Recurrent Quality Issues on Youtube Upload (1080p25)

[Alexique]

Hello everyone,

Here's my first post, cause I'm getting desperate to find a right way someday.

Which are the right settings, and/or What is the trick, to have the most accurate result when uploading a video on Youtube?

If I may precise:
* I mainly edit videoclips in HD 1080p, 25fps, on Premiere Pro or Avid MC.
* I've been this recurring problem: no matter how good is the visual quality of my video export, then it becomes pixelated with a loss of contrast and saturation when uploaded on Youtube.
* I've read Youtube video specifications, I've read and watched tutorials to try some new ways: converted and uploaded in H264 in VBR 2 pass 50Mbps, in H265/HEVC, or directly in a ProRes HQ, upscaled in 4K/UHD to force youtube to use its vp9 codec. And I did that both from Avid and from Premiere/MediaEncoder.

Yet, most of the videos I see on Youtube don't suffer from this issue.

I can anticipate the contrast/sat loss by enhancing effects in my timeline before conversion, but I can't do anything about the resolution loss. Herewith you can find an example: one screenshot is from my video file, the other one is from the Youtube upload (here I adjusted the color so it's mainly about quality).
(Sorry I had to blurry the eyes, which are the place where we feel the difference the most.)

Would you have a solid explanation or a magic solution for me?


Thanks a lot for your help!

A

[Swedaniel]

Hello everyone,

Here's my first post, cause I'm getting desperate to find a right way someday.

Which are the right settings, and/or What is the trick, to have the most accurate result when uploading a video on Youtube?

If I may precise:
* I mainly edit videoclips in HD 1080p, 25fps, on Premiere Pro or Avid MC.
* I've been this recurring problem: no matter how good is the visual quality of my video export, then it becomes pixelated with a loss of contrast and saturation when uploaded on Youtube.
* I've read Youtube video specifications, I've read and watched tutorials to try some new ways: converted and uploaded in H264 in VBR 2 pass 50Mbps, in H265/HEVC, or directly in a ProRes HQ, upscaled in 4K/UHD to force youtube to use its vp9 codec. And I did that both from Avid and from Premiere/MediaEncoder.

Yet, most of the videos I see on Youtube don't suffer from this issue.

I can anticipate the contrast/sat loss by enhancing effects in my timeline before conversion, but I can't do anything about the resolution loss. Herewith you can find an example: one screenshot is from my video file, the other one is from the Youtube upload (here I adjusted the color so it's mainly about quality).
(Sorry I had to blurry the eyes, which are the place where we feel the difference the most.)

Would you have a solid explanation or a magic solution for me?


Thanks a lot for your help!

A

Have you tried increasing the video bitrate? Try boost the video bitrate to maybe 400-800kbps.. if it is allready above theese values, try add more video bitrate to your settings which can help remove pixelation.

[Alexique]

Hi Swedaniel,

Thanks for your answer. Theoretically the max bitrate into Youtube is 80Mbps (80000kbps) I was currently at 50Mbps and I tried 100, but same result here...

(Something I didn't note: I did check that the play settings in Youtube were set on HD 1080p.)

[Swedaniel]

Hi Swedaniel,

Thanks for your answer. Theoretically the max bitrate into Youtube is 80Mbps (80000kbps) I was currently at 50Mbps and I tried 100, but same result here...

(Something I didn't note: I did check that the play settings in Youtube were set on HD 1080p.)

Ok, then im not sure! What i would try myself then is to increase the frame rate, which will create more file size, and then instead decrease the video bit rate..

Having videos with various frame rates make it difficult for the eyes, maybe the youtube player compensate with the bitrate which cause pixelation and basicly just an video frame rate + bit rate that does not cause the youtube player to use compensation is all that is needed..

[MicroMagic]

Ok, then I'm not sure! What I would try myself then is to increase the frame rate, which will create more file size, and then instead decrease the video bit rate. Having videos with various frame rates make it difficult for the eyes, maybe the YouTube player compensate with the bitrate which cause pixelation and basically just a video frame rate + bit rate that doesn't cause the YouTube player to use compensation is all that is needed.

Actually, I'm not necessarily the best guide when it comes to practice, but I've invested a lot of time in the last few weeks to find out how the technology behind YouTube works. I'm not a YouTuber myself because I'm not a Google member and would therefore rather upload to Dailymotion. But that's just my personal point of view.

In principle, it is a rather bad idea to work with less than 30 fps, as the image content then appears less fluid in movement. It should also be noted that the resolution of the upload plays an extremely important role, as even here conversion losses are unavoidable. However, these will be kept within limits if the resolution is kept as high as possible (e.g. 4K). For example, an upload in 1024p @ 30 fps will always lead to conversion losses on the YouTube player with the same playback resolution. In addition, there are also compression artifacts that appear sequentially in the image, since all codecs used at the moment are more or less lossy, except VP9 in special cases. It should also be noted that YouTube expressly recommends using MPEG-4 Part 10 (H.264), also known as AVC, for uploads. The choice of codec for conversion depends on several factors on YouTube, on the one hand it's the resolution, on the other hand the size of the file, and the access frequency of the respective channel is also a decisive factor. This means that the more popular a channel, the more likely it is that YouTube will use a lossless format.

YouTube does not want to pay license fees, so MPEG-H Part 2 (H.265), also known as HEVC, is out of the question. There are actually three codecs currently in use:

MP4 Part 10 (H.264),
WebM VP9 (lossy and lossless)
AOMedia Video 1 (AV1).

But that's not all, luckily there's a lot more to come, because developments continue. In order to finally eliminate the previously known shortcomings, a new standard has now been established: Versatile Video Coding (VVC), also known as MPEG-I Part 3 (H.266), ISO/IEC 23090-3. The new standard offers approximately 50% better compression rate compared to HEVC with the same perceptual quality and supports lossless and subjective lossless compression. It supports resolutions from very low resolution up to 4K and further up to 16K as well as 360° videos. In addition, it offers BT.2100 wide colour gamut and high dynamic range (HDR), variable and fractional frame rates from 0 to 120 Hz, and support for higher and high bit depths (12 to 16 bits per component). There are countless more details to mention about this part, but I'll keep it brief. In this case too, license fees are expected from some places.

I hope my explanations are at least a basis for your future clips.

[Swedaniel]

Ok, then I'm not sure! What I would try myself then is to increase the frame rate, which will create more file size, and then instead decrease the video bit rate. Having videos with various frame rates make it difficult for the eyes, maybe the YouTube player compensate with the bitrate which cause pixelation and basically just a video frame rate + bit rate that doesn't cause the YouTube player to use compensation is all that is needed.

Actually, I'm not necessarily the best guide when it comes to practice, but I've invested a lot of time in the last few weeks to find out how the technology behind YouTube works. I'm not a YouTuber myself because I'm not a Google member and would therefore rather upload to Dailymotion. But that's just my personal point of view.

In principle, it is a rather bad idea to work with less than 30 fps, as the image content then appears less fluid in movement. It should also be noted that the resolution of the upload plays an extremely important role, as even here conversion losses are unavoidable. However, these will be kept within limits if the resolution is kept as high as possible (e.g. 4K). For example, an upload in 1024p @ 30 fps will always lead to conversion losses on the YouTube player with the same playback resolution. In addition, there are also compression artifacts that appear sequentially in the image, since all codecs used at the moment are more or less lossy, except VP9 in special cases. It should also be noted that YouTube expressly recommends using MPEG-4 Part 10 (H.264), also known as AVC, for uploads. The choice of codec for conversion depends on several factors on YouTube, on the one hand it's the resolution, on the other hand the size of the file, and the access frequency of the respective channel is also a decisive factor. This means that the more popular a channel, the more likely it is that YouTube will use a lossless format.

YouTube does not want to pay license fees, so MPEG-H Part 2 (H.265), also known as HEVC, is out of the question. There are actually three codecs currently in use:

MP4 Part 10 (H.264),
WebM VP9 (lossy and lossless)
AOMedia Video 1 (AV1).

But that's not all, luckily there's a lot more to come, because developments continue. In order to finally eliminate the previously known shortcomings, a new standard has now been established: Versatile Video Coding (VVC), also known as MPEG-I Part 3 (H.266), ISO/IEC 23090-3. The new standard offers approximately 50% better compression rate compared to HEVC with the same perceptual quality and supports lossless and subjective lossless compression. It supports resolutions from very low resolution up to 4K and further up to 16K as well as 360° videos. In addition, it offers BT.2100 wide colour gamut and high dynamic range (HDR), variable and fractional frame rates from 0 to 120 Hz, and support for higher and high bit depths (12 to 16 bits per component). There are countless more details to mention about this part, but I'll keep it brief. In this case too, license fees are expected from some places.

I hope my explanations are at least a basis for your future clips.

Yep ok! I havent uploaded many videos to YouTube, but ive been around web video the players hosting explanations, which one of them was YouTube.. Which the explanation was, instead of causing the video to stop and give this loading circle because of low bandwith or to high quality on the video, it would instead compensate by adding pixelation to the video to keep the video playing and not stop/lagg! So what i was thinking is this pixelation appear because the video is not appearing on the screen within the demands of the youtube player, since without it it ill cause uncomfort for the viewer... (possible because of to low frame rate/fps).

Edit: H.266 sound cool! 360 degrees videos, is that the videos that can be recorded in an enviroment that some buisnesses use on their information websites so that customers can check the surroundings from the internet before visiting? And sometimes allso some similar type of videos is avalible from an airplane/helicopter recording above the buisness area? I know like 10 people IRL that have real companies and sell such videos to companies, theyre the coolest videos that excist...

[MicroMagic]

Yep ok! I haven't uploaded many videos to YouTube, but I've been around web video the players hosting explanations, which one of them was YouTube. Which the explanation was, instead of causing the video to stop and give this loading circle because of low bandwidth or to high quality on the video, it would instead compensate by adding pixelation to the video to keep the video playing and not stop/lag! So what I was thinking is this pixelation appear because the video is not appearing on the screen within the demands of the YouTube player, since without it it will cause discomfort for the viewer (possibly because of a too low frame rate/fps).

As far as I know, this explanation is incorrect. There is currently not a single platform on which such a procedure is used. There are basically two bandwidths to consider here: that of the server that offers the videos for download and that of the Internet provider that serves the end user. In most cases, the end user has enough bandwidth for 4K videos, although this is not guaranteed on the respective video platforms. So you have to expect that a video can get stuck for at least a few seconds. How this looks in detail depends on the playback software used, such as the browser. In the event of a corrupted video file, the image would be distorted rather than being corrected by software in any way. In all cases, the playback software is therefore equipped with a buffer memory that is usually large enough to allow the stream to run without interruption. However, this also has its limits if the download becomes too slow, as the amount of data becomes too large in relation to the bandwidth.

What should also be mentioned in this context is that there are two types of conversions, a primary and a secondary one. The primary takes place on the download server, the secondary, however, takes place within the playback software. So if a video with a playback resolution of 1080p @ 60 fps starts to get stuck, then there is still the option of changing it to another format such as 270p @ 60 fps. Some servers in Asia are affected by this, but those in Europe and the Americas are actually not. Last but not least, the size of the display also plays a role. 4K one-to-one on the monitor requires a correspondingly large device, and not everyone has that. Therefore, most users will not be able to avoid a secondary conversion by downscaling in the playback software. From a technical point of view, the developers' primary concern is to keep conversion losses as small as possible at all levels.

I wish you a pleasant Sunday!

[Swedaniel]

Yep ok! I haven't uploaded many videos to YouTube, but I've been around web video the players hosting explanations, which one of them was YouTube. Which the explanation was, instead of causing the video to stop and give this loading circle because of low bandwidth or to high quality on the video, it would instead compensate by adding pixelation to the video to keep the video playing and not stop/lag! So what I was thinking is this pixelation appear because the video is not appearing on the screen within the demands of the YouTube player, since without it it will cause discomfort for the viewer (possibly because of a too low frame rate/fps).

As far as I know, this explanation is incorrect. There is currently not a single platform on which such a procedure is used. There are basically two bandwidths to consider here: that of the server that offers the videos for download and that of the Internet provider that serves the end user. In most cases, the end user has enough bandwidth for 4K videos, although this is not guaranteed on the respective video platforms. So you have to expect that a video can get stuck for at least a few seconds. How this looks in detail depends on the playback software used, such as the browser. In the event of a corrupted video file, the image would be distorted rather than being corrected by software in any way. In all cases, the playback software is therefore equipped with a buffer memory that is usually large enough to allow the stream to run without interruption. However, this also has its limits if the download becomes too slow, as the amount of data becomes too large in relation to the bandwidth.

What should also be mentioned in this context is that there are two types of conversions, a primary and a secondary one. The primary takes place on the download server, the secondary, however, takes place within the playback software. So if a video with a playback resolution of 1080p @ 60 fps starts to get stuck, then there is still the option of changing it to another format such as 270p @ 60 fps. Some servers in Asia are affected by this, but those in Europe and the Americas are actually not. Last but not least, the size of the display also plays a role. 4K one-to-one on the monitor requires a correspondingly large device, and not everyone has that. Therefore, most users will not be able to avoid a secondary conversion by downscaling in the playback software. From a technical point of view, the developers' primary concern is to keep conversion losses as small as possible at all levels.

I wish you a pleasant Sunday!

Ok, well all i know i went to an explanation hosted and created by YouTube themselfs and a few other video websites and they all said the same thing basicly, it included information videos! this is many years ago tough! Im pretty sure any video player is using some functions to automaticly protect the viewer from uploaded videos that may cause uncomfort! Ellse it would less likely be an world wide approved web browser video player! Aswell when installing some web browser video players it is allso possible to select how the player should function in case the viewer have to low bandwith, either add pixelation or to wait for the buffer to complete which will stop the video and give this loading circle and aswell to increase the amount of buffering, automaticly force and decrease the quality and so on.. sometimes it is even possible to change the appearance of this loading circle, some even offer to help create your own with your brand/logo to be used as loading logo..

If one would want to upload an video to such an website it would seem obvious to customize the settings of this video for an specific site to get the best video for this website.. It seem pretty obivious if an website allow users to upload their own videos that they add security measures to automaticly protect the viewer from videos that cause uncomfort for their eyes, or give headache etc, which this allso must be added to concideration to customize the video settings to get the best effect/quality and as low loading times as possible... One of theese solutions to give less uncomfort is to add pixelation.

Edit: By not having such built in solutions directly in the video player, this would mean 1. either during the upload the video must be scanned and altered to remove this uncomfort 2. the video does not get altered and it gets automaticly blocked from uploading for it cause uncomfort 3. the website does none of this and uploaded videos cause uncomfort (less likely to get world wide approved then and people get headache/uncomfort etc) 4. instead of all this the web browser video player have built in functions to protect against all of this and it will be less failed uploads and the demands of the server hardware is lower during uploads because the web browser video player take care of the security which lead to fewer and cheaper server costs and with all of this more likely to get more successfull and approved uploaded videos which will increase chances for becoming an successfull website such as YouTube.. 5. maybe more!

Im pretty sure the most video website is going with #4.. either that or they set demand using an specific codec which this codec takes care of such security mesaures, which funnily could add pixelation during the encoding instead, and possibly even add restrictions to specific settings to be controlled during upload, which, funnily here we go again, by adding an web browser video player that can during playing take care of some issues will give the uploading user more flexibility in the settings and also less failed uploads, increasing the successfull ones ! Really, who wouldnt want an website with millions of users to upload videos, and not have an flexible way of getting successfull uploads!

Even if one would just want to create an small personal website it would seem important to offer an video for an friend that you could yourself just easily add the video and be sure this web browser video player have built in functions to protect the friend from an video that may cause uncomfort.. Which leads allso to no need of beeing some sort of video, eye and headache expert to just share a few videos! Makes life much easier when creating or converting an video knowing that the video player take care of some things for you, making life simple and not complicated! If this video player would then add pixelation to create less uncomfort it it easy to understand that customizing the settings for the video will remove this pxelation created by the video player or the codec.

If one would add an video to youtube that somehow had colors or lightning that was very strong and bright so it cause uncomfort for an viewer, it is easy to understand that this would be automaticly altered during upload, or by the video player! Ellse it seem like an easy terrorist act if people could just add videos that cause such uncomfort! To then say there is no platform that use such things seem like an faulty thing to say imo!

Atleast this is what ive learned about web browser video players many years ago, and it is my answer, hopefully magical enough as the person asking the quiestion asked for, hoping to give an slight clue or hint to what could allso cause pixelation or other issues with uploading an video, hopefully leading to solving the issue to get an perfect video uploaded to YouTube!

There is allso alot of people who have made videos, for educational purposes that have a few things displayed in their videos which will demonstrate and display some things that is real easy to understand how it could cause uncomfort and many other things, and with this allso how for example automatic pixelation during such an event removes the uncomfort! Some do videos of people falling or beeing tricked and some do theese type of videos! theyre definetely worth an view if you would want to understand video players to the point that any error/failure/pixelation would make sence to yourself to keep you motivated and faster guide you into the solution based on your own knowledge during your video creations, uploads and conversions!

[MicroMagic]

As the operator of a platform, who should judge what is perceived as uncomfortable and what is not? Personal feelings are not a valid benchmark here. If a video file does not meet the minimum technical requirements, it will be rejected rather than corrected by subsequent measures. There are now several procedures that make reading files safer, including automated repetitions during reading and other redundancies. This applies not only to the provider of such services, but also to the playback software. All processes used today have only one goal, namely to keep the video running in the best possible quality. This is understandable, as it would be a very bad idea to dictate to the viewer what is perceived as comfortable and safe and what is not. I, for example, find 24 fps unpleasant. Should we now ban everything that is uploaded with this frame rate?! In any case, temporarily repairing a video during its runtime through replacement measures is not an adequate method and should therefore be rejected. An acceptable measure would be to automatically cut out individual passages rather than using other processes to bridge or otherwise change the content. Luckily for us, almost all providers today do without these outdated methods from the video stone age.

I'm also sure that most authors weren't happy with such relief measures! The software used today is generally secure enough to make such unpopular measures largely superfluous. There are already too many miserable videos on the internet, so it would make more sense to raise the minimum standards than to follow up on poor material!

[Swedaniel]

As the operator of a platform, who should judge what is perceived as uncomfortable and what is not? Personal feelings are not a valid benchmark here. If a video file does not meet the minimum technical requirements, it will be rejected rather than corrected by subsequent measures. There are now several procedures that make reading files safer, including automated repetitions during reading and other redundancies. This applies not only to the provider of such services, but also to the playback software. All processes used today have only one goal, namely to keep the video running in the best possible quality. This is understandable, as it would be a very bad idea to dictate to the viewer what is perceived as comfortable and safe and what is not. I, for example, find 24 fps unpleasant. Should we now ban everything that is uploaded with this frame rate?! In any case, temporarily repairing a video during its runtime through replacement measures is not an adequate method and should therefore be rejected. An acceptable measure would be to automatically cut out individual passages rather than using other processes to bridge or otherwise change the content. Luckily for us, almost all providers today do without these outdated methods from the video stone age.

I'm also sure that most authors weren't happy with such relief measures! The software used today is generally secure enough to make such unpopular measures largely superfluous. There are already too many miserable videos on the internet, so it would make more sense to raise the minimum standards than to follow up on poor material!

I dont know who should judge within theese things, for things such as videos, framerates, bitrates, eyes, headaches, disturbing? an optician? maybe an specialized IT Expert with some sort of optician education that somehow set an legal rule system together with some sort of medical people? Or maybe just an overly good person who can download an approved web browser video player knowing it is all taken care of! I guess that would be ideal instead of having to live in an world where things are complicated, maybe if one would want to purchase an bisycle we must allso all of us invent the tire over and over and allso be an expert to even get the air into the tires!

If someone would want to use an 20 year old nokia to record its freshly caught crab and upload it to youtube, it dont seem like an answer to "increase" the standards and only use Ultra HD/4K with shiny, glossy no pixel videos, it sound like an moment to simplify the upload and increase the support for video formats with as much flexibility within the settings as possible so that everyone can check out this crab recorded with this old nokia!

Many people that say such things to increase standard etc is many times jealous people for some 5 yeard old kid have uploaded an video to youtube and recieved millions of viewers even tough the video was just an simple thing, while some other person with super advanced codecs, specialized, and puchased settings, Ultra HD/4K and what not, still does not manage to scrap up 5 viewers and come later crying, saying the standards must increase!

Edit: I dont know if the 24 fps videos should be banned, maybe those medical experts know that! Maybe one day the video player add pixelation or something ellse to automaticly make such an video not-uncomfortable! Because when i used Windows XP a real long time ago, this feature was basic thing for an video player to have avalible for its viewers/customers!

Some people watch videos all day, every day, it is an part of their life! If 4K is what they want lets give them 4K, but if it cause uncomfort for them after 30 minutes , lets add some pixelation if it help!

[MicroMagic]

Some people watch videos all day, every day, it is a part of their life! If 4K is what they want let's give them 4K, but if it cause discomfort for them after 30 minutes, let's add some pixelation if it helps!

No, it doesn't help at all. If someone don't like what a platform has to offer, then he or she is free to look for something else. It's completely irrelevant what habits the audience develops, as it's not primarily oriented towards the quality, but rather towards the content and that is the decisive factor. It would also be completely absurd if, in addition to the already known side effects that arise during the conversion, additional ones were introduced that would increase the losses that have already occurred even further by artificially changing the content. At the same time, it would also not be legally justifiable to change the content without the author's knowledge and consent. Twenty years ago things looked different, but back then neither the codecs nor the playback software were so advanced that a battered video file could be made to work without artificial intervention. But such tricks are now passé, the way a video file is rendered is now much more complicated and accompanied by highly developed algorithms that are intended to ensure the end result is of the best possible quality. The effort required to keep the compression losses that still exist as low as possible is already immense, so it would be a decidedly retarded idea to artificially induce further losses here!

It cannot be that the technical background of a development is oriented towards what a few confused minds find comfortable or safe. It's better to distance yourself from this seemingly fixed idea. Fortunately, the reality looks different, there is no distortion due to incorrect pixels, not even on Chinese or Russian platforms!