Poor scans - May explain why some switch to Digital

[mauro_franic]

<p>I do want one. I just want Nikon to make it for me.......</p>

[mauro_franic]

<p>Scott, I understand your point about the brain's completing an image. That is why sometimes street signs taken with a DSLR look ok and it seems you can read them but when you zoom in there is nothing there.</p>

<p>There is a limit though. I inserted a crop from the 5D2 over the Coolscan scan of Velvia 35mm. Can your eye complete the lines there?</p>

<div></div>

[rishij]

<p>Mauro-- huh? Your '4' mark corresponds to 3450 LPPH (4 x 863). I don't know where the heck you got that 5D Mark II crop that just looks like a grey box from, but, here's a proper comparison of the 5D Mark II's performance at 3450LPPH vs. your 35mm Velvia scan:</p>

<p><img src="http://staff.washington.edu/rjsanyal/Photography/FilmVsDigital/35mmVelvia_vs_5DMarkII-3450LPPH.jpg" alt="" width="317" height="556" /><br>

And that's precisely why I think that a 5D Mark II will outperform 35mm Velvia (scanned, anyway).</p>

<p>Again, be fair to digital, Mauro :)<br>

-Rishi</p>

[rishij]

<p>As proof that this was selected from somewhere near 3450LPPH for the 5D Mark II, here's a crop of the RAW resolution test chart shot for the 5D Mark II, blown up 400%, with sharpening applied in ACR (if you're sharpening your film scan, it's only fair to also sharpen the digital shot):<br>

<img src="http://staff.washington.edu/rjsanyal/Photography/FilmVsDigital/5DMarkII_3450LPPH.jpg" alt="" width="800" height="583" /><br>

Cheers,<br>

Rishi</p>

[mauro_franic]

<p>We must be using different chats. I am using the one from DPreview (I was comparing to the 4.2 mark equal to 3730).</p><div></div>

[mauro_franic]

<p>Here with Bicubic UPs:</p>

<div></div>

[mauro_franic]

<p>If we are using the same chart from DPreview, the 5d2 must resolve less in one direction than another???</p>

[rishij]

<p>You using the RAW version, Mauro? You can download the RAW conversions vis-a-vis ACR, DPP, or Camera 1.</p>

<p>Next, please sharpen before you upsample. You and/or NeatImage performed sharpening on the film scan.<br>

<br /> Finally, you really shouldn't place the 3730 LPPH result from the 5D II over the 3520 LPPH (sorry, I remeasured your chart height in Photoshop to get a new multiplication factor of 880, not 863) result from your film... it, again, makes the digital look much worse than it really is.</p>

<p>Here's the vertical resolution around 3600 LPPH for the 5D Mark II, ACR conversion, 200%:<br>

<img src="http://staff.washington.edu/rjsanyal/Photography/FilmVsDigital/5DMarkII_Horizontal3600LPPH.jpg" alt="" width="489" height="501" /><br>

-Rishi</p>

[mauro_franic]

<p>I measured 890 so there's no difference.</p>

[mauro_franic]

<p>I don't want to get confused with so meany numbers.</p>

<p>How much are you saying Velvia 35mm resolves and how much does the 5D2 in lpph?</p>

[rishij]

<p>Right, that's why I was correcting myself, because I was underestimating your Velvia resolution.</p>

<p>So with 880 as the multiplier, I'm getting:</p>

 

<ul>

<li>35mm Velvia 50 under microscope: <strong>26.2MP</strong> or <strong>4180 LPPH</strong> </li>

<li>35mm Velvia 50 + Minolta DSE 5400: <strong>20.5MP</strong> or <strong>3700 LPPH*</strong> </li>

<li>5D Mark II: <strong>18.9MP</strong> or <strong>3550 LPPH*</strong> </li>

</ul>

<p><strong>*</strong> : <em>absolute limiting resolution after oversharpening</em></p>

<p>I say 'oversharpening' because I'd never sharpen my film nor my digital SLR shots to the extent to which you & I have sharpened those Velvia test chart shots... :)</p>

<p>-Rishi</p>

[swilson]

<p>Mauro, could you post a link to the full Coolscan scan of Velvia 35mm?<br>

I am trying to make sense of the scaling between the chart and having the full scan would sure help.<br>

Thanks</p>

 

[mauro_franic]

<p>Oversharpening does not add detail but a raw converter maybe guessing - beyond me. </p>

<p>We all agree Velvia 35mm outresolves the 5D2, just figured it would have been by a larger margin.</p>

[mauro_franic]

<p>I only have the crop here (at 100% so you can measure in photoshop and calculate how many times the chart fits). Rishi do you have a full scan for Scott?</p>

<p> </p><div></div>

[rishij]

<p>This is a little cropped, sorry, quick & dirty:<br>

<img src="http://staff.washington.edu/rjsanyal/Photography/35mm_Scanners/Velvia-MauroResolutionChart_FullFrame.jpg" alt="" width="604" height="413" /><br>

My Minolta does the least bit of cropping... but I don't have that scan with me right now.<br>

-Rishi</p>

[DB_Gallery]

<p>Holy cow, this thread needs a table of contents!</p>

[swilson]

<p>Mauro, Rishi, thanks much that helps.<br>

 

<p>First off I am pretty impressed with the resolution of the Velvia scan. But it brings back my old frustration of scanned images not looking at sharp as what their resolution would lead one to believe they should look.</p>

<br>

 

<p>When I first started scanning film I figured my film scans would kill my digital photos, after all the film scans were around 10MP and the digital was only around 3MP. I was thinking that my film camera + scanner would be the same as having a 10MP digital camera, but it soon became clear that this was not the case.<br>

What ruins the photos for me is the noise in the film scans, whether is it from grain or the surface of the film does not matter, it makes it hard for me to enjoy the images at full resolution.</p>

<p>In the image below the film scan on the left most likely has a bit more resolution at the pixels level then the digitl image on the right. The one on the right is a 100% crop from my 350D, which has pixels very close to the same size as the 5D II.<br>

<img src="http://www.pbase.com/konascott/image/110636788/original.jpg" alt="" /><br>

What I seem to be finding over and over is to get a film scan to look really sharp at the pixel level it needs to be downsized to about 2000 ppi. In many cases detail will be lost going from 4000 ppi to 2000 ppi, but at the pixel level the downsized image is going to look far sharper. This is why MF scanned at 2000 PPI is going to blow 35mm scanned at 4000 ppi out of the water. </p>

<p>

<p>

<p>The scanner that impresses me will not be the one that resolves the most lines / inch but rather the one that produces the cleanest looking pixels. This is very much like the digital camera. </p>

 

 

 

 

</p>

</p>

 

</p>

[mauro_franic]

<p>Scott,</p>

<p>You have to compare same subjects side by side. You can sharpen or reduce noise (if it bothers you on the monitor) on film. Digital is already sharpened and has noise reduction applied for you, in addition data (color and luminosity) has an abrupt end, that make it look sharper but plasticy.</p>

<p>Here is the same velvia shot next to a 10MP DSLR (different cans because one died over dinner but exact same setting):</p>

<div></div>

[mauro_franic]

<p>And here is the result of several people trying to make best of the 40D raw:</p><div></div>

[mauro_franic]

<p>I still don't think it is possible to get results similar to 35mm film with a 10MP DSLR.</p>

<p>I have multiple related extensive threads and never had one person (although they relly tried hard) being able to produce a side by side example.</p>

<p>Also there is no grain/noise on the picture above that would show on any print.</p>

[mauro_franic]

<p>Obviously, the debate is between DSLRs and 35mm properly scanned.</p>

<p>6x7 film properly scanned (the purpose of the thread) is in a different league.</p>

[mauro_franic]

<p>Specifically to your comparison. I believe there will be situations where the 5D2 will outperform (detail wise) 35mm film. Also when a 40MP FF DSLR comes out with the same pixel density as the 50D, it will take an edge on film.</p>

<p>In the latter, since the information captured through a good medium format lens is larger, I still don't expect it to get close to MF. Dynamic range may suffer even more.</p>

[hary_diax]

<p>This whole discussion is really very interesting. Maybe someone with a good scanner is interested in doing a test to measure the number of pixels a good film / good scan holds.</p>

<p>Take 1 photo with a sharp film Test chart in the middle. Do a very good scan of it.<br>

Take some photo's of same object with a 10 mp dslr. Change the distance to the subject. Make distance smaller and stop when the objects in the middle are about the same sharpness as the photo with the film/scan<br>

Calculate how many pixels the whole scene would be if it had the same dimensions as the film.</p>

<p>Example<br>

Say you make photo of something 2 meters wide.<br>

With dslr you have to go closer to obtain same sharpness. Say you see same sharpness and at that moment you see only 1.3 meter of the scene<br>

If dslr = 10 mp. The film holds 2 meter/1.3 meter= 1.53 * 1.53 = 2,36 x 10 mp= 23 mp</p>

<p>It is easy to do.</p>

<p>Of course all with tripod and good light</p>

[mauro_franic]

<p>There is a great guy who did that already. Let me see if I can find the link.</p>

[mauro_franic]

<p>here:<br>

http://www.photo.net/film-and-processing-forum/00RV7N</p>