My Lab Toolbox: Frame Grabbers
This article is the third in a series of articles describing hardware and software tools I’m using in my test lab. Previous articles covered the RDS tool SenseConnector and how to build a white label test server. This article highlights frame grabbers used to capture sequences of computer screen images (frames) which then can be stored as video recordings. Traditional use cases of modern frame grabbers are capturing videos from healthcare or astronomy applications and providing video input for pick and place machines in manufacturing. New use cases are capturing videos from interactive computer games and recording end-user experience of interactive desktop sessions to measure perceived performance. The latter is exactly what I’m doing in my lab and here are the lessons I have learned.
A frame grabber is a physical device designed to consume and record signals generated by the VGA, DVI or HDMI output. The purpose of such a device in a virtual/remote desktop test lab is to record video sequences while user are interacting with the system or while automated application test runs are performed. The advantage of a frame grabber in such a scenario is that it does not generate any extra load on the client device – it’s non-intrusive. The alternative to a frame grabber is using video recording software, such as Camtasia Studio on the client. But this may either be impossible (like on a smart phone) or influence the visual results in such a way that they are not representing the real endpoint performance. So in essence, only videos recorded with frame grabbers can be used for unbiased visual comparisons.
Modern frame grabbers have the ability to capture screen images at high resolutions and frame rates, such as full HD (1920 x 1080 pixels) at 60 Hz. Some frame grabbers are able to capture audio at the same time, depending on the hardware capabilities. Active frame grabber perform compression on the video frames while passive frame grabbers just capture raw video data and forward the video stream for further processing.
I’m using several frame grabbers manufactured by Epiphan in my lab. The same is true for my fellow benchmarking experts Shawn Bass and Ruben Spruijt. But Epiphan is not the only company providing great frame grabbers, so here’s a short overview of what’s available in the market:
- Epiphan DVI2PCIe or DVI2USB: Passive high-end VGA/DVI frame grabbers that require video recording software with appropriate codecs. (http://www.epiphan.com)
- Elgato Game Capture HD: Active frame grabber with HDMI input, hardware H.264 encoder and USB2 connector. (http://www.elgato.com)
- Hauppauge HD PVR 2: Active frame grabber with HDMI and video input, hardware H.264 encoder and USB2 connector. (http://www.hauppauge.com)
- Blackmagic Design Intensity Shuttle: Passive frame grabber with HDMI and video input and USB3 connector. (https://www.blackmagicdesign.com)
- Vision Dimension: USB 3.0 and PCIe video capture – HDMI, DVI, VGA and video (up to 4k resolution). (http://www.vd-shop.de)
- AVerMedia DarkCrystal HD: Active frame grabber with HDMI and video input and hardware H.264 encoder. (http://avertv.avermedia.com)
In July 2014, Rachel Berry form Citrix published a great article about using video capture hardware for HDX product demonstration. Rachel’s article features two of the frame grabbers mentioned in my list. She also notes that some of the frame grabbers are consumer grade hardware. But today, due to the popular habit of recording computer game sequences and uploading them to YouTube, consumer video capture hardware is very mature. The typical price range from $150 to $1,000 also blurs the line between professional and consumer products.
It is important to note that using a frame grabber typically requires a dedicated recorder PC or laptop. Depending on the compression algorithms, video file sizes can be in a wide range. The codecs also define how good the resulting video quality is. Ideally, everything would be recorded in a raw format without any compression and for further processing at a later stage. But this generates huge files, such as 372 megabytes per second (!!!) when recording full HD at 60 frames per second. So it’s better to compress the videos, either with an active frame grabber or with a software encoder on the recorder PC. Popular codecs are H.264 and MPEG4 v2. The most critical factors are identifying the adequate codec package (such as ffdshow) and finding the right balance between file size and video quality. It’s very important to avoid visual degradations and artifacts introduced by the codec if a serious comparison of different video sequences is the primary goal. And make sure you don’t change the codec settings for all recordings you want to compare.