Maximum bitrate for Twitch. Video encoder settings, bitrate and live broadcast resolution Checking the bitrate

1st direction.
Obtaining the highest possible quality. In this case, a stream with a minimum of preprocessing is selected and the encoding quality is compared with the original.
2nd direction.
Preprocessing is selected for a strictly limited flow. In this case, comparison with the original is not provided.
In due time, namely when DVD format I had to seriously deal with all aspects of MPEG encoding and the conditions for obtaining maximum quality with a minimum flow rate. Naturally, the first attempt was very simple - by selecting a codec. After several attempts, the hardware codecs were thrown into the trash heap. Their scope of application can be described somewhere like this: “If you have absolutely no time, then you use hardware codecs, in all other cases, software codecs are used.”
The main advantage of software codecs is their flexibility. You get the opportunity to select filtering (and the filtering quality is very high), quantization scales, the number of passes (up to 20 passes), and most importantly, you get the opportunity to recode individual sections of the stream fragment by fragment to minimize artifacts.
To test encoding quality, we made a synthetic test based on 2D graphics. The test was made taking into account the visual assessment of the encoding quality not only of the luminance channel, but also made it possible to understand how color channels are encoded. Since we are not codec manufacturers, the test was designed to be extremely tough, with the minimum possible redundancy.
The result of this test gave very revealing data. Without using a low-pass filter and reducing the quantization scale, the test passed with a minimum of artifacts, starting from 14 Mbit/sec with multi-pass encoding, and multi-pass ceased to have an effect after 3 passes. The results were assessed on a professional 21-inch television monitor and a 21-inch computer LCD monitor from a distance of about 30-40 cm.
All this suggests that a conditionally high-quality signal in full bandwidth and standard resolution in MPEG2 SDTV can be obtained starting from 16 Mbit/s if there is an uncompressed signal at the input.
The next step was to understand which MPEG2 stream, without the use of preprocessing and quantization reduction, allows us to obtain the form of an uncompressed SDTV signal. We tried only software codecs that allow you to work with streams above 16 Mbit/sec. The result was very interesting - the peak value was 40 - 50 Mbit/sec, the average was about 30 Mbit/sec. It is best if GOP = 3 – 6.
Thus, the resulting value is very similar to the BETACAM IMX format.
All these mini-studies were carried out in principle to solve another practical problem.
Quite often it is necessary to create promotional video materials for display at exhibitions. The budget for such videos is not very large, since they are most often used once or twice. The shooting is accordingly carried out on DVCAM SONY DSR-400 (for its price it has excellent dynamic range control, detail and color in the highlight area, of course with the appropriate settings). The video material at the exhibition is shown on a fairly large plasma panel. Today you can install an inexpensive computer and show 50 Mbit/s, and before the advent of SATA-2 hard drives the most simple solution There were DVDs. But the DVCAM format itself, naturally, due to its 25 Mbit/s, is not very suitable for direct encoding into DVD with the necessary detail. You can, of course, blur everything, but you are asked to make an advertisement with all its inherent visual features.
From this moment it became clear that the most important role in coding is not played by the coding itself or the processes of killing quality before coding. The most important thing is the quality of the source material (or rather, how much it was compressed, since even noise on uncompressed video is easier to encode without a pre-process).
Planning and organizing filming taking into account the algorithms of the MPEG format gives a very great effect. Probably, here we can immediately note some features - the inadmissibility of working with a zoom lens, moving strictly horizontally or vertically, working with a small depth of field, etc. In general - as little movement on the screen as possible. Best option in general, just show photographs (and these are already the favorite frames of demo material manufacturers to show the quality of codecs and compressed video formats). This is the kind of creative lack of freedom we have because of the passion to compress everything to unimaginable states.
By the way, on the website of the British television company BBC, you can find one interesting document in the public domain. These are recommendations from the Air Force R&D department regarding what is desirable and what is not desirable to do if everything turns into MPEG. Yes, yes, I know the facial expression some of our television workers have when they mention the BBC. Well, who else has such an effective R&D department?
So now about high definition television (HDTV).
From all this experience of working in standard definition (SD), there can be only one conclusion - according to the principle of operation, HDTV encoding cannot be any different from SDS encoding.
Yes, I understand that the H.264 format is now being massively promoted under various names and promises of miracles with minimal streams.
If you carefully read the specifications of this encoding method, you will get somewhere around 20 - 25% of the gain compared to MPEG2, a little better in the 2-4 Mbps region. But with such a stream it will hardly resemble even a professionally converted signal to 720p standard definition.
For one Moscow company planning to release Blue-Ray, we carried out a number of experiments in the field of upconverting HDTV signals into HDTV with subsequent encoding into formats supported by Blue-Ray.
Here are some results when viewed on a 26-inch computer monitor from a distance of 30 cm with an average flow of 15 Mbps and a maximum of 30 Mbps.

1. without preprocessing, artifacts appear even on static scenes due to residual noise from the BETACAM analog signal.
2. When using a low-pass filter, very noticeable artifacts disappear, but the overall sense of clarity in low-contrast scenes also decreases.
3. when using a low-pass filter and Snell & Wilcox hardware noise reduction, the overall result is a good clean picture, but sometimes there is a certain feeling of plasticine (however, there are still artifacts in some scenes)

It is worth mentioning here that encoding was carried out in MPEG2, H.264 and VC-1 formats using a two-pass software codec with the possibility of subsequent fragment-by-fragment recoding to optimize quality. The codec was created by one very well-known company specifically for professional authoring of Blue-Ray discs and, in our opinion, works great for a compression ratio of 1:50 (probably only a black square can have such redundancy).
The result we obtained is almost independent of the use of various encoding formats (MPEG2, H.264, MS-1). If the signal quality is high, then the encoding quality differs minimally from format to format. Codecs behave the same way during aggressive preprocessing. The difference only appears at extremely low bitrates, where the resulting HDTV quality can be distinguished from HDTV quality only by the frame size.
Still, I must admit that when using H.264 the quality was slightly better. But, most likely, without comparison, the average consumer on a consumer monitor is unlikely to be able to see the difference in quality, say, compared to MPEG2.
And here it is worth recalling that this was a signal received from the TSC.
During test encoding of the XDCAM HD signal (Mbit/sec, 1440X1080i50), for example, the situation developed noticeably worse - artifacts during recoding with a smaller stream on complex fragments increased sharply (flash flashes at a fashion show). This suggests that when filming these types of events, you should always be mindful of secondary shots that minimize the overall amount of movement in the frame.
In general, there is a rather strange and not very simple way to get the illusion of uncompressed video from compressed video, at least for some kind of deception of the codec. This method is a type of post-process. By the way, this is exactly what they do on the BBC, Discovery Channel and a number of other foreign channels, for which quality is a trademark, both in terms of image quality and in terms of program content. The video is stylized, as a result of which you change as much as possible dynamic parameters pictures. At one time, this is how we solved the problem of using DVCAM to create advertising materials under relatively large screens using DVD. All this, of course, is not a matter of five minutes, but with a well-established process, the main thing is to approve the direction of styling with the customer.
Very often you can hear that the post-process is very expensive. Unfortunately, in order to look adequate in the video production market today, you have to do this too. Well, if you have customers for advertising materials who do not demand bluer skies, greener grass and tanned faces even more tanned, then store, protect and please them. They are unique and just lay golden eggs.
So let's get back to threads.
Of course, the most important aspect for reducing flow is resolution. AND digital television it just seems like it was created for these tricks. Usually no one ever mentions this possibility.
Yes, indeed, the human eye sees not the number of pixels, but the dynamic range. And therefore, everyone is very careless with resolution, especially since household televisions in the digital frequency range often show no more than 2-3 MHz. There are many examples of forced resolution reduction. Pixel-shift in video cameras, resolution reduction even in professional video formats, and simply broadcasting on air and satellite channels with a significant reduction in frame size. When viewed by a consumer, any high-frequency artifacts are still masked by the low bandwidth of the TVs themselves.
Based on all this, it becomes clear why in the United States the average stream for non-sports HDTV is 19 Mbps (based on data from viewing video clips from hacked channels). Maybe this average value is less, but apparently then no one hacks low-flow channels and posts content on the Internet, which in itself is indicative.
If we assume that uncompressed video is used at the input, plans with high texturing are adapted, and complex backgrounds are not used, then it is probably possible to obtain a generally tolerable output image with such a stream (but still, not for an LCD display with a diagonal 52 inches).
Probably, by following some rules and reasonably degrading the quality at the codec input, you can achieve quite decent quality of non-sports video, also using H.264 at a stream of about 13 Mbit/sec at 720p25 (the stream is also smaller due to the fact that in Russia there are not 30 , and 25 frames per second). But still, the most important thing will be the size of the screen and the distance to the viewer.
Since we also have to film concerts, according to our current understanding, the stream for further encoding of concerts (against the backdrop of LED screens and with metallic confetti falling on top) should ultimately be no worse than 35 Mbit/sec for the consumer. Naturally, when encoding from a signal with minimal compression. This will allow you to watch such video signals on screens probably up to 52 inches from a distance of 4-5 meters without very noticeable artifacts (although low-pass filtering will have to be selected in any case).

In general, if we switch to HDTV, then first of all it is necessary to solve the main question - we want to make sure that people have HDTV in their homes, or so that HDTV is available only in our television centers and studios, and viewers have only what that they will call HDTV with bewilderment. Having chosen the direction in solving this dilemma, we will choose the appropriate flow.
A small stream, of course, has a lot of advantages. You don’t have to think much about the content, because upconverting from TSC is quite suitable; when shooting, you don’t need to spend money, for example, on expensive makeup (if you don’t shoot a lot of detailed and close-ups), anyway, compression will kill the dynamic range in skin tones and etc. And the main thing is that a small stream is easier to transmit and easier to store. But you won’t be able to save money on lighting equipment, although TV shows have long been trying to convince us otherwise.
Maybe for a free channel, low flow is a forced decision. But not for paid ones. Probably, in order to force viewers to pay for 8 Mbit/s, you first need to show people a 0.5 Mbit/sec stream in HDTV for two years, take away all their home camcorders, turn off the Internet, and only then give them such a low-stream HDTV.
By the way, I personally think that it is the Internet that today determines the promotion of HDTV in our country. On the Internet you can find a large number of films and music programs in various HDTV formats. Considering that they are watched on computer monitors, the requirements for streams of this content are low, since these are mostly monitors no larger than 24 inches. In the future, the quality of these video materials from the Internet will increase up to the original, corresponding to the satellite source or Blue-Ray disc. It is possible that with the incoming Internet speed of about 6 Mbit/sec. An advanced user will simply leave the computer to download when he goes to work, and when he comes home from work for dinner with beer, he will watch a freshly downloaded movie in Blue-Ray quality. In Moscow today this is already a reality. Cost – monthly fee only unlimited internet, and of course, brains working to get freebies. But ingenuity in Russia has always been fine. For those who are not strong in Internet technologies, there are always friends who will happily send you a mobile phone hard drive what they downloaded in the past 24 hours. This is also purely our Russian peculiarity.
And yet, it must be said that the supply of high-quality video signals to the average consumer is apparently not held back by technical aspects. More like marketing.
In the era of “developed standard definition television”, in which we now live, the quality availability scheme looks something like this:

1. signals formerly called "broadcast quality" (from uncompressed to 25 Mbit/s)
2. DVD (maximum stream 8-9 Mbit/s, software codecs high quality, encoding from uncompressed video signals or Digital BETACAM)
3. Satellite digital streams (the stream is usually no more than 6 Mbit/s and reaches up to 1.5 Mbit/s, hardware codecs or real-time software codecs)
4. Broadcasting (any level at the input, but the quality of the viewer is determined mainly by the quality of the distribution channel)

If, suddenly, we switch to HDTV, the structure of quality availability will remain the same. But visually, due to existing trends, the image quality will most likely only become worse; only the screen size will change.

By the way, for reference. When you come to a store to buy an LCD panel, in the sales area, as a rule, with hard drive reproduce video material provided for demonstration purposes by one of the LCD display manufacturers. We were able to obtain these files from various manufacturers. So, producers not only select a special video sequence, but also carry out specific post-production to minimize encoding losses. This can be seen upon closer inspection. And we did not find a stream below 25 Mbit/sec in these files, most are 36-38 Mbit/sec. This is such a concern that we purchase the best and do not doubt the quality of the displays.
There is another interesting observation.
The maximum flow for a Blue-ray disc is about 40 Mbit/sec. But with streams ranging from 30 to 38 Mbps, we were only able to see demo discs. All commercial releases are 15-22 Mbit/s. Maximum flow peaks are no more than 30 Mbit/sec. Again, this is only based on those streams that fell into our hands. Comments are, as they say, unnecessary.

If you decide to broadcast live, you need to prepare for it in advance. Find out the outgoing bandwidth of your Internet connection and select settings that will allow the broadcast to go smoothly. You can check the download speed using special online services.

If you create a broadcast in the Broadcast Control Panel or in the "Start Broadcast" section, the system will automatically detect what settings are specified in the video encoder. You will only need to specify the resolution, frame rate and bitrate.

If you schedule a broadcast in the “All Broadcasts” section, you can independently set the resolution and frame rate. Another option is to select a broadcast key and let the system determine the settings for you.

The broadcast will be automatically re-encoded: thanks to different output formats, it can be watched on any device and regardless of the network to which you are connected.

We advise you to conduct a test broadcast - this will allow you to make sure that the image and sound are broadcast without interruptions. After starting the broadcast, keep an eye on the control panel: it will display all error messages and data about the quality of the stream. given full list possible problems.

Note. For video in 4K / 2160 pixels. You cannot configure a short delay. For such broadcasts, the standard delay will be automatically set.

Variable bitrate and custom broadcast keys in the Broadcast Control Panel

If you use your own broadcast keys, you can select the option Variable bitrate, and then the video encoder will automatically set the resolution. This value can also be set manually.

4K/2160p (60fps)

• Resolution: 3840 x 2160
• Video stream bitrate range: 20,000–51,000 kbps

4K/2160p (30fps)

• Resolution: 3840 x 2160
• Video stream bitrate range: 13,000–34,000 kbps

• When planning your broadcast, be sure to check the "Enable 60 fps mode" checkbox on the "Broadcast Settings" tab. If you created a broadcast in the "Start Broadcast" section, the frame rate and resolution will be selected automatically.
• Resolution: 2560 x 1440
• Video stream bitrate range: 9000–18,000 kbps

• Resolution: 2560 x 1440
• Video stream bitrate range: 6000–13,000 kbps

• When planning your broadcast, be sure to check the "Enable 60 fps mode" checkbox on the "Broadcast Settings" tab. If you created a broadcast in the "Start Broadcast" section, the frame rate and resolution will be selected automatically.
• Resolution: 1920 x 1080
• Video stream bitrate range: 4500–9000 kbps

• Resolution: 1920 x 1080
• Video stream bitrate range: 3000–6000 kbps

• When planning your broadcast, be sure to check the "Enable 60 fps mode" checkbox on the "Broadcast Settings" tab. If you created a broadcast in the "Start Broadcast" section, the frame rate and resolution will be selected automatically.
• Resolution: 1280 x 720
• Video stream bitrate range: 2250–6000 kbps

• Resolution: 1280 x 720
• Video stream bitrate range: 1500–4000 kbps

• Resolution: 854 x 480
• Video stream bitrate range: 500–2000 kbps

• Resolution: 640 x 360
• Video stream bitrate range: 400–1000 kbps

• Resolution: 426 x 240
• Video stream bitrate range: 300–700 kbps

Video encoder settings

Beginning users are often interested in what video bitrate is, how to calculate it correctly, and why it is needed at all. Well, let's answer this question. Bitrate, or otherwise the width of a video stream, is the amount of information transmitted or processed in one second of real time. The bitrate is measured in kilobits per second and is denoted as kbps. Please note that in kilobits, not kilobytes. A kilobit is 1/8 of a kilobyte.

The more useful information transmitted in a video stream per unit of time, the higher the video bitrate and, accordingly, the better its quality. At the same time, the higher the bitrate, the larger the video file size. From here it becomes clear why you even need to calculate the bitrate when converting. Calculating the bitrate is necessary to achieve the optimal balance between video size and quality.

Let's say you need to record to standard DVD a video file whose size exceeds the disk capacity, while maintaining the format, aspect ratio and resolution of the image. In order not to make mistakes when converting and not to compress the video too much or, on the contrary, "put the squeeze" it to the required size, calculation is required.

How to determine bitrate

To determine the bitrate, it is advisable to use special utilities, of which the best is.

This is very powerful tool, allowing you to receive the most detailed information about audio or video. You can also calculate the bitrate manually by dividing the video file size in kilobits by its playback duration in seconds. Let's go back to our DVD. We have a film size 5.2 GB and it needs to be recorded on a blank 4.7 GB . What bitrate should I set in the converter?

Let's do the calculation. Let's assume that the length of the movie is 2 and a half hours or 9000 seconds, the actual capacity of the DVD is approximately 4480 MB. Let's use the following formula:

(MBs) * 8000 = Result

That is, we divide the disk size available to us by the time in seconds, and then convert the received data into kilobits.

(4480/9000) * 8000 = 3982 kbps

Everything seems to be correct, but we did not take into account the audio stream. Its bitrate also needs to be calculated. That is why in such cases it is better to use special utilities - bitrate calculators, which are available both as native applications and as online services. In them you can set the size final file and, by specifying the duration of the video and the quality of the audio track, get the video bitrate, which needs to be entered into the converter settings.

Note: if the converter has a function for calculating the size of the final file, take the time to make sure that the bitrate calculations you made correspond to the expected video size.