8bit vs 16bit - Why most PROs get Bit Depth WRONG?
ฝัง
- เผยแพร่เมื่อ 21 ก.ค. 2019
- Bit Depth SIMPLIFIED In-Depth and 2 Biggest MYTHS Debunked in this Photoshop Tutorial. Witness Theory vs Reality of Bit Depth to see what almost Everyone fails to Understand. The only video you will ever need to watch about Bit Depth...
► Skip to ANY section ⏰ TIME STAMPS:
0:10 Introducing Bit Depth/Color Depth/Pixel Depth
0:26 Concept of Bit Depth
0:47 Bit Depth definition Visually explained
1:18 Understanding Bits per Channel(BPC) in Photoshop
1:54 What exactly are 'Bits'?
2:31 Why 8 bits have maximum value of 255
3:06 Different Bit Depth have different tone/RGB values
3:38 Photoshop Bit Depth comparison 1bit B&W/8bit grayscale/8bit RGB/16bit RGB
5:17 1st Reason for no visible Difference between 8bit & 16bit RGB
6:16 2nd Reason for no visible Difference between 8bit & 16bit RGB
7:25 Spot the Color difference challenge in Photoshop + Human Visibility explained
8:11 Why Photoshop shows 255 value for both 8 & 16 bits
8:28 Which is Better for retouching - 8bit or16 bit
8:59 Myth 1 debunked (To fully understand, watch the full video)
10:06 Myth 2 debunked (To fully understand, watch the full video)
------
📸 Start Retouching 10x FASTER with Post Pro Photoshop Panel:
www.postprotools.com/
-----
#photoshop #bitdepth #16bits - แนวปฏิบัติและการใช้ชีวิต
Im blown away by your channel! All your videos strike the right balance between technical explanations that the digital platform needs and the art theory that would follow any endeavor to make content in Photoshop. Great video💯
Part of the reason you don't see the gradation is the DACs that converts the digital values to analog is 8 bit DACs for each of the color channels just as it were for an Amiga 1200 and Amiga 4000 back in 1991. This means the gradation of each of the color channels like greyscale. The more shades per channel is like the shades of the 8bpp grey scale vs 16bpp grayscale but now apply that to each of the three color channels. Most TVs and monitors are using 8bit DACs (DAC = Digital to Analog Converters). Some newer TVs with the touted high dynamic range, in order to do it not only widens the gamut but also widens the range from darkest levels of the color channel to the brightest level of the color channel. Then you have that divided by 1024 shades of the component colors of R, G, and B. Hence, 10 Bit DACs per channel being used. The most extreme end of TVs currently that I have heard of uses 12 bits per channel. 16 bits can be used and you would notice it more IF you look solely in the gradation of that channel and the smooth transitions along it just as you compare between 8bit greyscale, 12 bit grey scale and 16 bit greyscale. Human eyes will notice most on the green channel and the red channel. Human eyes are weakest on the blue channel. This is due to the area of color spectrum our eyes are most sensitive. As creatures of nature, our eyes evolved to perceive the shades of green more and that's due to evolving from our earliest ancestors in woods, forests, and jungles. This is the natural spectrum our eyes are sensitive to. Red and Blue are near the ends of our visual spectrum. Hardware will generally mixes the channels blindly within a minimum and maximum output level in each channel in steps. Be they, 256 steps, 1024 steps, 4096 steps, or 65536 steps. If the DACs are 8 Bit DACs for each component, the limits are 16,777,216 colors. You could have had 9 bits of each of the RGB color channels and you wouldn't see any more than 256 shades of red, green, or blue. This is due to HW. Your eyes could perceive more than 256 shades of red or green or blue. However, your display won't output more than the DACs bit resolution. If you have 10 bit DACs, it's limited to 1024 shades of red or green or blue. The human eye could conceivably see as many shades of each of those mixes as it can perceptively discern with grey scale based on steps of brightness or intensity from lowest output level to highest output level. However, once you mix colors into pixels it becomes less and less distinguishable. At some point, you reach a point of practical use. There is a diminishing perceptable return. Going from 16bpp to 24bpp was less perceptible than it was to go from 8bpp to 16bpp. Anything more than 24bpp (RGB888) is more wasting bit size in most end use practical purposes especially in something like games. Heck, even game graphics were sometimes still 16bpp (65536 colors) to save space and look good. You'd be hardpressed to tell a 16bpp picture of ultra4K resolution (3840x2160) from one that is 24bpp. Visually, won't make much difference but saves you 1/3 in data space. Animate it and have lots of frames at that resolution, you can save space and have quality. Further techniques of image data compression and you'll save more space. With good computing power, you can decompress in real time and still have realtime video yet saved space. 16 BPC (Bits per Channel) is probably overkill and won't be needed on something animated. If something was static, maybe if you look at it long enough, you might see a slight shade difference but when you are above 4096 shades of red, or green, or blue, it would be hard.
That was a great video! I didn't learn anything new, but you covered all the bases in a way anyone can understand. Well done, sir!
The difference of a higher internal/working bit depth is noticeable on an 8-bit screen on clean digital gradients, such as the ending screen on your video, especially grey where all the color channels switch nearly simultaneously. The final 8-bit output will approximate the gradient better through dithering (error diffusion). Dithering noise is often misunderstood as only masking quantization noise, but it actually prevents it if it is added to a more accurate value. JPEG/MPEG has around 7-bit depth and will usually remove high frequency dithering and create banding again.
OMG...i can understand everything you've just teach here , thanks!
Love the explanation. Good thing I watched from the beginning because it can get very complicated very fast.
Thank You For Such An Informative Video. I Cleared My Many Misconceptions About Bit. Thank You So Much!
EXCELLENT . I love your perfectly clear explanation ! TNX
Superbly. Explained in detail.Got concept Clear by this video only . thank you so much sir
latterly I saw a very few video on TH-cam that completely worth it to watch, explain in very simple and nice way, you did it man, Great Respect ❤❤❤❤❣
This video couldn't be any more clear! This was awesome!
@StyleMyPic Well explained, any mention of monitor limitation display?
Great video. Only one minor issue to point out. At 5:47 you have 2^(15+1). This should be (2^15) + 1. As written, it's the same as 2^16. It's hard to write here in a comment with lack of exponents, but the +1 should be removed from the exponent and just added to the 2^15.
Very minor issue. Thanks, though, this is a great video. Just subscribed. I really want to check out this Pro Workflow X you have going on there.
Matt Williams Thanks for pointing it out.
Yes i noticed that as well. you beat me to it by a year
yea, i was like, you know 2^(15+1) is the same as 2^16 or 2^15 * 2^1
@@Stylemypic very good video anyway
Such a great explanation, Thank you very much my friend!!!!
Your video should get more than 1 million views! Outstanding !TY !
Very underrated channel. Subbed man! thanksssss
This was insanely helpful, thank you so much!
Great video! Can you please do a video explaining bit depth in after effects which has the option for 32?
Brilliant. Have had a few problems converting black & white photos into colour and then working on the images has produced compression problems in Jpg. SO am now trying same with TIFF files and doing photoshop in 16bit from 48 bit scan of black and white negs... This video is GREAT.
This is a lot of info for me.. I just wanted to know which looks better and does it affect rendering times much? Thanks :)
You're the best. ACCURATE, concise, and intelligible.
Fabulous video, my friend. You know how to explain things the RIGHT way.
You've got my sub. :-)
This is just amazing explanation! Thank you very much!
Glad it was helpful!
You should have mentioned that the size of the color space you are working in, could demand a bit depth of 16bits. E.g. if you choose Prophoto RGB, you need 16bit as the distances between the three corners spanning the colorspace are so huge, 256 steps often are not enough for smooth gradients.
You are absolutely right. This is briefly mentioned at 8:36 of this video and covered in full details in the previous color space video: th-cam.com/video/ujE5bBXIKJE/w-d-xo.html
You're really good at Explaining, Great job mate!✨
Thank for this very informative and engaging video. I'll consider adding 8 bit image files into the mix after editing for online only use esp on social media.
Great explanation, Sir. Thank you!
great vid. is there any difference in sharpening quality between 8bit vs 16bit? I personally don't see any difference. thanks
I am less than 1:00 in and this is already one of the best tutorial videos out there explaining this really important topic, thank you
Glad it was helpful!
the main problem with 8 bit colours arises when we look at dark colours and/or black backgrounds. Tom Scott has a really good video on this but basically, if we do a gradient from red to black, then we will start to see the limitations, and we will begin to see the individual lines of each colour change. especially when done over smaller pixel distances.
Nicely explained ❤
This is EXCELLENT content. Thank you for spending your valuable time to help educate people.
At last, found the exact reason ! thanks bro!
8 bit files fall apart very fast with a lot of adjustments. With lots of adjustments, even 16-bit images run into artifacts and noticeable degrading MUCH faster than most realize (especially if you make critical enlargements). 32-bit RAW Layers, and 32 bit rasterized images are a TOTALLY different beast altogether! The ultimate quality workflow.
Absolutely true. Unfortunately PS has many important adjustments and tools disabled in 32bit.
@@Stylemypic like?
Thanks. some questions:
If im editing for web, should the end image be always 8 bit (srgb jpg)? And do I have to manually convert it to 8 bit? And can there be any perceivable changes in images after I limit it to 8 bit, like colors changing?
A web image till now is compatible only with 8 bit images, so you have to convert your 16bit image to 8 bit. Color gradation is affected only if you compress the 8 bit image to a medium to low quality jpeg file. And color change has nothing to do with bit depth, watch this video to see how and why color is affected...
Color Profiles (Assign Profile vs Convert to Profile) Video: th-cam.com/video/nB7aEiKQw6E/w-d-xo.html
My God!!!! I became your lover in the first visit!❤❤❤❤ How simply could you define the complex technicalities with beau.....tiful presentation!!! I cant stop myself to SUBSCRIBE your channel. Hope, the rest videos will definitely be superb!!! Bravo, Dear!!
What a great exclamation I hope this video gets a ton of use from its exposure on peta pixel
Wow. That was fantastic. Thanks man.
Your explanation is so great!!! You are a real teacher!!!
can you please elaborate a bit on why there's no need to change the mode for raw photos?
Excellent! Subscribed.
Question: If you are a digital artist using a tablet where the pen has let's say around 8000 levels of pressure sensitivity. Would working in 8 bit just restrict that to 256 levels of pressure? Perhaps a stupid question. But let's say I'm painting something in Black and white there are 256 possible shades, right? In other words nothing to represent those 8000 levels of pressure? Or am I getting it wrong?
This is why I believe digital artists and creators should have some knowledge on how data works! Else everyone thinks more bit depth and large file sizes means better quality. Thanks for the great explanation.
this has been very helpful, thank you
high bit depth or sample rate or whatever usually is imperceptible but matters more if you do things like distortion, stretching, amplification, anything that stretches out those things. slow down an audio file significantly and suddenly the difference between a flac and a 192 mp3 should be much more noticeable, same for pixel depth, or resolution; make certain kinds of effects and they become more important, like if you substantially brighten+ raise contrast suddenly 000000 vs 010000 becomes absolutely massive
so essentially when it comes to a final file, lower quality is perfectly fine, but for something someone wants to edit, suddenly they might matter- though of course, in some cases that crunchiness might be desirable to someone
great neat explanation dude 😍
Might make sense to go up to 12 or 14 bits per channel point and click, and for really gaudy levels of colour mapping/correction/grading maybe 16, unlikely ever more than that out side of special cases such as scientific data, for example single channel images from space telescopes. If you do want severe color mapping then you might need really low noise images, for this take the photos in really cold conditions.
wow, great explanation, thank you
Very informative!
What is the relation between display bit depth and picture bit depth. On a 6 bit display how 8 bit would be displayed?
Very well explained, but in some cases, if you raw has a very smooth transition and work on 8 bit you gonna see banding very soon, without doing extreme retouching, some example are blue skies without clouds and very smooth transition from brighter to darker tones, and studio work with very smooth light transitions on the background, also if you are printing those same images with very smooth transitions, printing in 8bit can again show banding depending on the ink and the paper, but it's also worth to mention that printing in 16bit with a printer that does not support 16bit, will not give you any benefit.
perfect explanation, thank you!
Great videos! Please, more, because there are hardly any videos that deal with more complex topics. Maybe a little less information per video for a little more time and examples and more in depth.
not less info per video, keep it like this. I was picking up what you were putting down and alot of other people were too.
Thank you for going in depth with the technicalities of everything! a m a z i n. g
This is an amazing video. Thank you.
Amazing video!
Excellent. Thank you.
great explanation!
Does all this applies for video too?
Because the it's just 1 frame for a photo and 24 frame for a video.
Wow you're such a good teacher! Thank you ✨
WOW.. good job man!!!
Excellent video. Bravo.
Amazing. Thank you
Nice to understand
Any possibility that we can recreate the eye and implant a chip in head that makes ous support (see) more then 10 million colors, so we can get the full experiance
Great Explanation..
amazing...thanks for sharing
Really helpful.
Nice and awesome, thanks you
Thank you so much.
Wow, great presentation.
Thank you so much for your valuable info
Glad you found it useful.
Superb.......
Thank you
So what should we do?? Work on 16 bit and export in 8 bit?? If yes then how can we export in 8 bit??
Fantastic teacher!!
What have you been using- 8bit or 16bit?
well 32bit and 16bit
ive been using 8bit to retouch my landscape photos for soo long i cant believe i just realized i need to put it in 16 lol smh
Which algorithm is used to convert picture from 16-bit to 8-bit? Just simple rounding?
I'd have to dig deeper, but I'm going to assume it's a simple truncation. Usually when you reduce precision, you simply discard the extra bits. You really can do the same thing here. The first 8-bits are still a perfectly valid sample. You're just loosing the next 8-bits that made it more precise.
For instance, take his red gradient for example. In 16-bit, Red 255 to Red 254 had 129 shades, but if you converted that down to 8-bit, 128 of those shades would now be Red 255, with the last being 254, because you truncated the additional bits that gave the precision for those 128 extra shades.
More informative than my Photoshop lessons at university
brillant video !!!
thanks a lot
but i can see the difference between the two reds. Not every shade, but still loosing colour difference.
wow nice thankyou :D
that was amazing
i just want to know if I just export my png files in 16 or 8 bit :(
Edited: Okey I just finished the video. Thanks :)
wow...great content
FAAAAANTASTIC!!!!!! MUCHAS GRACIAS!!!!!!!
@StyleMyPic. Relates well to photo editing.. but not for digital painters.
Open a fresh canvas with black background, "airbrush" a grey over the top and watch everything you just described about imperceptibility become redundant.
How am I supposed to paint anything, when a soft airbrush looks like a topography map from 1800's? No seriously, that's a genuine question
correction, your "8 bit" at 0:28 is really 4 bit, same at 3:40
Besides that you won't see a difference between 8bit and higher content just if you have standard 8bit screen.
Other story would be working on calibrated 10 bit oled screen.
thanks
awesome :O
Damn dude ! You are an excellent teacher.
Nice walkthrough into bitdepth. The conclusion, however, is wrong. It's true that there won't be significant difference between final images in 8-bit and 16-bit, but having extra bitdepth when editing is something you might definitely want.
For instance, create a linear gradient from black to 20% gray in 8 bit and check the result: dithering sort of hides it, but you are going to see a lot of banding.
Any editing you do on the shadows of your image is likely to be affected if you don't have enough bitdepth.
Just multiply an identity 8-bit image by a low value and you'll see how much color precision is eaten out by 8-bit!
super..
Legend says, *_"Great images are not taken. They are created."_*
You don't take a photograph, you make it. - Ansel Adams
Dude's good.
To see a diff between 8 and 16 bit images you need a 16 bit [or more] monitor. And that is true 16 bits. AFAIK there are none. Monitors are 8 bit plus some fuzzy algorithms to get perceived 10 bit.
whats going on with your audio
Thanks 👍 get it that 16 bit has more colour in depth and more necessary when editing a photo that has large variety of colour than 8 bit but still I don't understand why it need to be 8 bit to use filter gallery in photoshop?
Most of Photoshop Filters have limitations in 16 bits and so have always been available only in 8bits.