Frequently Asked Questions

Which encoders are available?

For now, we focus on reference models implementation of actual and future codecs: AVC, HEVC, VVC, AV1, AOM post-AV1 codec (AVM), and AVS3. More codecs to come!

Which decoders are available?

Reference model implementations that correspond to the above-mentioned encoders reference models.

Also, decoding by our quality measure tool is available too. More details here.

Which metrics are available?

PSNR, PSNR-HVS, SSIM, MS-SSIM, VMAF, VMAF-NEG, CIEDE2000. VQProbe is used under the hood.

Is it possible to get raw data of session results?

Yes. For your sessions, you could always get raw results. For our data, it could be purchased separately.

Can I use my own codecs and streams?

It’s not possible currently, but we plan to add it in future releases. Sign up to keep updated!

What streams are available?

We use AOM CTC stream set from https://media.xiph.org/video/aomctc/test_set/ as well as several well-known open-source streams like Big Buck Bunny and Tears of Steel.

What is the “create” service?

This service allows you to run your own custom pipelines: specify codec and its config, stream set, and other parameters based on your interest.

What is the “compare” service?

This service allows building comparative quality reports based on available data: public or private ones shared with you.

Stream gallery

It is a database of available media streams to run pipelines on. There is general information about them, and you can also download them.

BDRate calculator

It is a small handy service in which you can set your bitrate/quality data as well as get RD-curve and BD-rate values.

Workspaces

It is an entry point to navigate across all running sessions and their statuses.

What are the quality and smoke tabs in the workspace?

Currently, we support two types of pipelines:

• “smoke” is a fast and small sanity check pipeline to check the correctness of a new unique session before actual execution.
• “quality” is a general working session to check the quality performance of a video codec.

Dashboard

It is an entry point for all the information and statistics related to your account. You get various spending statistics and may top up your balance. Also, favorites are placed there.

Is there any pre- or post-processing on streams?

No, the stream passes as is into the encoder without any processing.

What could be configured for the pipeline?

You could configure the following:

1. Encoder implementation parameters: coding tools, rate control, and so on;
2. Stream set;
3. Values of variable parameter for RD-curve points;
4. Quality metrics to be used.

What configurations are used in the pipeline?

We have prepared RA and LD configs used from the relevant CTC documents. You may use this as a template for your own configs.

How to comprehend that my config is correct?

There is a mandatory test run on a very small data set before the actual execution of the pipeline to check that configuration is viable.

How is the codec structure organized?

As a basic entity, we work with specific codec implementation, which could have one or more revisions – release tag, build commit hash, etc. Each codec implementation is associated with one or more configurations which can be used as a template for your own configs.

And on the top level, we group implementations by their standards for better navigation.

What are private and public sessions?

Private sessions are visible only to the owner or within groups/users with whom it’s explicitly shared.

Public sessions are visible to everyone.