Detail Settings with cinema lens adapters and possible film out...

[Michael Maier]

Originally Posted by Sean Adair

My feeling is if in doubt, shoot with less detail processing in camera.
It would be a lot harder to remove edge and artifacts from too much detail, while post-production sharpening and/or edge enhancement should be just as effective (I suspect much better), and adjustable to your specific shots and final delivery format. Especially for a film-out project.
Correct me if I'm wrong - is there anything about the signal path or method of in camera detail that is inherently superior to post-proccessing or is it just time? I realize you wouldn't want to do this in a HDV workflow with re-compression to mpg2.

Originally Posted by Paolo Ciccone

Sean, you are correct. The so-called "detail" is an edge detection function that creates an outline for high-contrast areas. It's fast because it's implemented in hardware and it uses the higher color resolution of the sensor but it's an artificial edge enhancement that adds noise to the image, reduces the amount of compression achievable and it's irreversible. The camera does not create blurry images with the detail set to off or MIN. You can add edge enhancement in post and in doing so you can take advantage of advanced tools like masking.

If you have access to a waveform monitor I suggest that you shoot a chart, even a print out of a 11-step grayscale pattern will do. Once you have the chart correctly lit and centered you will see the characteristic "X" shape. Now set the detail to off, take a snapshot and then turn the detail to 0 or +3. Even at 0 you sill see a lot of "fuzz" around each square. That is all noise. On the other side, lower the detail and you will see that each square look actually sharper.

I'm not sure it's easier and better to add sharpness in post than in camera or if it's easier and better to add sharpness in post than to make it softer if needed. Tiffen has a new software version of their filters and pretty much any NLE will have some sort of soft filter. If it end up too sharp just use one of them as you would a diffusion FX for example. In my opinion, the in camera detail besides more accurate is also easier and faster. Now if you are doing stuff for DVD, I wouldn't see a reason to add sharpness at all, but for theatrical release and specially if using a 35mm adapter, turning detail off or to MIN will probably look out of focus on a big screen, even if it may look better on a studio monitor or in a gray scale test. Just my opinion.

[Sean Adair]

Originally Posted by Michael Maier

I'm not sure it's easier and better to add sharpness in post than in camera or if it's easier and better to add sharpness in post than to make it softer if needed. Tiffen has a new software version of their filters and pretty much any NLE will have some sort of soft filter. If it end up too sharp just use one of them as you would a diffusion FX for example. In my opinion, the in camera detail besides more accurate is also easier and faster. Now if you are doing stuff for DVD, I wouldn't see a reason to add sharpness at all, but for theatrical release and specially if using a 35mm adapter, turning detail off or to MIN will probably look out of focus on a big screen, even if it may look better on a studio monitor or in a gray scale test. Just my opinion.

It's definitely not easier in post! However once ideal parameters are found, it can be batch processed. I'm also very confident that you do not want to process footage to be sharp, followed by processing it it with a soft filter.
Artifacts get generated, and the 2 direction interpolation cannot be as accurate as unprocessed footage.

The camera detail you think is more accurate than what? There is a plethora of different approaches to adding sharpness. I wouldn't recommend just throwing your NLE sharpen filter at it. Edge enhancement, unsharp masking, and very specific algorithms exist.

Most important is dialing in the optimum amount - for the output. If it's being printed to film, this takes place at the final digital stage. You might even be uprezzing to a 2K intermediate for instance - that's where sharpening would ideally be done. Often there will be shot by shot decisions for the amount of sharpening = according to image content as well as the mood or intent of the shot. It's even conceivable you would be masking shots for selective sharpening.

I can tell you from personal experience that when downconvertiong and outputting to SD that there can still be benefits of sharpening at that output resolution and it's more effective than sharpening at high resolution first.
I believe it optimizes edge and detail interpolation more than the resizing.

[Michael Maier]

Yet, any F900 shooters out there will tell you do dial in some detail if going to the big screen. It's been done since the F900 started being used for features. Why is that then?

[Paolo Ciccone]

Originally Posted by Michael Maier

Yet, any F900 shooters out there will tell you do dial in some detail if going to the big screen.

That's interesting. Do you have any more information about this?

[Steve Oakley]

having put HD100 images up on a 16ftX9ft screen, post sharpening is a big no-no. **ANY** and **ALL* compression artifacts will jump off the screen big time. hugely ugly. while the small sharpening looked good on a down converted component SD monitor, it was horrible on the big screen. I had to remove the sharpen filters on all the shots I had used it on. post sharpening is not an option unless its going to SD only.

somewhere between off and normal, find a sweet spot which I think is about -4. the difference between off ( pure mush ) and min to normal is more like 16mm vs 35mm. I could easily pass off hd100 footage at normal thru about -2 settings as 35, go below and its more like 16mm. video cameras are designed to have a little sharpening because at off or min, they are not really showing the resolution they are capable of. its not evil to turn up a little, but I would not normally turn it above the NORM position because then it starts to make some edge artifacting because its over-sharpening.

you need to play a bit with it. for key work, I've shot at several settings and the last stuff I shot was at normal or -1 and it worked fine. progressive imaging is diferrent then interlace where there can and normally is a difference between fields that can make for problems.

[Michael Maier]

Originally Posted by Steve Oakley

having put HD100 images up on a 16ftX9ft screen, post sharpening is a big no-no. **ANY** and **ALL* compression artifacts will jump off the screen big time. hugely ugly. while the small sharpening looked good on a down converted component SD monitor, it was horrible on the big screen. I had to remove the sharpen filters on all the shots I had used it on. post sharpening is not an option unless its going to SD only.

At first I thought you were talking about any sharpening at all, including in camera. But I guess you mean post sharpening, in which case, I agree.

Originally Posted by Steve Oakley

somewhere between off and normal, find a sweet spot which I think is about -4. the difference between off ( pure mush ) and min to normal is more like 16mm vs 35mm. I could easily pass off hd100 footage at normal thru about -2 settings as 35, go below and its more like 16mm. video cameras are designed to have a little sharpening because at off or min, they are not really showing the resolution they are capable of. its not evil to turn up a little, but I would not normally turn it above the NORM position because then it starts to make some edge artifacting because its over-sharpening.

you need to play a bit with it. for key work, I've shot at several settings and the last stuff I shot was at normal or -1 and it worked fine. progressive imaging is diferrent then interlace where there can and normally is a difference between fields that can make for problems.

That's what I think too. -4 is a good compromise. But OFF or even MIN. is too soft. HD needs sharpening by nature.

By the way, I also think that when using a 35mm adapter sharpness could be slightly increased above that, specially for big screen release.

[Michael Maier]

Originally Posted by Paolo Ciccone

That's interesting. Do you have any more information about this?

Just ask anybody who had shot a feature for theatrical release on the old F900 or search for articles maybe American Cinematographer on Hollywood films shot on the F900. You will see that sharpness was always dialed in. Not sure about the newer generation cameras like the Genesis or even the F23. But the F900 needed it. If the F900 needs it the HD100 needs it even more.

[Paolo Ciccone]

Originally Posted by Michael Maier

Just ask anybody who had shot a feature for theatrical release on the old F900...

What I was hoping to get is a description of the phenomenon from a technical point of view. With all due respect, I am not interested in the aesthetic opinions, I'm trying to gather facts about how HD cameras work and I'm a bit surprised that a camera, any camera, would deliver soft images without additional "help" and I was interested in the rationale behind this.

If the F900 needs it the HD100 needs it even more.

That might be but it's not a logical consequence. Testing needs to be done. The cameras use different technology, different sensors etc. From my testing, MIN or -9, -8, with the HD100, works best.

I don't mean to be confrontational about this :), just "digging" for information on a topic that seems to be of interest for many of us.

[Michael Maier]

Originally Posted by Paolo Ciccone

What I was hoping to get is a description of the phenomenon from a technical point of view. With all due respect, I am not interested in the aesthetic opinions, I'm trying to gather facts about how HD cameras work and I'm a bit surprised that a camera, any camera, would deliver soft images without additional "help" and I was interested in the rationale behind this.

I guess I'm the other way around. I barely ever worry about rationale or numbers. If it looks good for my eyes, it's scientific enough for me. One can discuss numbers and technicalities the hole day, but in the end of the day is what looks good for the eye that counts and not the why and how it looks good. All I said is just my opinion.

Originally Posted by Paolo Ciccone

That might be but it's not a logical consequence. Testing needs to be done. The cameras use different technology, different sensors etc. From my testing, MIN or -9, -8, with the HD100, works best.

I don't mean to be confrontational about this :), just "digging" for information on a topic that seems to be of interest for many of us.

Not a problem at all Paolo. It has been a pleasure.

[Ted Ramasola]

Giuseppe Pugliese

I developed my own lens adapter utilizing both 645 and 35mm lenses. I used them on my DVX and HD100.

I made several notes on this endeavor. After a couple of years on it i'm still making discoveries everytime i do some test shots. After getting sharpness, i dealt with falloff, after that grain and so forth.

To make a long story short, i found out that an increased detail on the hd100 when using a lens device with a ground glass, tend to exaggerate the grain structure since the detail cercuit will also try to enhance this. I wont preach a particular setting. Its a matter of taste. My suggestion is you view this on a monitor while changing the detail settings. shoot somthing black. You will notice dramatic changes in amount of noise or "enhanced grain" with the detail changes you make. In my case, I stayed at +1 and normal. Grain is there a wee bit, but then again, its a matter of taste.
The choice is yours.


Ted

[Paolo Ciccone]

Originally Posted by Michael Maier

I guess I'm the other way around. I barely ever worry about rationale or numbers. If it looks good for my eyes, it's scientific enough for me.

That's the same for me, our eyes are the final judges but what I want to know is what the camera is recording. Because it seems weird that with detail set to OFF the camera generates blurry images. It doesn't make sense to me. Once the subject is in focus and the lens is installed properly, the camera should give a perfectly sharp image. I'd like to hear the definite answer, possibly from the manufacturer on what's going on. Is the sensor able to capture an image in full sharpness or it needs the help of an edge sharpener...

[Ted Ramasola]

I second that paolo.

It is blurry to my eyes when detail is off.

[Eric Ramahatra]

i always wanted to start a topic about my HD200 sharpening but i thought i always had bad settings:
- i'm using TC3
- back focus was done with a chart
- stock lens used

Sharpening set to Norm gives a not so sharp image and set to min, it seems like out of focus !
why do my Hd200 does not give those "HD look" at all?!
It is much more like a DV scaled to HD !
i will link some frame grab but i'm very disappointed sometimes.

Just one time i used some testing settings, everything was nice! But didn't remember how i did it!
I have to shot a short within the next days so any advice will be appreciated.

[Claude Mangold]

OK Giuseppe I've done a rough edit of the lens test, but the QT H264 file is 137 MB for 3m46s and I don't know how to upload a file this size.
Can anyone help ?

[Sean Adair]

Michael & Steve - thanks for the added explanation. It does make sense that sharpening post-compression would cause issues with "enhancing" artifacts - which are even more of an issue with HDV than they'd be with the F900. I'm still learning the situations where artifacts start raising their ugly head. It was strange to discover they could be more of an issue with "simple" scenes like a uniform dark area, yet interpolated scenes containing more detail and motion with better results. I can understand that sharpening would not be good for the shots with artifacts. I guess you are saying that post-sharpening (taking place in a non-compressed environment from converted HDV footage) will actually bring out compression artifacts in footage which appeared clean before.
Steve - I'm curious about your tests on the 16x9 screen. Were you looking at actual film transfers from the JVC, or digital projection - if the latter, is it from a deck or streaming from a video card?
I've just started to look more critically at my footage going to a roughly 4x7 foot screen from the BR-HD50 to a 720p projector.
In short, some things look fantastic, others terrible...