[Bf-cycles] Fw: Ideas

Thu Oct 27 20:45:20 CEST 2011

Hi Davis,

I posted my reply (below) to Julian several hours ago, didn't notice it was only sent to his personal e-mail address, instead of including the bf-cycles at blender.org address.  Must have been doing that with all my replies, hope people don't think I have not been replying to responses.

Thank you,

David

----- Forwarded Message -----
From: David <db4tech at yahoo.co.uk>
To: Julian Hhhhhh <julian.hhhhhh at googlemail.com>
Sent: Thursday, 27 October 2011, 15:45
Subject: Re: [Bf-cycles] Ideas

Hi Julian,

Thank you for your reply.

As a photographer I fully agree that it should be an accurate emulation of a physical lens, I possibly worded my reply badly, I meant to keep it simple for end users without need for them to carry out calculations, so (as with a camera) they choose focal length, focal point and f/ stop, all calculations would be are handled invisibly for them.

Sorry if I gave the wrong impression.

David

________________________________
From: Julian Hhhhhh <julian.hhhhhh at googlemail.com>
To: David <db4tech at yahoo.co.uk>; bf-cycles at blender.org
Sent: Thursday, 27 October 2011, 14:27
Subject: Re: [Bf-cycles] Ideas

Hi David,

just a short comment on the f-numer - focal length - dof discussion:

2011/10/27 David <db4tech at yahoo.co.uk>

Hi Davis,
>
>
>Thank you for your reply.
>
>
>I agree, it is also worth considering that the f/ stop should not change due to focal length changes (well not as far as the end user is concerned or within the Blender interface), unless simulating lower grade camera lenses which alter their f/ stop dependent on focal length, I don't think that level of simulation is required.

I think it would be best to emulate the actual, physical behaviour of a lens, which is that the f-number increases with the focal length. This happens due to the fact that the f-number is calculated from both the focal length and the effective entrance pupil diameter.

The difference between "lower end" camera lenses and "higher end" lenses is only that some of the latter ones automatically open the aperture when zooming in, which helps to shoot at a constant f-number but artificially limits the aperture size when zoomed out (which is no problem because most of them would not be of high optical quality at the f-number you then can't reach any more).

You are correct, the formula would require both aperture and focal length to be calculated but also focal point (lens distance to target object), to produce expected real world results.  From tests I carried out last night, it seems as if this is being replicated to a small degree, the depth of field does seem to change slightly dependent on focal length, just not as much as expected.

This is a relevant effect that should be noticeable. There is even an influence of the focus distance on the focal length (the given f.l. of a lens only applies when focusing to infinite distance).

Btw, there are tools like these to calculate the correlation between f.l. / dof: http://www.dofmaster.com/dofjs.html

Thank you,
>
>
>David
>
>
>
>
>________________________________
>From: Davis Sorenson <davis.sorenson at gmail.com>
>To: bf-cycles at blender.org
>Sent: Thursday, 27 October 2011, 12:52
>Subject: Re: [Bf-cycles] Ideas
>
>
>
>Hi David,
>It's worth noting that the F-number is a value dependent on both the aperture of the lens and the focal length, so as you zoom the
 lens, the f-number becomes higher. (F number = focal length / aperture). So it's actually a ratio. That said, it would be nice to have the option to choose to use f-stops, but the value would have to update as you zoomed in/out.
>
>Davis
>
>
>On Thu, Oct 27, 2011 at 2:32 PM, Daniel Salazar - 3Developer.com <zanqdo at gmail.com> wrote:
>
>Simpler solution to close by mistake or crash is not saving a flat png
>>but a package with the samples data like luxrender does, so you can
>>simply resume your render like nothing happened. This idea is already
>>in the todo :)
>>
>>Daniel Salazar
>>3Developer.com
>>
>>
>>
>>
>>On Wed, Oct 26, 2011 at 10:04 PM, David <db4tech at yahoo.co.uk> wrote:
>>> Hi development team Brecht & Thomas (sorry don't know everybody's names),
>>> Idea 1:
>>> Since render times can be quite long to clean up indirectly lit areas...
>>> How do you feel about the idea of adding the option of automatic progressive
>>> render saves?
>>> A user can choose between either a time value, e.g. once an hour,
>>> or after x amount of sample passes, e.g every thousand passes,
>>> the ongoing render is automatically saved and rendering continues.
>>> Then if there is computer crash, or power-cut, while the computer is left
>>> rendering, many hours rendering will not be lost.
>>> The automatically saved image file name, could be same as the .blend file
>>> name, plus date & time?
>>> That way, each image name will be unique, without the chance of images being
>>> overwritten when a user starts another render with the same .blend file.
>>>
>>>
>>> Idea 2:
>>> Future render optimizations...
>>> Is it possible to focus sample passes based upon light intensity within a
>>> scene,
>>> darker areas receiving more sample passes than lighter areas (which already
>>> clean up very quickly)?
>>> If this isn't already happening with Cycles, would it provide a significant
>>> speed boost?
>>>
>>> Idea 3:
>>> While rendering, trying to close Blender shows a popup warning window in the
>>> centre of the screen,
>>> "Blender is currently rendering, are you sure you want to close the
>>> program?" Yes/No
>>> I did this by mistake after just over 10 hours of rendering... :(
>>> While closing other open program windows (which were in front of Blender),
>>> I accidentally closed Blender before I had saved the ongoing render.
>>>
>>>
>>> Idea 4:
>>> Aperture size scale...
>>> 'f/ stop' as an option to 'Aperture size', so users can use which they feel
>>> most comfortable with?
>>> (Shallow depth of field) f/0.9, f/1.4, f/2, f/2.8, f/3.2, f/4, f/5.6, f/6.3,
>>> f/7.1, f/8, f/11, f/16, f/22 and f/32 (Depth depth of field).
>>> ***
>>>
>>> I am presently doing a BA (Hons) and using Blender & Cycles for my Major A,
>>> do you mind if I submit responses as part of my degree work?  Hoping to post
>>> my finished render results online.
>>>
>>> Thank you greatly for your time and fantastic work producing Cycles and
>>> Blender!
>>> David Black
>>>
>>> _______________________________________________
>>> Bf-cycles mailing list
>>> Bf-cycles at blender.org
>>> http://lists.blender.org/mailman/listinfo/bf-cycles
>>>
>>>
>>_______________________________________________
>>Bf-cycles mailing list
>>Bf-cycles at blender.org
>>http://lists.blender.org/mailman/listinfo/bf-cycles
>>
>
>_______________________________________________
>Bf-cycles mailing list
>Bf-cycles at blender.org
>http://lists.blender.org/mailman/listinfo/bf-cycles
>
>
>
>_______________________________________________
>Bf-cycles mailing list
>Bf-cycles at blender.org
>http://lists.blender.org/mailman/listinfo/bf-cycles
>
>

-- 
Julian Herzog
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/bf-cycles/attachments/20111027/2f4695b9/attachment.htm 