Hi.
Here’s an example of using HQX as an intermediate master file for uploading to YouTube. For this video I set HQX to its maximum output quality settings with a resolution of 4K/UHD at 25FPS.
While it’s safe to say that it’s very diminishing returns to use such a high quality codec for YouTube, it does however provide the opportunity for the best possible transcode at YouTube’s end.
I’ve done many such tests over the years and the results are that HQX will provide the exact same visual quality compared to using a completely uncompressed codec.
The results of using HQX for such things can be bettered when using intra-frame source material as there is no compounding of the temporal redundancy characteristics from inter-frame sources to the final inter-frame output from YouTube when using an inter-frame intermediate.
Although the source codec here is inter-frame, using HQX and its intra-frame properties, does remove another inter-frame stage during the workflow. In this instance HQX, in practice, removes one more destructive stage in delivery. The padding out of the source’s limited colour and temporal information into HQX effectively makes the HQX transcode visually identical to the source. Which although doesn’t negate the limits of the source, it does however add extra robustness for future post manipulation.
I’ll show another similar example at some point where I use high quality 4K intra-frame source material within a HQX delivery. There’ll also be an example of CG/game play captured 4K/UHD using HQX and delivered from the capture file.
Like I’ve already mentioned, this type of workflow is subject to diminishing returns with the final YouTube transcode process, which includes the reduction to 8 Bit for any rec.709 10 Bit deliveries and the extremely low bit rate inter-frame outputs. There’s also the large file sizes of a HQX master to be considered but using HQX does represent an option for those seeking absolute video quality when delivering to YouTube.
Despite the small visual gains with such a workflow, there’s a very practical outcome to delivering 4K/UHD and higher resolutions to YouTube. With these types of high resolution deliveries, regardless of the codec used for the intermediate. YouTube will always use VP9 for the main streaming codec for all down scaled outputs from such a delivery to any compatible client, be it via software such as a browser or app or hardware such as a streaming device. This benefit is very important as it avoids H.264 for the lower resolutions, including 1080. And while a comparative VP9 stream uses a slightly less bit rate compared to the already very low bit rate of H.264. The VP9 stream does benefit from better motion prediction and other better pre and post processing, that results in a better looking encode that very noticeably helps with complex picture detail and motion.
On that last point I’ll be making a tutorial soon showing the benefits of using Edius and Lanczos to up scale a 1080 project to 4K/UHD, in order to gain the above benefits for 1080 YouTube streams. While this may sound counter intuitive due to what would appear to be the unnecessary processing of the up and down scaling and the reasonable assumption of the compounding negative effects, the process does indeed work extremely well.
Anyway, here’s the example of the HQX 4K/UHD upload. The file is best watched on YouTube with the 4K/UHD stream option in full screen.
Cheers,
Dave.
Here’s an example of using HQX as an intermediate master file for uploading to YouTube. For this video I set HQX to its maximum output quality settings with a resolution of 4K/UHD at 25FPS.
While it’s safe to say that it’s very diminishing returns to use such a high quality codec for YouTube, it does however provide the opportunity for the best possible transcode at YouTube’s end.
I’ve done many such tests over the years and the results are that HQX will provide the exact same visual quality compared to using a completely uncompressed codec.
The results of using HQX for such things can be bettered when using intra-frame source material as there is no compounding of the temporal redundancy characteristics from inter-frame sources to the final inter-frame output from YouTube when using an inter-frame intermediate.
Although the source codec here is inter-frame, using HQX and its intra-frame properties, does remove another inter-frame stage during the workflow. In this instance HQX, in practice, removes one more destructive stage in delivery. The padding out of the source’s limited colour and temporal information into HQX effectively makes the HQX transcode visually identical to the source. Which although doesn’t negate the limits of the source, it does however add extra robustness for future post manipulation.
I’ll show another similar example at some point where I use high quality 4K intra-frame source material within a HQX delivery. There’ll also be an example of CG/game play captured 4K/UHD using HQX and delivered from the capture file.
Like I’ve already mentioned, this type of workflow is subject to diminishing returns with the final YouTube transcode process, which includes the reduction to 8 Bit for any rec.709 10 Bit deliveries and the extremely low bit rate inter-frame outputs. There’s also the large file sizes of a HQX master to be considered but using HQX does represent an option for those seeking absolute video quality when delivering to YouTube.
Despite the small visual gains with such a workflow, there’s a very practical outcome to delivering 4K/UHD and higher resolutions to YouTube. With these types of high resolution deliveries, regardless of the codec used for the intermediate. YouTube will always use VP9 for the main streaming codec for all down scaled outputs from such a delivery to any compatible client, be it via software such as a browser or app or hardware such as a streaming device. This benefit is very important as it avoids H.264 for the lower resolutions, including 1080. And while a comparative VP9 stream uses a slightly less bit rate compared to the already very low bit rate of H.264. The VP9 stream does benefit from better motion prediction and other better pre and post processing, that results in a better looking encode that very noticeably helps with complex picture detail and motion.
On that last point I’ll be making a tutorial soon showing the benefits of using Edius and Lanczos to up scale a 1080 project to 4K/UHD, in order to gain the above benefits for 1080 YouTube streams. While this may sound counter intuitive due to what would appear to be the unnecessary processing of the up and down scaling and the reasonable assumption of the compounding negative effects, the process does indeed work extremely well.
Anyway, here’s the example of the HQX 4K/UHD upload. The file is best watched on YouTube with the 4K/UHD stream option in full screen.
Cheers,
Dave.
Comment