Nice explanation, but you should have mentioned that the bit counts are per Color channel, R=red, B=blue, G=green. So the actual bit count is (almost) always three times the count you mentioned. So your 8 bits are actually 24 bits. 😊
This was super useful, I never knew what that meant, I assumed it was something to do with the information captured but I hadn’t realised it was to do with the amount of colour’s. Thank you!
He is wrong. Bit depth pertains to the precision of the ADC, and even then, most of it is false precision because very few cameras capture enough dynamic range to take advantage of it. A higher bit depth itself does not produce more colours but rather, it means the tone curve has more steps (greater precision), making banding/contouring artifacts are less visible. The comparison at the beginning showing different colour grades is misleading and has nothing to do bit depth.
My experience as video engineer is that 10 bit typically allows better compression due to less bit banding. The extra colors over 8-bit aren't noticeable for the average person, but the extra colors are useful to provide smoother transitions, thus adding less compression noise.
Ah. This explains why when I create a LUT based on a 14 bit photo in Lightroom and convert the preset to a .cube file, I can’t usually crank up the new LUT to more than 50% on my 8 bit footage without the prior stated banding.
I'm pretty sure that the difference in the 2 photos is just compression levels. On non-compressed photo, you CAN have super smooth color transition with only 8bit.
@frstylol Many thanks for replying to my comment and contextualising the pictures. Since not all people read comments, it would be good for the OP to add a "illustration only" or anything like that on future videos, to avoid misleading anybody...
@@enricesThese guys put more effort into telling you what and what not to buy than actually teaching you how to master your craft, most people typically don’t look that closely at the screen to notice anyway
You can't really see the bit banding, the one shown was exaggerated to illustrate the issue. The main advantage is to reduce the noise on the image, making compression more efficient.
@@Shockload I am a video engineer myself and never noticed any bit banding with a naked eye at 8 bits, even in dark scenes. It does, however, causes issue with video compression, so 10 bit is typically preferred to have better compression with less losses.
Whoa! This is a huge new thing to learn! I had no idea they were that different. I thought bits just referred to how much info the camera gathered per photo or video or something like that.
I think speaking of more colours is confusing. It's better to say the transition steps between the colours are finer and thus allow for more shifting without loosing the fine differences between those steps.
Light is analogue and digital is discreet so the additional bits allows tha abiltiy to chop up the visible light spectrum into more discreet amounts. 8 bits is 2^8 and 10 bits 2^10.
Even most phone cameras see 10 bit but they encode 8 bit for some reason. Most likely because you don't have the same features as the pro versions. Just like, although it has the power to shoot 4k 60 FPS, they added this to the pro versions.
10bit only makes sense in processing or if you're going to display it on an HDR screen. Otherwise, you're displaying 10bit colors on 8bit display and the banding will still show up. It's the same for audio: Use 24 or 32bit for processing and 16bit for playback.
Banding happens in any bit depth, a lot due to compression. To "fix" this you should add some gaussian noise to the areas that need it. This will solve it even in 8bit
Quick question to all the photo nerds: Am I correct that this is 8-bit for per R, G or B value? A byte can store maximum of 2^8 or 256 different configurations, so the total amount of different combinations would add up to 256^3 or 16,777,216 different colors. Also, this means that Hexadecimal color codes are also in 8-bit colors, right?
Yes, 8 bit per channel, represented by three hex numbers. Having a 10 bit source is only advantages when you are changing brightness and contrast of the image in post. There is 0% difference in quality of 8 and 10 bit if you are using the photo unedited.. 8 bit gradient per color is already better then your eyes can detect...
8 bit video captures the same range of colours as 10 bit video but there are more gradients steps in-between and the difference is not nearly as bad as shown in the video, but can be kinda shitty in very dark scenes. 8 and 10 might not be so far away but given that each additional bit doubles the amount of steps for a colour channel per pixel, it makes sense. having 256*256*256 color steps is less than having 512*512*512 color steps
I have a ¥23,000 10bit 2K 165hz hdr monitor. The color is noticeable compared to my Huion Kamvas Pro 19 8bit when connected via hdmi. My gpu nor my motherboard doesn’t have type c alt mode (type c displayport) so I ordered a “Wacom Link Plus”. Hopefully I’ll finally be able to get 10bit on my drawing tablet. It’ll arrive next week. I’m so excited! 😝
And then everything on TV is rendered in something like 720p or some shit with compression so bad that camera quality doesn't matter 😂👍🏻 Same with streaming services
Wanna have your 8 bit to look 12 bit? Use vivid colors. That's exactly what vivid does, as you see on Tom hardy. He ended up looking like a powder puff...
That stressed me out that he uses 8 bit (for each primary color) instead of 24 bits (for all possible color) I never used the number of bits according to the primary color
Dude, you got it wrong. As far as I know you define bits per pixel, not bits per RGB color because video doesn't have to be RGB encoded. 2^24 colors is 24bit color, not 8bit. Besides that yo you have shown 8bit image with 8bits/pixel not 8bits per each composite color of RGB. That's why there are bands. 8bits image has bands because it uses ony 2^8 colors, where 16bit image uses 2^16 colors and 24bit image uses 2^24 colors PER PIXEL. Another reason for bands is lossy compression.
Correction! 16 Bit video is actually 281 TRILLION colors! I got my Bs and Ts mixed up 😅
Can the human eye even perceive that?
@@nalingupta4005Don’t believe so.
@@nalingupta4005I think we can perceive like 10 million?
Most colours are invisible to humans( light)
Nice explanation, but you should have mentioned that the bit counts are per Color channel, R=red, B=blue, G=green. So the actual bit count is (almost) always three times the count you mentioned. So your 8 bits are actually 24 bits. 😊
I love how people can see the difference viewing 10 or more bit footage on 8 bit panel 😅
Çoğu panel 32-bit değil mi?
I love how humans can't even see that many colors.
Because the effect is exaggerated for the video.
@@TinaNewtonArt Maybe not recognize them individually, but you can definitely tell the colorbanding when there Arent enough colors in a gradient.
@@mimimi9170 onlar işletim sistemi, 64 bit ve 32bit. Burada monitör hakkında konuşuyorlar
This was super useful, I never knew what that meant, I assumed it was something to do with the information captured but I hadn’t realised it was to do with the amount of colour’s. Thank you!
He is wrong. Bit depth pertains to the precision of the ADC, and even then, most of it is false precision because very few cameras capture enough dynamic range to take advantage of it. A higher bit depth itself does not produce more colours but rather, it means the tone curve has more steps (greater precision), making banding/contouring artifacts are less visible. The comparison at the beginning showing different colour grades is misleading and has nothing to do bit depth.
@@Amaraldothis
“Some high end cameras” laughs in EOS M, iykyk
My experience as video engineer is that 10 bit typically allows better compression due to less bit banding. The extra colors over 8-bit aren't noticeable for the average person, but the extra colors are useful to provide smoother transitions, thus adding less compression noise.
Ah. This explains why when I create a LUT based on a 14 bit photo in Lightroom and convert the preset to a .cube file, I can’t usually crank up the new LUT to more than 50% on my 8 bit footage without the prior stated banding.
Just remember the best camera is the one you have!
My eyesight is -4
Mine is -7 @@ibie27
I'm pretty sure that the difference in the 2 photos is just compression levels. On non-compressed photo, you CAN have super smooth color transition with only 8bit.
It was just a reference image, the actually difference is very minimal for everyday photos and videos. But matters a lot for professional work.
@frstylol Many thanks for replying to my comment and contextualising the pictures. Since not all people read comments, it would be good for the OP to add a "illustration only" or anything like that on future videos, to avoid misleading anybody...
@@enricesThese guys put more effort into telling you what and what not to buy than actually teaching you how to master your craft, most people typically don’t look that closely at the screen to notice anyway
8bit looks super smooth, and the banding only happens in if you crack up the exposure,
You can't really see the bit banding, the one shown was exaggerated to illustrate the issue. The main advantage is to reduce the noise on the image, making compression more efficient.
@@gteixeirain darker scenes especially in movies you can clearly see the banding with the shades of gray in 8bit displays
@@Shockload I am a video engineer myself and never noticed any bit banding with a naked eye at 8 bits, even in dark scenes. It does, however, causes issue with video compression, so 10 bit is typically preferred to have better compression with less losses.
Whoa! This is a huge new thing to learn! I had no idea they were that different. I thought bits just referred to how much info the camera gathered per photo or video or something like that.
There are NOT that different. Read my comment, the 2 pictures were not really showing 8bits vs 16bits
I think speaking of more colours is confusing. It's better to say the transition steps between the colours are finer and thus allow for more shifting without loosing the fine differences between those steps.
And the most impressive things is that our brain can recognise all these colours
until you edit that 8 bit footage LMAO
😂😂😂 Google how many colors humans can see.
Light is analogue and digital is discreet so the additional bits allows tha abiltiy to chop up the visible light spectrum into more discreet amounts. 8 bits is 2^8 and 10 bits 2^10.
Im so glad cameras with this still use the gold standard of a 2 inch lcd screen thats 360p and shows all of 3 colors built in
Super easily digestible information thank you man!
Very clear! You’re brilliant thanks for the video.
And I still remember my old PC monitor with 256 colours... and I can enjoy movies at that time. 😂
Dude! This was just awesome!
This helped me out a lot. Thanks
Thanks for this. It's a question going through my mind.
Thank you broo keep up the good work thanks so much
Had no idea abt that till today.
Thanks
Wow! This was super helpful!
Thanks for the explanation
thank you so much! short explained and i understood everything!
Meanwhile TH-cam shorts is crippling the potential of my gigabit internet on my 2023 flagship phone
Perfect explanation 🎉
People with their 8bit devices.
Me watching on a black-white screen
👁👄👁
Great explained for a short video
Wow saviour ❤️📸🖤 thanks a ton crisp and clear content
Even most phone cameras see 10 bit but they encode 8 bit for some reason. Most likely because you don't have the same features as the pro versions. Just like, although it has the power to shoot 4k 60 FPS, they added this to the pro versions.
per bit it should be represented as 2^(3*n) because there is 3 color channels of rgb
Ahhh
10bit only makes sense in processing or if you're going to display it on an HDR screen. Otherwise, you're displaying 10bit colors on 8bit display and the banding will still show up. It's the same for audio: Use 24 or 32bit for processing and 16bit for playback.
Banding happens in any bit depth, a lot due to compression. To "fix" this you should add some gaussian noise to the areas that need it. This will solve it even in 8bit
Quick question to all the photo nerds: Am I correct that this is 8-bit for per R, G or B value?
A byte can store maximum of 2^8 or 256 different configurations, so the total amount of different combinations would add up to 256^3 or 16,777,216 different colors.
Also, this means that Hexadecimal color codes are also in 8-bit colors, right?
Yes, 8 bit per channel, represented by three hex numbers.
Having a 10 bit source is only advantages when you are changing brightness and contrast of the image in post.
There is 0% difference in quality of 8 and 10 bit if you are using the photo unedited.. 8 bit gradient per color is already better then your eyes can detect...
Thanks. I didn't think I needed this. And i did not need this actually. So as I was saying, thanks.
I instantly "Liked and Subscribed!!"
Best explanation so far for me
is your camera a canon? what kinda cam you use for this?
8 bit video captures the same range of colours as 10 bit video but there are more gradients steps in-between and the difference is not nearly as bad as shown in the video, but can be kinda shitty in very dark scenes. 8 and 10 might not be so far away but given that each additional bit doubles the amount of steps for a colour channel per pixel, it makes sense. having 256*256*256 color steps is less than having 512*512*512 color steps
Been into photography for a few years now but never knew what reviewers meant when they talk about bits. Now I know.
how lol even I as a non photographer knew that
@@Alex-tu5vu ok lol
Should we consider the resolution capability of TV or Phone Screen to see these on video???
My eyes can see all those colours.
I still dont get this
rgb color space is still only 3 characters so logically how is that even gonna work
A picture speaks a thousand words😮😮.
I wish TH-cam were like Vimeo where I can reshoot and replace the same content then post a new correction content
Nice: 👍🏿 I see you can be a good teacher... 😅😊
What's background music track?
I wonder how many bit of colors a human can usually see...
I have a ¥23,000 10bit 2K 165hz hdr monitor. The color is noticeable compared to my Huion Kamvas Pro 19 8bit when connected via hdmi. My gpu nor my motherboard doesn’t have type c alt mode (type c displayport) so I ordered a “Wacom Link Plus”. Hopefully I’ll finally be able to get 10bit on my drawing tablet. It’ll arrive next week. I’m so excited! 😝
Can't wait until 64 bit comes out or even 128 bit.
bro immediately dropped the decimal after million, shame
Need help, when I play my 10bit videos on pc, it freezes or lags 😢
so if i put 10 bit and 8 bit video side by side on same 10bit panel, you should tell which one is which right? i bet you can't
Oh wow that's cool now ik my Lgv30 value
Pixel phones can shoot HDR 10-bit photos and videos 🎉
As far as i remeber i know only 12 colours in colour pencil box😂
How many bits do we see in? 🤔
In other words the higher the bit the better the quality.
My LCD screen doesn't support 10 bit so they basically look the same bro 😂😢
And then everything on TV is rendered in something like 720p or some shit with compression so bad that camera quality doesn't matter 😂👍🏻
Same with streaming services
How set in camera 10 bit
Is this true when posting on YT and IG?
Wanna have your 8 bit to look 12 bit? Use vivid colors. That's exactly what vivid does, as you see on Tom hardy. He ended up looking like a powder puff...
well it's not quite what the camera can see, more what it can process and encode
So banding is the name, i won't forget that
how do i know how many my camera has
So what's 64 bit like?
Which GPU can Rander 16 bit H256 4k footage
I have shot 8bit video all my life and my video never looked like the crap sunset video you just showed us.
How many colours can the human eye see?
now you just need a 10 or 12 bit monitor.
How many Bit dose see our eyes?
BITS have more luma information. Bitdepth sampling has more chroma
Me using 32 bit in Blender 😎
Uh but I can see only seven colours 😂
My phone can capture nice looking photos btw 🌝
Two more decimals diference
I dont get how 16.7 million colors isnt enough to get smooth gradients.
It can. That image was posterized (artificially banded).
long story short if youre only planning on posting online, more than 8 bit is not necessary at all
That stressed me out that he uses 8 bit (for each primary color) instead of 24 bits (for all possible color) I never used the number of bits according to the primary color
But the thing is human eyes can see only 1million colours😂
But there's only three colours needed to create all the colours 😅
Ah.. Thats the reason my picture seem bending..
Imagine bought a camera to record movie in 16 bit video and people will reproduce the movie in 8bit 480p Facebook livestream. 😅
Dude, you got it wrong. As far as I know you define bits per pixel, not bits per RGB color because video doesn't have to be RGB encoded.
2^24 colors is 24bit color, not 8bit. Besides that yo you have shown 8bit image with 8bits/pixel not 8bits per each composite color of RGB. That's why there are bands. 8bits image has bands because it uses ony 2^8 colors, where 16bit image uses 2^16 colors and 24bit image uses 2^24 colors PER PIXEL. Another reason for bands is lossy compression.
I bought my first camera and i think i went a little overboard
The real question is: can my eyes actually see those colors?
My canon eos m record 14 bit. 200 bucks camera
Why do I see everyone talking about eos m rn I have it for like two years and it's amazing but did someone make like a video about it?
maybe just photos
8-bit ALSO means 320x200 resolution
NGL, I can tell almost no difference between an 8 and 10 bit OUTPUT when shown on their own...
Helium?
I wonder how many bits does an iPhone camera has
I didn't know that there's more than 10 colors
Then why is my pastel have only 6 colours, it's cheating broh.
My 15 year old tv can read 12 bit signals
tnx but does not really help much of OSMO ACTION 3 bave focus issues hahahaha
At the End of the Day Who TF cares about how many million it is.
😮💨
Do this much colour even exist? First question I got