====== Basic Photoshop Workflow ====== - (optional) * method 1: RAW development to image with simple color correction (no sharp, no contrast process), - then export as original image * method 2: directly export 16 bit tiff as original image - original image (optional in 16-bit tiff for more color pulling and pushing) - Clone, Fix, Retouch layers - Color Correction layers - Color Grading layers - duplicate result layers for sharpen and blur ref: https://fstoppers.com/education/speed-your-retouching-six-easy-tips-75929 ===== RAW vs 16-bit tiff ===== * watch this video first: http://tv.adobe.com/watch/the-complete-picture-with-julieanne-kost/the-advantages-of-the-dng-file-format/ * http://photo.net/learn/raw/ * ref: http://forum.luminous-landscape.com/index.php?topic=57894.msg467946#msg467946 * "A raw file contains an image just as does a TIFF. The raw image contains data in a mosiac format, whereas the TIFF uses a separate channel for each of the RGB components. Shown below is a straight dump of a raw image as shown by Rawnalize--it looks to me like an image. However, it needs to be demosiaced by a raw converter just as the three separate channels of a TIFF need to be combined in Photoshop." * also for RAW file * it contains lens and camera info for database lookup lens correction and optional to tune any camera settings related to camera manufacture setting * smaller size because each pixel is actually a single 12-16bit "color" rather than three 16bit colors, thus is faster to write to disk * RAW quality is actually higher. TIFF export requires interpolation of color data in camera, which can cause artifacts in post. * RAW has more useful data when in high contrast scenes. RAW is in sensor read data linearity while TIFF is linear in the display space. For some situations a 12bit RAW will have better data (not necessarily more) than a 16bit/color TIFF made in the same camera and scene. * RAW guarantees data integrity. While you can overwrite a TIFF you can't do the same with RAW * ref: http://www.tomsguide.com/forum/56235-5-tiff ref: * http://www.thephoblographer.com/2011/09/26/are-tiffs-and-raws-really-the-same-thing/ ===== Rotoscope - Vector Masking Workflow ===== - Use Pen tool to draw the outline of the object - for "out of focus" or "motion blur" area, the line should be either following the inner blur radius or outer blur radius (but overall should be consistent) - generate a bitmap/pixel version mask from it - use blur brush with controlled strength to smooth out the transition area, - for sharp edge, the transition area of mask should be around 0.5 to 1 pixel - for blurred transition area of mask, should be manual painted to a natural level - then, use level edit to controlling the shrink in and out of overall transition area, (basically shifting the black and white for the grey area; so you can use dodge and burn technique as well) ===== Retouch Part ===== * ref: * https://fstoppers.com/originals/frequency-separation-rehab-42129 * https://digitalphotoshopretouching.com/blog/my-beauty-retouching-workflow ====== Work Reference ====== * Portrait * https://www.instagram.com/risa.yu/ ====== Important Tips ====== * Whenever there is a gradient in the background in your image, make sure you are working in 16-bit color depth to avoid banding of colors. ====== Words about digital photo editing and creation ====== * if you can't get a wide-angle lens like 11mm, \\ you can always stitch a 180 degree panorama with several side-by-side shots taken with normal-angle lens like 28mm * if you can't get a 18MegaPixel camera \\ you can always put a grid shots of 3MegaPixel photos together. * if you can't get a dynamic range from your sensor, \\ you can always shoot mutiple shots with different exposure and stitch together. * if you can't get a large aperture lens to create a background-blurry portrait shot, \\ you can always blur the background manually in photo editing softwares. Because - Wide angles lens are expensive - large aperture lens are expensive - MegaPixel sensor camera used to be expensive (my 1st camera is 3MP) - High sensitivity sensor camera are expensive We achieve the same thing digitally with free photo eiditing software, [[https://krita.org/en/|Krita]] or [[http://www.gimp.org/|GIMP]]. ====== Digital Only stuff ====== * Moire pattern effect and AA (anti-alias) process and Low Pass filter * http://www.wfu.edu/~matthews/misc/DigPhotog/alias/ * http://mansurovs.com/what-is-low-pass-filter ====== HDR Merge ====== * Tone mapping tips: http://www.lightstalking.com/5-important-tone-mapping-tips * Exposure Value and Exposure stops: http://www.reedbeta.com/blog/2014/06/04/artist-friendly-hdr-with-exposure-values/ * pixel brightness and exposure stops: http://www.scratchapixel.com/lessons/3d-advanced-lessons/things-to-know-about-the-cg-lighting-pipeline/0-18-stop-and-exposure-explained/ ====== Color Grading ====== * video tut: cn TangWan: http://www.chinahunpai.com/play1435.html * style 1 (white/cyan/contrast): ([[http://sosli.poco.cn|ref]]) * cleanup and sharpen * lighten the face before clipping (drag highlight in level) * cyan the shadow * skin soften by soft light blend of BW filter (high contrast red filter) * saturate facial color * style 2 (yellow/cyan/low contrat) ([[http://www.swcool.com/html/article/sanwen/333.html|ref]]) * skin: (red/yellow) - cyan - magenta - yellow - black + saturate + light * shadow: (blue/black) -cyan + magenta + yellow + black * neutral: + yellow * 2nd pass color tuning * lower highlight curve to reduce contrast * lighten highlight in level to brighten, middle point to balance color * channel mix to boost color channel a bit more * color contrast: white/yellow - cyan - yellow - black; ====== Common Technique ====== * Green Leaf to Orange Yellow Leaf * yellow (Abs -70% cyan=+70% red (affect light green part), -50% yellow (-saturation)) * green (Abs -75% cyan (affect deep green part), +10% meganta (+aging+saturation), +50% yellow (+saturation)) * black (Abs, -cyan (dark->red), -meganta(dark->green)) * color balance: shadow+mid (+megenta); hightlight (+yellow) * soft light look * 40% color-filled top layer in screen mode * color replacement * paint layer with hue mode * color style * white light cyan: pure, light, clean * orange warm: energy, warm, home * cyan white cold: modern, stylish, fashion * control of Hightlight, Middle Tone, Shadow Area * selection method: * load from RGB channel (as white is more selected, dark is not selected) * color range (white, grey, black) * calculation (math way for selection making) * hightlight: grey in screen; * shadow: inverted grey in screen; * middle tone: grey in exlusion * steps of a color-grading approach - decide your color style - make a duplicate layer - environment - apply Selective Color and Color Balance and Curve to get the rough color and exposure - make a duplicate layer - shadow - apply Selective Color to tune the shadow to the look - get Shadow Masking to that layer - make a duplicate layer - highlight - apply Selective Color to tune the shadow to the look - get Highlight Masking to that layer - make a duplicate layer - skin - apply selective color to tune the skin to the look - paint masking to show the skin area. - apply final grading and color tunes - if any parts need to be specifically tuned, duplicate and tune it seperately ====== Skin Perfection====== * skin perfection introduction: [[http://jingyan.baidu.com/album/f3ad7d0fc77a2d09c3345b34.html|link]] * method - [[http://jingyan.baidu.com/album/77b8dc7f2707456174eab6c0.html|channel curve mask]] Of course, before any below method, please use heal brush or patch tool to fix big holes of skins **Overlay + Invert + Highpass Filter (blur) + Gaussian Blur Filter (sharp) Method** - duplicate layer and apply overlay blend mode - invert layer - add highpass filter to blur it like 10 point setting - then Gaussian blur filter to sharp it a bit - mask **Highpass filter (3x,18), Gaussian Blur (1x, 6), Invert, Linear Light mode (corner blend option)** - duplicate layer - highpass filter (3x, 18) - Gaussian blur filter (1x, 6) - Invert - linear light mode (corner blend option) - mask **Median Filter (10) ** - duplicate layer - Noise: Median Filter (10) - mask **Surface Blur Filter (5,15)** - duplicate layer - Blur: Surface Blur Filter (5,15) - mask **Highpass Filter (10), De-saturate, Invert, Softlight/Hardlight mode ** - duplicate layer - highpass filter (10) to get details, - de-saturate to get detail line information only - Invert to get smooth area information only - softlight/hardlight mode to blend into smooth area - mask Summary of all above method and theories behind it: * ref: https://fstoppers.com/post-production/ultimate-guide-frequency-separation-technique-8699 **and more scientific accurate way of detail (high frequency image data) and smooth area (low frequency image data)seperation and process** - prepare layers: - high frequency layer: fill with 50% gray - low frequency layer: duplicated image with Gaussian Blur all fine detail. - write out into high frequency layer - select high frequency layer, choose Apply Image (8 bit, Source: LowFrequencyLayer, RGB, Subtract, scale 2, offset 128) - linear light mode - soften and even out color and tone transitions on the Low Frequency layer ===== Oil Spot removal ===== * clone tool * **steps**: - use 30%~50% clone the edge around it to surround it slowly to eat it out, - but the things are, when tons of little oil spots, there is no clean area to clone with - force the brush will cause bad brightness change overall, then you need to lower the opacity and flow of brush to ease the blends of clone * heal brush * **step** - use small brush and heal the dots out, - best of heal is auto blend the brightness of the spot, but sometimes it left with a patch outline ====== Use before Shots ====== * water bottle spread mirror for low angle sun shot ====== additional ref ====== * coloring: http://www.3lian.com/edu/photoshop/pstiaose/