Transcoding an Input Stream

zixi broadcaster offers live transcoding, enabling users to create multiple bitrate streams from a single input stream supported video and audio formats include video decoding – mpeg2, h 264/avc , h 265/hevc video encoding – mpeg2, h 264/avc , h 265/hevc audio decoding – aac, mpeg 1 audio layer i / ii / iii, mpeg 2 audio layer i / ii / iii, ac 3 (dolby digital)] for ac 3, which can contain up to six discrete channels of sound, the transcoder only supports the left and right channels and will ignore the other channels audio encoding – aac, aac he (high efficiency) and aac hev2; pass through (no encoding) – any audio codec (mpeg1/2 / aac / ac 3, e ac 3 (dolby digital plus) the transcoder comes with pre set common video profiles 1080i60 / 1080i59 94, 1080p30 / 1080p25, 720p60, 720p50 / 720p30, 576i60 / 576p25, 480i60 / 480p29 97 in order to transcode an input stream from the inputs screen you must first configure the transcoding profiles in the transcoder screen for more information, see broadcaster transcoder docid 7l8zpt7uyouyl6oseunzl section to transcode an input stream in the zixi broadcaster inputs page, click the desired input stream that you want to transcode click the ⚙️ icon on the right end of the selected row and then transcode the transcode < input stream name> window with the settings of the selected input stream is displayed enter the unique stream i d for the new transcoded input stream under stream parameters , in the video profile and audio profile fields, do one of the following select the desired video profile and/or audio profile that you have configured select keep original stream if you have not configured a video or audio profile or do not want it transcoded or you want to set specific transcoding parameters for the specific stream in this case, the video or audio input will not be transcoded select remove stream – you can use the transcoder to remove the video or audio track from the stream fill in additional fields (see table below for additional parameters) click ok the transcoded input stream is created and appears in an embedded window labeled transcoded sources under the original input stream from which it was created stream parameters parameter description advanced options force software decoding if the input video source comes in a 4 2 2 format, which is not supported by the nvidia gpu decoder, select this option so the decoding will be performed on the cpu (software decoding) and the encoding will be performed on the nvidia gpu video profile select an existing profile, select keep original stream if you do not wish to transcode the video stream, or select remove stream if you want to remove the video track audio profile select an existing profile, select keep original stream if you do not wish to transcode the audio stream, or select remove stream if you want to remove the audio track apply audio profile for all audio pids select this checkbox to apply the selected audio profile for all pids output type select one of the following output types the outputs are configured as part of the transcoding process, so there is no need to create the output from the zixi broadcaster outputs screen mpeg ts – the output will be available as mpeg transport stream ndi – the output will be available as ndi when selecting ndi output, you can select the following video/audio profiles video profile – remove stream or a raw profile audio profile remove stream or a raw profile webrtc – the output will be available as webrtc when selecting ndi output, you can select the following video/audio profiles video profile – remove stream, keep original stream, raw video profile, or if the input stream is not webrtc compliant, a transcode to webrtc compliant profile (e g h 264 codec with baseline or main profile and no b frames) audio profile remove stream, keep original stream, raw audio profile, or if the input stream is non opus, transcode to opus transport stream bitrate (kbps) specify the bitrate of the mpeg transport stream (ts) this parameter is optional if you do not specify a ts bitrate, the transport stream will be variable bitrate (vbr) with no additional null packets added if you specify a bitrate, the transport stream will be constant bitrate (cbr) where null packets are added as necessary to achieve the specified bitrate cbr transport streams are typically required when sending a stream to a satellite head end or to a professional integrated receiver decoder (ird) the overhead of mpeg ts is about 7% to determine the bitrate setting, sum the max video bitrate and audio bitrate from the encoding profile and then add the other pids (if choosing to pass all pids) and add 7% if the bitrate needed for the transport stream exceeds the value specified, the broadcaster will give a “bitrate too low” error message vbr transport streams are typically used for hls, dash, rtmp, and other similar live streaming workflows pass all pids select this checkbox to pass through all non transcoded pids add rtp headers select this checkbox to add rtp headers to the stream rtp adds a 12 byte header, which includes a sequence number and a timestamp that can be used to detect dropped packets the sequence number in the rtp headers can be used to do automatic reordering and the timestamps can help overcome inter packet gaps, also known as iat (inter arrival time) however, plain rtp does not feature error recovery or error protection by default, the rtp headers (which include the timestamps and sequences) are being used on the input side, but then they are stripped off and they no longer propagate to outputs force i frames on scte 35 markers select this checkbox to force an i frame on the frame with the pts mentioned in the scte35 message or on the first frame after that pts, if it doesn’t exist in the stream copy closed captions select this checkbox to copy cc (if exist on original stream) to the transcoded stream applicable to hevc/h264 sources and profiles only crop parameters select this checkbox to set the cropping parameters cropping mode set the units used for defining the crop options are pixels or percentages note depending on the type of content, in some cases using different crop values for low resolution outputs might be appropriate for example, if the content includes both video content and textual content, the text would not be readable at lower resolutions and thus might be better to exclude another use case for using different crop values for each output would be to slice a 360 or vr input video left, top specify the left and top positions for the crop width x height specify the width and height values for the crop dynamic overlay url the url of the html overlay feed this field will be displayed only if the cef software was installed the html overlay feature overlays a dynamic html page on the video stream during the transcoding process as part of this process, the images from the html page are “burned” into the video the process uses chromium embedded framework (cef) to render the html page and then the image is passed on to the transcoder to overlay on top of the raw video frame, which is then encoded for more information, see html overlay installation guide overview docid 6snc7fcknryvmtgzr06kb smoothing buffer (ms) specify how much of the transcoded stream (in milliseconds) to store before transmitting it back to the server in order to prevent bursts if the transport stream bitrate \[kbps] field is empty, the smoothing buffer is disabled if the transport stream bitrate \[kbps] field is not empty, it is recommended to enter around 200ms, which should be enough to prevent bursts enable time shift select this checkbox to record the stream content and broadcast it after a specified delay to deploy an input stream that uses the current stream's time shift, you need to configure a "delayed" input stream (see adding delayed streams docid 9yv8fx2x pdvy2qibluz3 ) maximum delay \[hh\ mm\ ss] specify the time of the delayed transmission of this stream maximum 24 hours transmit as multicast selecting this checkbox enables the transmission of this stream’s outputs as multicast a zixi receiver that will pull the stream will receive it in multicast by default zixi broadcaster is configured to allow transparent fallback to unicast if it is out of the lan requires enabling multicast pull in the settings this setting requires enabling multicast pool (settings > multicast pool) multicast only selecting this checkbox will force the transmission of this stream only in multicast enable encryption select this checkbox to encrypt the input stream for more information, see encrypting an input stream in zixi broadcaster docid 5sh sqkz1lgkmgzjsqrzq encryption type (for encryption enabled) specify the type of encryption (aes 128/192/256) encryption key (for encryption enabled) click generate to generate an encryption key the generated encryption key must be sent to the end user to decipher the received encoded stream recording parameters select this checkbox to customize the recording parameters destination type select the radio button for the desired storage type and then fill in the relevant parameters · record to disk – the recording is saved on the local disk · record to s3 – the file is saved to your aws s3 bucket record to disk parameters file name template specify the file name pattern that zixi will use to generate when saving the recorded content a separate file will be generated for each recording up to the time specified in the max recorded file duration parameter (see above) after this duration a new file will be generated the file name pattern can include any string and up to five variables listed below only the %t=hh mm ss creation time is mandatory you can change the order of these variables and add text as long as you maintain the legitimate structure of the file in linux and windows · %s=stream id optional · %y=year optional · %m=month optional · %d=day optional · %t=hh mm ss creation time mandatory for example, the following pattern %y recording%t ts will generate the following file name 2017 recording15 32 35 ts max recorded file duration \[hours] specify the maximum time allotted for recording a video stream default 2 hours keep recorded files for \[hours] specify the maximum time for storing the recording default 0 hours custom path optionally specify a storage location for the recorded files that is relative to the root folder this location bypasses the location specified as the root folder in the settings > general screen record to s3 parameters file name template specify the file name pattern that zixi will use to generate when saving the recorded content a separate file will be generated for each recording up to the time specified in the max recorded file duration parameter (see above) after this duration a new file will be generated the file name pattern can include any string and up to five variables listed below only the %t=hh mm ss creation time is mandatory you can change the order of these variables and add text as long as you maintain the legitimate structure of the file in linux and windows · %s=stream id optional · %y=year optional · %m=month optional · %d=day optional · %t=hh mm ss creation time mandatory for example, the following pattern %y recording%t ts will generate the following file name 2017 recording15 32 35 ts max recorded file duration \[hours] specify the maximum time allotted for recording a video stream default 2 hours url specify the url of the s3 bucket ignore tls certificate errors tls certificate for s3 bucket might be detected as faulty if bucket name contains dot\[s] when this option is selected, zixi broadcaster will ignore the tls certificate errors and transmit the stream to its destination access key the access key that is used for accessing the s3 bucket secret key the secret key that is used for accessing the s3 bucket