents a foot to scan it professionally. How much material do you have?

[mammo1789]

I tried to just stick the film on the scanner and scanned some strips which looks pretty good without nothing just film from one side to the other. Scaning seems not to be the problem the problem am having is ho to automatically like he to make individual frames I scanned them in 6400dpi 48 bit and they look much better quality than his if I can just find out to make this automatically I will post some slides. This is flatbed for film also not just regular one, I know that someone can make it for me ( in my country there is no such service) but I just want to do it my self,. What I did is scaned the strip croped in epson soft 6400dpi 48 bit tiff files ( the software suports only jpeg) and I want to avoid any compresion to jpeg prior to making avi file. Then I opened photoshop rotated the image as vertical and crop 15 frames per scan wich i will get indidividual 15 fps 1 sec video and if i scan by 15 every scan can be a sec of video.
Here some pictures
thanks

[smrpix]

Experiment with a few of the strips to be sure you aren't getting geometric or lighting distortion as you scan from one edge of the bed to the other.

[mammo1789]

My problem is extracting the frames from the film to make avi video not the scanning itself

The worst part of the color film it wasn't even watchable on projector or dslr camera, I used the backlight comp on the scanner
bellow some

[jagabo]

Here's an AviSynth script that takes the image in post #12 and converts it to a video. You could use VirtualDub to open the script and export it as a BMP, PNG or JPG sequence. Or as an AVI file.

global FRAME_WIDTH = 1408
global FRAME_HEIGHT = 1056
global NUM_FRAMES = 15

global TOP_LEFT_X = 192
global TOP_LEFT_Y = 8
global BOTTOM_LEFT_X = 226
global BOTTOM_LEFT_Y = 15840


function CropOne(clip v, int framenum)
{
    xpos = TOP_LEFT_X + framenum * (BOTTOM_LEFT_X - TOP_LEFT_X) / NUM_FRAMES
    ypos = TOP_LEFT_Y + framenum * (BOTTOM_LEFT_Y - TOP_LEFT_Y) / NUM_FRAMES

    Crop(v, xpos, ypos, FRAME_WIDTH, FRAME_HEIGHT)
}


ImageSource("kur018.jpg", start=0, end=NUM_FRAMES-1, fps=18) 
Animate(0,NUM_FRAMES-1, "CropOne", last,0, last,NUM_FRAMES-1)

FRAME_WIDTH and FRAME_HEIGHT are the width and height of the frames cropped from the full image. You'll want to keep this the same for the entire video.

NUM_FRAMES is the number of frames to crop out of the full image. This will probably be the same for each input image except the last one.

TOP_LEFT_X and TOP_LEFT_Y are the X and Y coordinates of the top left pixel of the first (top) frame. You have to get these values manually in an image editor. They will probably vary with each input image.

BOTTOM_LEFT_X and BOTTOM_LEFT_Y are the X and Y coordinates of the top left pixel of the frame past the last frame being saved. Again, you will have to get these manually in an image editor and the values will probably vary with each input image.

You can deshake, crop, and clean it up later. Sample raw video attached.

[mammo1789]

Thanks Jagabo that is exactly what I was looking for, The pictures that I uploaded are jpegs, I will use with tiff files this black and white film was just a test to see where to start I also have 5-6 color 8mm and super8mm films. Do you think it is better to use slides by 15 frames (like the one that I uploaded) and make lets say 1 min clip from 60 one sec clips use the script and then stitch them or do you think it is better to stitch the slides first and then use the script on the whole slide.

If i understood the script I used the same thing on the other slide color png (uncompressed file) and the frames go from top to botom not like the other black and white slide I cropped it and

BOTTOM_LEFT_X and BOTTOM_LEFT_Y are the X and Y coordinates of the top left pixel of the frame past the last frame being saved

The script looks very fast and without other software which is good
The questions I have is
Can I make true 32bit lagarith video from 48 bit pngs ( can script accept such high bit depth of pngs uncompressed) and keep it 16 (18) fps of the original film in true full hd lagarith 32 bit videos for archiving and make h264 copies for watching on tv.

I am still confused about the order of the frames in slide up down or left right, before using the script (example of another color slide bellow is sliding up down not like the original that you made in continuus left to right maner

Is there limitation of number of frames ( if not I can align and stitch individual slides in one full long slide and then use the script on that slide that way I can have the whole movie in one avi file mine films are not longer than 2 3 minutes home made )

[jagabo]

Do you think it is better to use slides by 15 frames (like the one that I uploaded) and make lets say 1 min clip from 60 one sec clips use the script and then stitch them or do you think it is better to stitch the slides first and then use the script on the whole slide.

AviSynth probably has some limit on the frame size. I'd stick with ~15 frame source images. That will also help with keeping the string of frames aligned. See this sample made from your stitched image.

global FRAME_WIDTH = 620
global FRAME_HEIGHT = 480
global NUM_FRAMES = 24
global TOP_LEFT_X = 980
global TOP_LEFT_Y = 310
global BOTTOM_LEFT_X = 1000
global BOTTOM_LEFT_Y = 11770


function CropOne(clip v, int framenum)
{
    xpos = TOP_LEFT_X + framenum * (BOTTOM_LEFT_X - TOP_LEFT_X) / NUM_FRAMES
    ypos = TOP_LEFT_Y + framenum * (BOTTOM_LEFT_Y - TOP_LEFT_Y) / NUM_FRAMES

    Crop(v, xpos, ypos, FRAME_WIDTH, FRAME_HEIGHT)
}


ImageSource("Stitched.jpg", start=0, end=NUM_FRAMES-1, fps=18)
TurnRight()
Animate(0,NUM_FRAMES-1, "CropOne", last,0, last,NUM_FRAMES-1)

Note the changed parameters and the addition of TurnRight. There's no need to crop the source image, the script does that. On the other hand, if you cropped the source image exactly the same way every time you wouldn't need to modify the script with each image.

BOTTOM_LEFT_X and BOTTOM_LEFT_Y are the X and Y coordinates of the top left pixel of the frame past the last frame being saved

this part I don't understand

The sample image had 15 frames. Those coordinates are the top left corner of the 16th frame. Or one pixel below the bottom left corner of the 15th frame.

[mammo1789]

I tried with different file and the result are on upper post but my video goes from up to down instead of left to right
Omg stitched one very good, How do you find out the coordinates in Photoshop or avspmode

"Note the changed parameters and the addition of TurnRight. There's no need to crop the source image, the script does that. On the other hand, if you cropped the source image exactly the same way every time you wouldn't need to modify the script with each image."

Yes I noticed it but if use it it says error cannot crop something, the last slide with the kid is giving me up down instead of left to right move

[poisondeathray]

How do you find out the coordinates in Photoshop or avspmode

Move the cursor with mouse . The location here you point to will show up as the x= , y= coordinate in the status bar

[jagabo]

Here's the script modified for "test col 1.jpg"

global FRAME_WIDTH = 1080
global FRAME_HEIGHT = 956
global NUM_FRAMES = 18
global TOP_LEFT_X = 0
global TOP_LEFT_Y = 40
global BOTTOM_LEFT_X = 0
global BOTTOM_LEFT_Y = 17164


function CropOne(clip v, int framenum)
{
    xpos = TOP_LEFT_X + framenum * (BOTTOM_LEFT_X - TOP_LEFT_X) / NUM_FRAMES
    ypos = TOP_LEFT_Y + framenum * (BOTTOM_LEFT_Y - TOP_LEFT_Y) / NUM_FRAMES

    Crop(v, xpos, ypos, FRAME_WIDTH, FRAME_HEIGHT)
}

ImageSource("test col 1.jpg", start=0, end=NUM_FRAMES-1, fps=18)
Animate(0,NUM_FRAMES-1, "CropOne", last,0, last,NUM_FRAMES-1)

Note that I'm only changing the variables at the top of the script, and the name of the source clip. The rest of the code doesn't have to change.

[mammo1789]

Move the cursor with mouse . The location here you point to will show up as the x= , y= coordinate in the status bar

Is this it and you are talking in photoshop yea poisondeathray?

global TOP_LEFT_X = ?
global TOP_LEFT_Y = ?
of the frame or the whole picture ?

[poisondeathray]

In avspmod . It makes sense to read the coordinates there since you can enter the values into the script directly

Push f5 for the preview. Mouse over a position and where ever you "point" to will be indicated as the x,y position . It might not be setup in your avspmod configuration, but you can program the status bar to show many things besides position like YUV values, RGB values, hex values etc....

[jagabo]

I originally wrote the script assuming the strip would be vertical. But it works for horizontal strips too. Here's the script for the long stitched image in the first post, without rotating it first:

global FRAME_WIDTH = 480
global FRAME_HEIGHT = 620
global NUM_FRAMES = 24
global TOP_LEFT_X = 320
global TOP_LEFT_Y = 1482
global BOTTOM_LEFT_X = 11770
global BOTTOM_LEFT_Y = 1456


function CropOne(clip v, int framenum)
{
    xpos = TOP_LEFT_X + framenum * (BOTTOM_LEFT_X - TOP_LEFT_X) / NUM_FRAMES
    ypos = TOP_LEFT_Y + framenum * (BOTTOM_LEFT_Y - TOP_LEFT_Y) / NUM_FRAMES

    Crop(v, xpos, ypos, FRAME_WIDTH, FRAME_HEIGHT)
}


ImageSource("Stitched.jpg", start=0, end=NUM_FRAMES-1, fps=18)
Animate(0,NUM_FRAMES-1, "CropOne", last,0, last,NUM_FRAMES-1)
TurnRight()

TurnRight() at the end rotates the final result so it's in the correct orientation.

[jagabo]

Here's how the script works. Here we are going to save three frames, NUM_FRAMES = 3:


[Attachment 40611 - Click to enlarge]

You provide the coordinates of the top left corner of the first frame and the fourth frame. They are marked in green in the above image (oops, frame 4 is marked in cyan). The program then linearly interprets the location of frames 2 and 3 between those two points, marked in red in the above image. It then crops out FRAME_WIDTH by FRAME_HEIGHT frames at those points.

If the frames are oriented horizontally rather than vertically it works the same:


[Attachment 40610 - Click to enlarge]

You give it the top left corners of frame 1 and 4, the program interpolates the position of frames 2 and 3.

[mammo1789]

Thanks very much Jagabo always helpful and knowledgeable


global BOTTOM_LEFT_X = 11770 I assume this are x= frame 1 right corner or ( frame 1 left low corner? )
global BOTTOM_LEFT_Y = 1456 ? y= frame 4 right up corner or low left corner

And I don't see the coordinates in avspmode?

[poisondeathray]

And I don't see the coordinates in avspmode?

Watch this screen capture. Look at the bottom status bar where it says pos(xxx,yyy) . That is your x,y position corresponding to where your cursor is pointing to


If your scans are all the same alignment, then you might be able to use the same script for them all in a batch. But if you "wiggle" it a bit, you're going have to manually enter the values for each sheet scan

[poisondeathray]

Didn't someone (John Meyer , or StainlessS ???) at some point try to tackle something like this by a sprocket alignment detection avs script ? For each reel that has it's own sprocket characteristics, I think they tried to automatically align and cut according to detection logic geared towards the light coming through the sprocket

[jagabo]

Since I originally wrote the code assuming the frames were stacked vertically I named the the coordinates at the start "TOP_" and at the end "BOTTOM_". But if the frames are stacked horizontally "TOP_" is really "LEFT", and "BOTTOM_" is really "RIGHT". The coordinate is always the top left corner of a frame.

Since you seem to be having problems understanding this I changed the code to reference the frames as FIRST and LAST, and changed the last frame to be the last frame that's saved, not the one after it. I also changed the filename to a variable so everything you might need to change is at the top of the script:

global FILENAME = "Stitched.jpg"
global NUM_FRAMES = 24
global FRAME_WIDTH = 480
global FRAME_HEIGHT = 620
global FIRST_TOPLEFT_X = 320
global FIRST_TOPLEFT_Y = 1482
global LAST_TOPLEFT_X = 11300
global LAST_TOPLEFT_Y = 1458


function CropOne(clip v, int framenum)
{
    xpos = FIRST_TOPLEFT_X + framenum * (LAST_TOPLEFT_X - FIRST_TOPLEFT_X) / (NUM_FRAMES-1)
    ypos = FIRST_TOPLEFT_Y + framenum * (LAST_TOPLEFT_Y - FIRST_TOPLEFT_Y) / (NUM_FRAMES-1)

    Crop(v, xpos, ypos, FRAME_WIDTH, FRAME_HEIGHT)
}


ImageSource(FILENAME, start=0, end=NUM_FRAMES-1, fps=18)
Animate(0,NUM_FRAMES-1, "CropOne", last,0, last,NUM_FRAMES-1)

This code works in any direction: left to right, right to left, top to bottom, or bottom to top. You could even use it diagonally but the framing won't be right.

Using a 3 frame example:


[Attachment 40613 - Click to enlarge]

You supply the coordinates of the top left corner of the first frame you want to crop and the last frame you want to crop, 1 and 3 in this example. The script interpolates the position of the second frame.

Note that AVSPMod and VirtualDub show frame numbers starting with zero. So the three frames are shown there as 0, 1, and 2.

[mammo1789]

Thanks Jagabo I tested again and the first strip I make adjustment on the script to be for that slide and the first stripe is made in avi just fine if I use the script on different strip ( of 16 frames ) almost identical as previous it came with error of crop the resolutions are the same and the w and h of the stripes are the same

[jagabo]

if I use the script on different strip ( of 16 frames ) almost identical as previous it came with error of crop the resolutions are the same and the w and h of the stripes are the same

Did you get an error something like "You cannot use Crop to enlarge or shift a clip"? You have to make sure the crop never extends outside the source image.

[mammo1789]

"You cannot use Crop to enlarge or shift a clip"

Exactly like that

[jagabo]

Exactly like what? Post the problematic script and image.

My guess is you wrote the script using image #1 and you're trying to use the same script on image #2 or #3. But those images are smaller so the last Crop is falling partially outside the source image.