Bit Depth: Where Filmmaking Meets Computer Science - Video Tech Explained

== Navigation ==
00:00 Opening
00:42 How Computers Store Numbers
04:27 The Precision Problem
06:05 Binary Counting
08:20 The Trouble with 8 bit
11:51 Gamma Curves
13:17 Which Bit Depth Should You Use?
15:47 Closing & Credits
== Links ==
Join the official VTE Discord server: / discord
Support me: / videotechexplained
My portfolio: camoncrocker.com

[@iComplainer]

i love this guy. no bull shit. no “hey guys, ad ad ad , lets get in to it, without further ado” etc. just gets to the meat and lets me geek tf out.

[@MiMEDesignStudio]

Exactly!

[@rubenlevisznajderman6284]

This is quickly becoming the best channel on video science.

[@NikiWinProd]

i agree

[@ouajbir]

This guy is UNDERRATED here in KZbin, he deserves more subs, seriously

[@CaptaNovious]

I used to skip computer science classes and go to art classes at uni - so I could learn about both. I am glad to have found an intersection of Art and Technology explained so well. Thank you.

[@MiMEDesignStudio]

You are the best in the biz.

[@totochandelier]

well edited, and every point of attention is highlighted . High level work : you're my reference now . Thank you!

[@cameradave1]

Brilliant. The best explanation of bit depth I've come across.

[@JeffBourke]

Wow. This kid has a big future! Outstanding method of explanation

[@rajatbanerjee3413]

C'mmon boy ,, restart making videos .. eagerly waiting

[@hikerbikeromar]

Wow, this was absolutely brilliant. I can't believe this has under 1k likes. Thanks you so much!

[@price561]

This my new one stop channel. I've learned more from 3 of his videos than my years working in Post Production lol

[@nathanabrahams5305]

Haven’t even watched it yet, but liked it cause I know it’s gonna be good 🔥

[@BekirSai]

Video is quite teaching and simple, thanks. I hope you'll get bigger in KZbin. 😊

[@mattwolcott3266]

Excellent. Your channel is simply undiscovered but it will definitely blow up with the quality you produce. Keep it up!

[@redstoneranger1404]

It was already mentioned: Best video science channel by far. Please continue!

[@yperboreus]

Thank you! Finally understood bit depth..

[@maenzeidan9154]

You are a great teacher, thank you 🙏

[@nathanielmcbride2645]

As usual, I learned something new! Hats off to you sir. Thanks for the stellar content!

[@NibuThomson]

Thank you so much for the notes👍

[@edrfitness]

Amazing work! Looking forward to you getting the attention you deserve. You've helped me up the quality of my videos massively since I discovered your content. If you have a course to take, I will buy it.

[@LEscaladeCaragol]

Well done. I want you to know that your explanations are touching the rim of the tech stuff I am able to understand. You have a very capable intellect. Congratulations!

[@eggs-benny]

awesome explanation, thanks!

[@thejubinbennett]

Great video !! you are really good at explaining in a simple way !! keep going.

[@filmcrafterscanada]

Keep up the good work! Subbed 🙏

[@matoflynn]

Thank you! Best explanation available.

[@Janken_Pro]

This video is insanely informative and easy to understand.

[@jannisnewiadomsky]

very well done !

[@DanSantanaBows]

Excellent! Thanks for the wonderful videos

[@AbhinavanandSingh]

Sincerely appreciated!

[@ArunKumar0923]

Good information Perfect explanation

[@99SBX]

Dude this video is awesome.

[@egjundis]

Another great video. I'm looking for more info on understanding log vs. linear in compositing. Hopefully, you have something already or will do so in the near future. Thank you!

[@12darinho]

What about shooting HLG3 on A6400? It has good dynamic range and won't fall apart as fast as SLOG3. As far as i know, HLG its linear + log (hybrid log gamma).

[@alphonseraj9434]

Hats off to you

It's a shame not more filmmakers....or 1 ACs watch these videos. Having this as a base to "dumb down" the tech helps a lot when you are getting into it.

[@shirmaman1886]

i cant be the only one who thinks this kid looks like little timmy from jurassic park right?....amazing content by the way.

[@shueibdahir]

Camera sensors including the cheapest ones are able to achieve a level of color detail that's professional 10k+ cinema camera level. Problem is no one wants to work with huge files and also camera companies have been locking down their firmware so you're unable to directly access the raw data, even with photography. Camera sensors aren't bound by bitdepth they're analog devices. The ADC and image processors are though. I believe no log video format or even hdr such as hlg is supposed to be graded. Anytime a camera writes a video file to a memory card, it has already processed the image for you. You wouldn't push a JPEG and expect good results now? So read the technical paper on the specific log curve you're using, follow its instructions and get the look you want in camera. You can do very slight adjustment but the problem is with video files, whether its 8bits or 12bits, anytime you push or pull a single color, it inevitably messes with every other color. Raw bayer Data, such as those out of RED, Sigma FP, Arri and Canon Magic lantern are the only ones i'd feel comfortable pushing around. Instead of the image processor, you do each and every step yourself. Those cameras are essentially data collection devices, not video cameras

[@choke_the_woke1179]

you need to make a vide explaining in depth the elation between Bit depth and gamma curve

[@skymakai]

Very well explained. I only would have asked that the distribution of data in the stops of a gamma curve be explained so as to wrap up. How is one more 'efficient' while another isn't? Why is the log gamma curve distributed the way it is, so that it retains more detail in shadows?

[@VideoTechExplained]

It mostly has to do with how much of the signal range is devoted to different parts of the image. For example, if you extend a standard gamma (2.4) out to 10,000 nits, then about 14% of the bits would be dedicated to values below 100 nits, and only 37% would be dedicated to values below 1,000 nits. This isn't ideal, since values above 1,000 nits are not very common yet they are occupying the majority of the signal range. The PQ curve, on the other hand, distributes bits much more efficiently for a typical HDR grade. 25% of the signal range is dedicated to values below 10 nits, 25% for values 10-100, 25% for 100-1000, and 25% for 1,000-10,000. This distribution much more closely matches the relative importance of these regions of an image. Most HDR grades still keep their shadows and midtones below 100 nits, with the rest of the signal reserved for highlights and extreme highlights. We don't need nearly as much information in the 9,000-10,000 range as we do in the 0-1,000 range Most gamma curves are tuned to distribute bits as efficiently as possible for their particular use case.

[@choke_the_woke1179]

we need a video about DaVinci delivery page, in depth, , specially Advanced Settings.

[@phenixnunlee372]

Base 10 is fine because quantization is just an issue with rounding to whole numbers

[@JoshWilliams123]

Camon, let's say you have the same scene and you take a photo and a single video frame of it. The photo is 14-bit RAW and you export it to sRGB JPG. The 12-bit RAW video frame has the log and gamma corrected and exported as an sRGB JPG. Why does the photo look objectively better than the video frame? On paper, the photo was captured with 2 extra bits per channel. Does the extra precision account for the better color rendition and dynamic range? Granted, photographers do not have shutter angle constraints to deal with. They also don't feel compelled to shoot at base ISO 800, log, or use ND filters. In fact, you never hear photographers talk about dynamic range. Curious what your thoughts are.

[@laustinspeiss]

Bit depth, the equivalent of dynamic range in analog audio.

[@robmcd]

I’m really interested to know why dark scenes in Netflix films turn to murky ugly banding despite their attempts to restrict camera types and specs. I assume it’s a scaling or compression artefact

[@harysviewty]

Keep doing HDR video

[@VideoTechExplained]

I was waiting to see if someone would notice! Yes, I've started uploading regular episodes in HDR and hope to continue doing so in the future

[@tem425]

camon u crush

[@jannisnewiadomsky]

are you sure that redraw stores data with liniar gamma ?

[@divyanshumishra1807]

Does shooting in 10 bit and after editing exporting in 8 bit SDR leads to color loss, if yes, then what's the point shooting in 10 bits?

[@VideoTechExplained]

If you're exporting at 8-bit, then yes, you're going to be losing color information compared to the original 10-bit footage. However, there is still a benefit to shooting 10-bit even for 8-bit delivery, because it gives you more flexibility to apply color corrections/grades without introducing noticeable banding in the final product

[@NecumNaTo]

I love your content, great videos, thanks. 💪 Maybe just shout little less…