3
Planning and Preparation

The Boy Scouts have it right: Always be prepared. Many first-time compressionists assume that preparing to compress video means choosing your equipment and firing it up or shooting and editing the video you intend to compress.

These are important topics, but the most important step you need to take before you start compressing video is to identify what type of compressionist you are. Taking a moment to ask yourself a few questions and understanding the various requirements of the tasks you’ll need to accomplish will allow you to make better choices along the way. Once you’ve established your position in the production pipeline and what kind of compression you need to do, you’ll be able to come up with a few workflows for how best to get your compression work done. Even though all compressionists ultimately use similar setups, different jobs have different needs, so in this chapter you’ll take a look at some of those needs. Once these preliminaries are out of the way, you can then turn a critical eye on your equipment and make some decisions about what gear you’ll need to compress your video effectively.

Once you do all that, you’ll be well prepared for video compression. In this chapter, we’ll also offer some suggestions for optimizing both your equipment and your workflow.

What Are Your Compression Needs?

Different types of content have different requirements in terms of how the content is handled and the type of equipment you’ll need to use. So, let’s figure out what type of compressionist you are and what kind of compression you’ll be doing by answering some questions.

Final Delivery or Work in Progress?

The first question to ask is where in the production process the majority of your work will be done. Will you be compressing video for the final delivery of finished titles? Are you facilitating viewing of works in progress? Both? Understanding a bit of the big picture will help you optimize your work for the task at hand. It’s useful to start your work by thinking about the end result and working backward. If your role includes delivering final outputs, you may tune your process toward delivering high quality at low data rates or efficiently encoding for multiformat delivery. If most of your work is distributing screening copies of work in progress, your process may be optimized for speed and security.

Are You the Master of Your Destiny?

How much control do you have of your decision-making process? If you are working as a one-man band, you’ll have freedom and flexibility that compressionists working in a more collaborative environment may not have. Are you generating your own content, or will someone else be handing it to you? Are you part of an established workflow, or are you creating a new pipeline from scratch? Knowing what external requirements that you don’t have control over will help inform the decisions for things you can control.

Long-Form or Short-Form Content?

Once you’ve established where your work exists in the production process and how much flexibility you have, determining the length and type of content will help you make some basic decisions, especially once you also know the target delivery medium. Different lengths of video lend themselves better to different delivery platforms. Generally speaking, video meant to be viewed in a web browser is shorter than video meant to be viewed on a TV. This is not a hard-and-fast rule, and every year the lines between online and offline delivery systems get a little fuzzier.

Let’s say you are planning on delivering long-form content that is meant for a progressive download player on the Web. You will have to factor in length and encode the video at a much lower data rate than you would for short-form content. This will help keep the file size down to a manageable download. After some experimentation, you might find that your long-form show is better suited to an adaptive bitrate streaming platform that requires specialized encoding tools. All of these considerations are length-dependent and may change dramatically if the total run time changes too.

What Is Your Final Delivery Target?

Knowing your target delivery medium will help you determine the video quality limitations of the final product. You may be asked to deliver for broadcast, theatrical, web, mobile, optical discs like Blu-ray and DVD, or some combination of those. Most likely, you are not just compressing your video for one particular targeted delivery. Identify all the places to which you’ll ultimately deliver your content and then determine what specific needs each destination has.

How Is Your Content Delivered to You?

If you’re working in a collaborative environment where content is not originating with you, you’ll likely receive content you haven’t seen before.

There are a lot of ways to deliver content to a compressionist. In some production environments (broadcast, for example), the most common delivery mechanism may be old-fashioned videotapes like Digibeta for standard definition, or HDCAM-SR for 1080i. In other realms, you may be asked to work with a digital source file, which might be delivered on a drive or over the Web. The medium used to deliver content to you matters for several reasons.

First, you must be prepared to accept the delivery format your clients or colleagues commonly use, and then you also have to know the practical limitations of it.

Do you have the right equipment to accept your content source’s format of choice? If not, you need to ask for something else or budget time into the process for troubleshooting issues that may come up when you try to access the delivered source material. For example, your client may prefer using Mac-formatted drives, but you only have a Windows PC that cannot read the Mac drive. You may need to buy special software to read the Mac-formatted disk or ask for the delivery drive to be formatted using a cross-platform file system before it gets to you.

Is the source material of high-enough quality for the output destination? Video compression is the art (and science, but mostly art) of eliminating information. If you don’t start with enough information in the source, it doesn’t matter what sort of golden touch you have; your output results will always be poor. This is a game of garbage in, garbage out. Make sure the client is delivering a file or tape format that is high-enough quality that you can encode a decent finished deliverable from it.

How Much Content Are You Compressing?

How many minutes or hours of content are you working with actively in a given week? What about at one time? Do you have enough equipment to do the workload?

Overcommitting your time and equipment will lead to delayed or rushed results. Determine how much work you have to accomplish and then make sure you have the appropriate amount of hardware and software available to accomplish it.

Some video clips take longer to compress than others, so budget your time based on the length of video, ingest format, export format, and complexity of compression to decide whether you can keep up with your workload. Unfortunately, there isn’t a fast or easy way to determine this—you have to learn the capabilities of your setup over time.

Is Compression All You’re Doing?

Where does compression fit into your job description? Is it your only responsibility in the project at hand? Are you also playing some other role? If you are going to be doing things other than encoding video, where does video compression rank in your job priorities?

Quality or Timeliness?

Which is more important to you and to the client: quality or on-time delivery?

There’s much more to video compression than subtracting or reassigning bits, and oftentimes the better the compression, the more difficult it is for the computer to encode. The tougher the encode, the longer it will take. When it comes to crunch time (and you will find that almost every project has hard deadlines), do you or your client simply need all the work done in a specified time frame, or is it better to take an extra few hours or days to get the right results? Rushing to complete a project when it will only be rejected because the results are suboptimal could mean you’ll be doing the work twice (and maybe getting paid only once).

How Knowledgeable Are Those with Whom You Work?

A little knowledge is super dangerous in any field, and compression is no exception. Some clients know a little of what they want, or their knowledge may be out of date.

This kind of knowledge usually does more damage than good. There are lots of examples, but a favorite is when a client asks for an “uncompressed H.264” file. Both of those are actual video compression terms, but they are contradictory. You might have to read between the lines or ask additional questions to figure out what a client or colleague is asking for. Help educate them so they can work with you to get what they need, instead of making arbitrary demands based on poor, outdated, or incomplete information. This, of course, means you need to be up to date and knowledgeable, too—which is as good a reason as any for you to be reading this book and asking these questions well before your actual compression work begins.

You may find that people have an idea of the results they ultimately want but not the specifics of how to get them. You as the compressionist—and, consequently, the person most knowledgeable about the compression process—should be in a position to make suggestions to the client or colleague who needs a particular end result. If you see something that doesn’t make sense, say something. You may be surprised to discover you have information that others in the chain may not have that will make the process easier or the results better.

Equipment and Workflows for Different Scenarios

Now that you’ve taken the time to establish answers to a lot of questions about the work you’ll be doing, let’s look at how to equip yourself for the job to be done. Though the equipment may not vary much from one compressionist to another, it’s good to identify the type of setup that will best suit your work needs. The following sections highlight some real-world scenarios to illustrate how to match encoding equipment to the demands of the job at hand.

Encoding as an Editor

Let’s say you’re an editor and you need to encode video. You might be sharing rough cuts with a client or delivering dozens of finalized short clips meant for the Web. Perhaps you’re a blogger packaging your own content for YouTube. Regardless of the reason, you’re working with video in a nonlinear editing (NLE) system, and your ultimate goal is to prepare content to go somewhere outside your edit suite. If this is the case, then video compression is almost always a second- or third-tier priority in your workflow. If you’re busy trying to carefully edit a video, get all the titles finished, and ensure the audio levels are good, chances are by the time you’re ready to encode, you will be running low on mental energy. Compression is an afterthought—something you do while you get up and stretch after a long editing session.

In this situation, you probably aren’t encoding heaps of content at once. Most likely you will edit a sequence together, encode it, send it off for review or to the next postproduction stage (whether that’s the graphic artist, the colorist, or the sound editor), and then carry on editing. The good news is that if you are using an NLE system, you don’t really need much more specialized equipment. Your NLE system will feature encode/export presets for all kinds of contingencies (Figure 3.1) and usually will have the ability to tweak them to your specifications if needed. One word of caution: Make sure you have enough disk space. Editors tend to use a lot of storage for source material, media caches, and renders. Confirming you have enough space on your drive before you export your compressed version will save you a lot of head- and heartache.

Figure 3.1 Here’s where you choose encoding options in Apple Compressor, the compression tool available from Apple’s App Store.

If you are a hobbyist just editing together home movies, you may not want to invest in an overly expensive setup. In fact, you may not have to invest any money at all. Most Macs and Windows machines ship with video-editing software that allows you to export compressed video suitable for many different purposes. These tools come with some video-compression capability—usually the ability to encode video for Windows Media Player or QuickTime Player, plus Blu-ray and DVD authoring. Your encoding options might be limited to good/better/best or similar categories, but this is enough to get you started compressing video that’s reasonably suited to your output choices.

Maybe you’re getting paid to edit and need a bit more than the free tools provide. Or perhaps you’re just the type of avid hobbyist who wants the best equipment possible. Then you’ll want gear categorized somewhere between the inexpensive consumer equipment and the high-end professional stuff. This in-between tier is called prosumer. Equipment manufacturers and software creators love prosumers because they tend to spend more on equipment than the average hobbyist-type consumer. Prosumer equipment tends to have a better set of features but not quite as many as a full-time professional might need (Figure 3.2).

Figure 3.2 Transcode settings in Adobe Media Encoder CC, the encoding tool in Adobe Creative Cloud.

Maybe you’re a professional editor and your equipment reflects it. You might have somewhere between $25,000 and $200,000 parked in a climate-controlled edit suite (or you just rent other people’s gear). You might even have multiple systems in your studio—the high-end setup for high-paying clients and maybe a smaller setup on a laptop that you use to bang out jobs on the go (Figure 3.3).

Figure 3.3 Professional editors may simply choose to use the default export settings within their NLE system of choice to expedite the compression process.

No matter the situation, in all these cases, you’re an editor first and a compressionist second (or third or fourth). Even though it might not be the top priority, you’ll still need to output video in a compressed format at some point. Generally speaking, the more expensive the tool, the more options you’ll have, and the more you’ll need to know to use it effectively. Sometimes this means that expensive toolsets can be difficult to use, but they can be powerful. You’ll want to know as much as possible to use your encoder to its full potential and achieve high-quality results.

Encoding as a Compressionist

Maybe you don’t bother with editing and content creation. Maybe your part of the pipeline is to take the stuff already made and turn it into something else. Well, you, my friend, are a compressionist—mazel tov. You probably don’t need quite the same equipment your friend the editor needs, but the setups will still look shockingly similar (although you will have a wider variety of tools on hand). Because editors typically are their own encoding customers, they usually only need to accept content from one primary source—their NLE system. However, you’ll need to be equipped to decipher lots of video formats and content from many sources.

What are the basics you need to consider? These are the fundamental requirements:

• You need to be able to successfully open and play back content that comes your way.

• You need to get that content into a format from which it’s easy to encode.

• You need to encode it (duh).

• You need to organize and archive it.

• You need to deliver it.

In short, if you’re encoding as a compressionist, you’ll need to know more about encoding than your editor counterparts and be prepared for many different scenarios.

In the ever-changing world of digital video, there’s at least one constant: the need for compression to make video manageable and deliverable to the end user. Entire content platforms rely on the work of compressionists who understand the ins and outs of video-encoding workflows.

Automating the Encoding Process

As a compressionist, you’ll likely be working with much higher volumes of content to encode. This means you’ll want to be as efficient as you can be and let the computer do as much of the work as possible.

Many encoding applications have various automation tools built into them that allow for higher efficiencies. In actuality, you’re not so much removing yourself from the process as allowing the computer to handle the simpler mundane tasks and jobs, which leaves you free to focus on the more difficult segments or to manage a larger volume of work.

The most common automation feature is the watch folder. The compressionist designates a folder on the computer through the encoding application as an incoming compression source and assigns a template or groups of templates to the folder. When the watch folder is activated, the encoding application monitors the folder looking for new files. When files are copied to this folder, the application automatically begins encoding the content based on the templates that were assigned.

Some systems further support automation by allowing finished files to be copied across the local network or remotely to FTP servers for delivery. Chaining together actions in an automated way can be crucial in time-sensitive environments. One example of how a watch folder might be useful is the creation of dailies by a postproduction house for overnight delivery to a director on set.

If the content is always in the same format (frame size, frame rate codec, etc.), a watch folder system can be relied on to deliver consistent results. But if you receive a variety of types of content, this type of automation may be less useful.

Enterprise Systems: The Big Leagues

Automation on a single encoding system may help you speed up a production environment, but in many cases it may not be enough. Some broadcast facilities and outfits dedicated to the business of encoding have to deal with upwards of 200 or more hours of content a week. When you hit that scale of work, it’s time to invest in an enterprise encoding solution that enlists multiple computers in the encoding process. Telestream’s Vantage and AWS Elemental Server are two popular on-premise enterprise solutions, but many others are also available in the cloud.

An enterprise solution means more than just being able to split compression jobs over several computers (known as grid encoding). It also means giving the compressionists running this system some options that might not be present in lower-end systems such as segment re-encoding or support for adaptive bitrate packaging. Such tools aren’t cheap—expect to spend at least $40,000 to $60,000 just to get into the entry-level models. But if you are encoding that much content at a time, you probably have a business case to spend the money.

Essential Encoding Equipment

Unless you’re working in a studio that specializes in high-end theatrical delivery, you’re probably not working with a hardware encoder. More likely, you’re doing your encoding with software, perhaps using a supporting application that shipped with your NLE system (such as Apple Compressor or Adobe Media Encoder). You might even be using a stand-alone encoding application such as Sorenson Squeeze, FFMPEG, or Apple QuickTime. But that doesn’t mean you aren’t using hardware to support your encoding work or that your hardware setup isn’t critical to the process. Make no mistake, video encoding is nothing if not a processor-intensive application, so making sure you have enough power in your system to support the process is a key step in your preplanning efforts.

Hardware

If you’re about to build (or buy) your encoding system, it’s a good time to be doing so. Most computers on the market today ship with two or more processors; these multicore systems deliver a noticeable improvement over older single-core and even hyperthreaded systems, which do a decent job of acting like multiprocessor systems but aren’t the real deal. Video encoding demands lots of encoding horsepower and, more to the point, needs the full attention of your processor, or as close to it as possible. Whenever possible, your encoding system—even a multiprocessor one—should be a dedicated one since asking your computer to perform other tasks while encoding just slows down the process. Multicore systems are great for focusing the full resources of one or more encoders on the encoding tasks at hand while relegating routine operations to other system resources that don’t interfere. Keep in mind that not all encoding software is optimized for multicore systems, but you can expect to see that change as new iterations of the tools emerge.

Another advantage of multicore systems is that the processors themselves don’t need to be as honkin’ fast; in other words, clock speed is less important than the number of processors you have. If you have multiple older, single-processor systems you want to use for encoding applications, you can network them into a cluster, but keep in mind that this process—which actually breaks up an encoding job into multiple segments, sends the pieces to the other systems on the network for encoding, and then retrieves the encoded segments and pieces them back together—might not increase your efficiency as much as you’d imagine. In many distributed encoding systems, the longer the clip to be encoded, the more time that is spent shipping it back and forth across the network, which cuts into your encoding time savings. In addition, breaking up the clip will reduce your encoder’s ability to identify redundancies in the data that make efficient compression possible.

Monitors aren’t as important as processor power, although you’ll want to make sure that you have enough resolution to view your encoded output at full resolution. If you’re encoding 4k material, you’ll want to be able to view the output without scaling it down. If you’re using multiple machines in a cluster, you can also use a keyboard-video-mouse (KVM) switch and let multiple machines use one on-screen keyboard if you can.

Storage

As for storage, there’s no way around it: You need a lot, and it needs to be fast (Figure 3.4). Fortunately, high-performance storage has gotten cheaper and faster over the years. In addition to capacity, you must consider the performance of the drive’s connection because you will be moving around very large files. The slower the drive, the longer your encodes will take. Table 3.1 shows the different transfer speeds for various common hard drive interfaces.

Figure 3.4 Storage may be the most important feature to the compressionist. All the encoding horsepower in the world doesn’t mean a thing if you do not have enough space to properly store the source material as well as all the files you plan to create in the encoding process.

High-Performance Storage

Depending on how high-quality the source is, you may need a RAID if your work requires playing back the video in real time. At the least, you’ll want a NAS solution for offloading work. A SAN can be annoying (well-specialized knowledge is needed to set up and manage one) but handy. All these systems share the same basic concept: using multiple hard drives combined to act as one large storage device (Figure 3.5).

Figure 3.5 If compression is your main job, consider investing in a high-capacity storage system such as a SAN.

The higher the volume of work you are performing at any given time, the larger-capacity storage your encoding system needs to be, and potentially the more encoding machines you need. The number of jobs and length of content will dictate how much active space you need, but 1 TB per CPU is a good start.

Note how we said active storage space. After you have been encoding video for a while, you’ll discover that doing so means creating a lot of files. Some files are tests, and others are intermediate or mezzanine files. It is quite easy to run out of space, lose files, or generally become confused about where everything is. As part of staying organized with your files, you should also create and follow a storage policy. A storage policy is a guideline for how files get treated as part of your workflow. Depending on how long your project cycles are, try storing files based on what is active and keeping those files close at hand. Then, in descending order, create a storage hierarchy for files older than 30, 90, and 120 days. Active files are regularly needed for work and are on the main system, as are files that are less than 30 days old. You can back up older files to external hard drives such as a NAS solution once you no longer need to access them regularly. Every six months or so, it is worth reviewing this cold storage to decide whether the files still need to be retained on a local drive or can be uploaded to a cloud backup service, LTO tape, or other form of even colder storage.

In addition to backing up the video and audio files you are working with, don’t forget to back up your settings and applications. If you can afford to do so, keep an external hard drive available as part of your backup system that is an exact copy or clone of your entire encoding system, and plan on recloning it once every 30 or 60 days. This way, if anything happens, you have a perfect image of your entire system that can be used to quickly get you back online. Apple’s operating system comes with a built-in backup system known as Time Machine, but many other backup systems are available for both the Mac and Windows machines. The best backup system is the one that gets used!

Input and Output

We’ve talked a little bit about storage connections (refer to Table 3.1), but it’s worth taking a moment to dive a little deeper. If you think about each step of the encode process as parts of a pipeline, the goal is to remove as many bottlenecks in the pipeline as possible. This will allow the system to reach its maximum encode speed.

Generally speaking, the steps of the encode process go something like this:

1. Read source data from disk.

2. Decode source data and store in RAM.

3. Encode decoded source data to destination format.

4. Write encoded data back to disk.

When you break it down like this, you can see that there are at least three factors that can affect overall encode speed: CPU clock speed, amount and speed of memory, and disk throughput. If you’ve purchased or built a new machine in the last couple of years, then it’s likely that your CPU and memory are up to speed. However, your disk speed (or lack thereof) might be slowing you down.

In many situations, your CPU is able to process data faster than you are able to feed it. This means the CPU might be wasting clock cycles waiting for the hard drive to give it more data to work on. Even the fastest and most expensive systems are only as strong as the weakest link, which could be your storage system. Note that we say storage system. This is because there can be many parts of a storage system depending on your exact hardware. Solid-state disks are typically faster than spinning hard drives, and drives that use a faster connection usually transfer data faster than those that use slower connections, but these aren’t hard-and-fast rules. A striped (RAID 0) set of ten spinning hard drives connected over Thunderbolt 2 could be faster than a single SSD connected over Thunderbolt 3.

When it comes to disk throughput, you can’t just look at the max speed of the connection used because the speed of the drive inside the enclosure could be the limiting factor. Real-world speeds of devices also tend to vary from the stated theoretical maximum speed. One rule that does tend to hold true is that the faster the performance of the drive, the more expensive it tends to be, especially at larger capacities. An economical storage strategy is to have a lower amount of high-performance active storage used for encoding, and a lower-cost, higher-capacity storage system for bulk storage of encoded files.

Although you may not specifically have to purchase storage separately, be conscious of your system’s hardware input/output (I/O) ports because each protocol typically has different supported read/write speeds, which will greatly affect your performance for copying files back and forth and compressing files on external drives.

Software

Hardware, of course, is only one piece of the compressionist’s toolkit. You’re also going to need a variety of software to help you get the job done.

Operating System Agnosticism

Although choosing an operating system is a religious issue for many people in this industry, no operating system is inherently superior for video encoding. The most important thing is to choose the one that best matches your existing workflow. There are lots of transcoding software options on the Mac side, and the software is typically a little cheaper; however, the hardware will cost more. Like many things, it’s a trade-off.

Players

The first part of being able to encode video is being able to play it back. With the range of file formats and codecs in use (see Chapter 2 for more details), you may often find yourself presented with a file that doesn’t conform to your production standards. Keeping a variety of video players handy is a useful way to verify that the video file you need to encode works and is in a format that will work in your particular workflow. Another thing to keep in mind is that many players interpret color differently (even on the same monitor!), so the ability to view material via multiple players can be crucial to the quality control (QC) process.

Disc Rippers

The term ripper sounds like the evil tools of pirates and misguided college kids, but the fact is they are a perfectly legal and necessary tool that should be part of any compressionist’s toolkit. A ripper is a specialized encoder that allows you to convert DVD or Blu-ray discs into a format readily understood and usable by your encoding workflow. Though not the best source material for encoding video, authored optical discs are a popular way of moving files around because they are small and lightweight (easy to ship) and nearly universally playable. The problem with using a DVD or Blu-ray Disc as source material for re-editing and re-encoding is not playability but access.

Although a DVD or Blu-ray Disc has the ability to act as just another data storage medium, like hard drives or flash media, these discs are most often written in a playable format. This means that the video is formatted specifically for set-top and software players. For DVDs, the video data is stored as MPEG-2 files in video objects (VOBs), which include metadata that makes the discs readable in consumer players. These VOBs need to be disassembled—a process often described as unpacking or ripping—to extract MPEG-2 video files that can be edited and re-encoded on your computer. Blu-ray Discs store the video data in .m2ts files, which stands for MPEG-2 Transport Stream. Transport streams are optimized for playback and also need to be converted for any other use.

A popular (and useful) DVD ripper on the Mac is DVDxDV (www.dvdxdv.com), which allows you to extract whole sections or just small clips by marking in and out points from a video DVD. Another option is the open source and cross-platform application Handbrake (https://handbrake.fr). With both of these tools, you can export the files as a number of high-resolution file types, making it easy to then transcode them to whatever formats are needed. Several PC ripping tools are also available; however, many automatically transcode the video to another editing or delivery format on the fly, which means they’re optimized for speed rather than quality. Although this seems appealing and can save time, we haven’t been impressed with the results of such direct-from-disc encoding and, as such, prefer to keep the extraction to a high-quality format and do the transcoding later. Something to keep in mind is that DVD is inherently a standard definition format, so the largest frame size you’ll be able to extract is 720 × 480, while Blu-ray supports the two popular HD frame sizes—1280 × 720 and 1920 × 1080—as well as 4k UHD (Table 3.2).

Compression Tools

Chapter 5 discusses compression applications in specific detail, but we’ll lay the groundwork for that chapter now by discussing the categories of compression tools in use today. Some applications—usually the free or lower-cost tools—are meant to encode only one video clip at a time. Some applications that are meant for other purposes, such as editing, creating motion-graphics, or authoring DVDs, can also encode video clips. These single-encode tools are useful for one-off projects and jobs that use simple settings but are typically used as part of some other workflow, not just video compression.

If you are reading this book, you probably have more than just one clip at a time that you need to deal with, so you’ll likely need a batch-encoding tool to get your compression jobs done. Using a batch encoder will allow you to save specific settings as a template that is stored separately from the videos, then apply them to lots of files at the same time, and finally send them off to encode. The application still encodes each video one format at a time but will keep the rest queued and begin encoding them one after the other until all files have been completed. Batch encoding is incredibly handy because hours of content can be queued at night or on dedicated machines while you accomplish more important or creative tasks.

Although you will most likely pick one compression tool as your primary workhorse, it’s a good idea to have a few of these tools installed so that you can switch between them as needed. Certain applications perform better than others for certain types of files and certain types of source material. For example, Andy used to regularly receive uncompressed music-video files from record labels in New York for encoding to various sizes meant for delivery on the Web. One band regularly sent videos that were shot and edited in such a way that the standard deinterlacing options available in his normal compression application produced poor results. He knew that when that band’s content came in, he’d want to use a different application that had a wider range of deinterlacing options available.

Another reason to be familiar with several different compression tools is that, over time, support and features may change enough to merit switching from one to another for your day-to-day work. The flexibility to switch between different applications to get the best performance and results will help set you apart as a top-notch compressionist.

Aspect Ratio Calculators

In the course of normal compression work, at some point you’ll be asked to change the frame size. In scaling and cropping all that video, you end up doing a lot of math to determine the new dimensions. Several handy tools have cropped up to make this easier on those of us trying to crunch the numbers quickly. Although it might not be the kind of tool you use daily in your workflow, this type of product can come in handy from time to time. Several aspect ratio calculation tools can be used online, and others are available as small downloadable applets. Some of the better ones we’ve tried include the following:

• The Aspect Ratio Calculator (cross-platform),

  www.wideopendoors.net/design/aspect_ratio_calculator.html

• The Screen Aspect Ratio & Dimension Calculator (online tool),

  www.silisoftware.com/tools/screen.php

• Andrew Hedges Aspect Ratio Converter,

  https://andrew.hedges.name/experiments/aspect_ratio/

Video Storage and Data Rate Calculators

Because storage is one of the most important parts of the compression process, tools to help calculate storage requirements are also super handy.

VideoSpace is a useful macOS dashboard widget that can help calculate the amount of disk space required given a duration, frame size, frame rate, number of audio channels, and codec.

If you’re more inclined to use a mobile app, AJA’s DataCalc for iOS or Android can be found in Apple’s App Store or on Google Play, respectively.

Both apps also work in both directions, so you can calculate time to space, or space to time. Having a data rate calculator in your toolbox can prevent you from creating files that are too large for your destination and are a must have for any compressionist.

Analysis and Quality Control

It’s not enough just to encode the video. You need to make sure it actually works as it should and that it doesn’t contain unacceptable artifacts. This is the fundamental premise of quality control. QC isn’t just about making sure the video file opens and plays back either. True QC means replicating the environment in which the video will be consumed and testing its performance there. As a compressionist, this means having access to a variety of DVD or Blu-ray players, computers, mobile devices—or whatever gives you the broadest representation of your target platforms—to assure that the content you are creating will play back on them. Ideally, the files should be tested in their entirety; it’s tempting to test just the first few seconds of any clip, but until you’ve reviewed the entire video to verify full playback without losing audio/video sync, dropping frames, and so on, your compressed video has not passed QC. After all, nothing can be more embarrassing than delivering files that do not work.

As you can imagine, this process can be quite time-consuming. It may be impossible to check every single frame of every video you encode, so you have to decide what level of spot-checking you are comfortable with in order to catch any serious problems. Some encoding tools, such as ProMedia’s Carbon Coder, have some analysis tools built into them, allowing the QC process to be done as part of the encoding process. Stand-alone products such as Tektronix’s Aurora or Telestream’s VidChecker analyze video after the encode process to verify specific settings. Several companies have also developed software and hardware designed to perform the same QC in a matter of seconds that would take a human all day to do. However, these industrial-strength products can be quite expensive, so again it’s a matter of how important QC is to your work when deciding how much time and money to invest in performing it.

But QC in the lab or studio environment isn’t enough on its own. It’s important that the encoded video be tested under the real-world conditions that the end user will encounter. For example, video meant for progressive download via the Internet should not only be checked for visual and audio quality but also for the appropriate file size that allows the user to watch the video without buffering during the download. Regardless of how great the video looks, if the file is too big to download and watch simultaneously, then the video ultimately fails to deliver the desired experience. It may be better to reduce the quality slightly to get the file down to a manageable size and thus make the viewing experience more enjoyable.

Productivity Tips

Now that you have your compression setup and workflow in place, let’s review some best practices to follow while encoding.

Minimizing Quality Loss Without Overdoing It

Typically, though not always, the goal of video compression is to get the best-possible quality out of the smallest amount of data (the most bang for your bits, as it were). So, what exactly does that mean in terms of encoding?

To begin with, you want the best-quality source material you can get your hands on. We can use a painting metaphor to describe the bits that make up a video clip. The more paint (or bits) making up the source, the better the end result will be, regardless of the file sizes involved.

Until the moment you are creating your delivery format (the clip the end user will actually view), you want to stay in what is referred to as production or authoring formats and codecs that will retain as much data as possible. Production formats typically take a fairly powerful machine to open and play back, but that’s OK—you aren’t sharing this version outside your postproduction workflow, so leave it big and buy more storage if you need it. You want all the data possible available right up until the moment you create the final version. At that point, you are going to subject this super-high-quality, super-high-data-rate file to all the magic, trickery, and science you can to maintain superior video and audio quality while removing as much of the data (bits, paint, whatever) from the process as possible.

That said, don’t be silly and take low-resolution source files and convert them to some high-quality format in the hopes that you will get a superior end result. You can’t invent bits that weren’t there to begin with. You will waste your time and end up with just-as-ugly video as the end result—just in a lot larger file. The quality of the source you get to work with, whether it is uncompressed or something else, is the highest priority you should concern yourself with for the duration of the job.

Stay Organized

As mentioned earlier in the “Storage” section of this chapter, if you are doing any kind of serious encoding, regardless of the volume of work, you are going to make a lot of new files. If you’re not organized, it becomes easy to lose track of files or, worse, erase something you didn’t mean to erase. How do you need to stay organized when it comes to encoding?

• Don’t work on your desktop. It is tempting to keep all the files you are working with on your desktop. Do not do this! Your desktop should be a temporary space, and it can get cluttered quickly. Instead, set up a working folder, preferably on a dedicated media storage drive. If you are working on a laptop or do not have a dedicated media drive, create this folder in your local user account.

• Use source and export folders. Within your work folder, set up both an incoming work area and an outgoing work area (just like the little baskets people used to have on their desks). This way, all your source material has an assigned spot in which to live for the time you need it, which can be regularly erased to make more space. Your newly created files have a special location to go until you have time to back them up, upload them, or do whatever needs to happen after encoding. This is an area that will also be regularly deleted to make room, but since you know all the files that live here were newly created, make sure they have been archived somehow before deleting.

• Create client or project folders. If you have a lot of different unrelated projects, maybe across multiple clients, it’s also a good idea to give each project or client its own unique folder within the incoming and outgoing sections. Nothing is more embarrassing than delivering the wrong content to the wrong person.

• Develop a storage policy. As described in the “Storage” section, make keeping your storage organized part of your daily or weekly schedule. Regularly move files you no longer need on your active system to some sort of backup and then review the backup from time to time to make sure it is also organized.

Filenaming Conventions

Choose a method for naming your files, and stick to it as best you can. It’s easy to get lazy about naming files and give them generic names like Finished Movie.mov, but will you remember what that was in a month when you need to use the file again? Probably not—especially if you’ve used similar names for other video files. Even just a few years ago, we were limited to only a few characters for filenames. Both OS X and Windows XP made it possible to have much longer filenames, but often only the first dozen characters or so will be visible in the operating system, making it hard to differentiate between similarly named files.

Filenames should be long enough to help you describe what the file is without opening it but short enough to work in the computer environment. Spaces in filenames were also once taboo but have become acceptable to use on local machines. However, files on remote servers such as a web or FTP server typically cannot read across a space in a filename, so substituting a dash or underscore is a good policy.

Also, remember that at minimum when encoding, you are going to have two similar files: the source and the export. And these days you will typically have more than just one export; you may have half a dozen exports from one source clip, formatted for a variety of destinations with various data rates, formats, and resolutions.

File extensions will help you easily differentiate many of these different formats, but in some cases, the file type isn’t enough. For example, all the movies exported for the Web may be in the same format but at different data rates meant to target different qualities of the same source. In this scenario, it’s common to denote the different qualities with a numeric value of the target data rate. So, if you have three encodes of the same movie—one meant for low-bandwidth or mobile connections, one for middle-of-the-road broadband, and one super-high-quality movie meant for high bandwidth—you can easily differentiate between them. By adding the target data rate to the filename (which, in this example, would mean adding something like _56, _200, and _800 to the end of the filenames), you can tell which file is meant for which delivery without opening each one. It’s also a good idea to keep the numeric value in the same base, so convert the earlier data rates in megabits per second to their equivalents in kilobits per second (for example, 1.5 Mbps would become 1500 Kbps for naming purposes).

Last but not least, try not to use superlatives like the word final in your filenames. Many professionals disdain the use of this word—for good reason! Almost every project will come back for a revision of some sort after you create your deliverables. Files named final will almost always be out of date at some point in the future, and thus the word final loses its meaning. A better practice is to stamp files with a version number or a date (Table 3.3).

Experimenting Can Be a Good Thing

Even the most proficient compressionist has to do some trial-and-error encoding from time to time to optimize the settings for a specific video. The less experience you have with encoding, the more trial-and-error work you may need to perform to get the exact results you want from a clip. Save yourself a lot of time and effort by performing these dry runs on short clips rather than on your entire video. Many encoding tools allow you to mark in and out points around a specific part of a video, but if yours doesn’t, find a way (through an editing tool) to create a short, self-contained clip you can use to tweak and perfect your settings. Take care to alter only the length of the source clip so that the short clip you’ve created represents the full-length source in every other way.

If the image of your video content changes dramatically from beginning to end (for example, cutting between fast-moving outdoor scenery and indoor headshot interviews), then you may want to perform experiments on one or two sections of the video to make sure you aren’t optimizing it for one specific type of scene at the detriment of the others. Using these short clips allows you to quickly adjust and re-encode your content several times quickly until you get the results you want. Then set the entire video (or multiple videos) to encode using the final setting.

Keep a few clips around as test encodes for new applications as they come out (or even upgrades of the same application). These clips should run the gamut of the same types of video you are likely to receive for encoding (for some, this may mean from super-high-quality, gorgeous 4k content all the way down to video shot on a mobile device). When new products or upgrades release, it’s useful to run this baseline test video through once at some fairly standardized settings so you can see what type of results you may expect from the application during your regular work. This basic diagnostic experiment can quickly help identify pitfalls in your workflow before they cost you time and money in the middle of a big project.

Mezzanine, or In-Between, Files

It isn’t practical for most people to try to keep uncompressed versions of every video they create; it is possible to “lightly” compress the same video to some format in between the finished deliverable and the source. These files are known as mezzanine versions. Mezzanine is an architectural term for an intermediate floor found between two main floors of a building. In the encoding world, it describes a file that is high-enough quality to act as an archive and encoding master but sufficiently smaller than the original source to be stored and handled practically.

Some people refer to this as an intermediate format, but this can get confusing because the term digital intermediate (DI) is used in the postproduction world to describe a digital color-correction session. To avoid confusion, compressionists have adopted the term mezzanine, and though it hasn’t been accepted as a standard, it’s used widely enough that others should generally know what you’re talking about if you use it.

I-frame-only MPEG-2 video is a popular mezzanine format for many encoding facilities because it is still edit friendly but relatively small in file size when compared to 10-bit uncompressed video. Apple’s ProRes 422 and GoPro’s Cineform are also high-quality codecs that are popular both in postproduction and in compression as mezzanine formats.

Archiving and Transcoding

Video clips that get used as part of marketing such as movie trailers or commercials may be first released in one format or size but, over the life of the marketing initiative, may need to be sent out in as many as 20 or more different resolutions and formats. Rather than repeatedly ingesting the same source again and again, you can save a step by storing a mezzanine version of the content somewhere safe and then transcoding from it repeatedly whenever needed. This can potentially save a great deal of time as part of your workflow.

Conclusion

By now you should have some idea of what type of compressionist you are and the gear you need to get the job done. But there is still one area of knowledge compressionists need if they are to be considered world-class professionals. This area is the difference between just being able to convert video simply and efficiently from one format to another and being able to take any video and draw out the best quality and the most detail from each frame when compressing or transcoding it to a deliverable format. Skill in this area is invariably referred to as the “magic of video compression” because the results are sometimes nothing short of a miracle. While some might call it a “dark art,” we in video compression just call it preprocessing, and although it can be a little specialized (and sometimes a hit-or-miss process), it can be taught to mere mortals. How do you think we learned it?