Video transcoding refers to converting a previously compressed video signal into another one with different format,such as different bit rate, frame rate, frame size, or even compression standard. Until today, video transcoding is one of the most compute intensive operations running on our PC.
The conventional way of transcoding is done through the brain of our PCs, that is the central processing unit(CPU) i.e Pentium 4, Athlon, Core 2, Phenom and others. Our CPUs can handle the transcoding job without any problem. The main problem of transcoding on the CPU is that it takes ages to complete and while the process is running, it eats up all the processing prowess of your cpu rendering your PC useless until the process is completed. Nowadays, most transcoding software is tuned to take advantage of multicore and multithreaded CPU i.e Pentium D, Athlon X2, Core 2/Q/i7, Phenom and others, which somehow manage to reduce the transcoding time than their single cpu counterpart which is expected.
The new way of transcoding is by elevating the processing requirement to your PC's Graphic Processing Unit (GPU). In simple terms, GPU is a subsystem in your pc which renders images and then output it to your screen. However, today's GPU can do much more than just drawing images and 3Ds, it can handle brute force arithmetic operations that have never seen before on a CPU and software developer can tap this ultimate power to increase the overall performance.
Why GPU is faster than CPU?
CPU is more to a general processor for everday use such as email, web browsing, chatting, word processing, and others. In other words, the CPU can handle any job given to them but they do not specialize in doing their job. GPU on the other hand is a specialize processor which is designed to carry out their specialize job. So, GPU is fast in doing their work but their scope of work is not as wide as CPU.
GPU is fast because it contains multiple cores(in this case it is called stream processor) which handles multiple work at the same time. Although the each GPU's core is not as complex as a CPU core, but the more the merrier. On a budget GPU, it contains 32 stream processors vs the fastest CPU with 4 processors.
What type of GPUs is supported?
For Nvidia GPUs - Geforce 8 series, 9 series and GT1xx/GT2xx
For ATi GPUs - Radeon HD2000 series, HD3000 series and HD4000 series
Is the performance difference is noticeable?
Lets find out!!
Central Processing Unit
AMD Athlon X2 5600 @ 3.2Ghz
256KB L1 Cache, 2MB L2 Cache @ 3.2Ghz
2GHz Hypertransport system bus, 2GB Dual DDR667Mhz Memory
Graphic Processing Unit
Nvidia Geforce 9600gt 64 stream processors
Core @ 700Mhz Shader Domain @ 1.75Ghz
512MB GDDR3 @ 1.9Ghz
GPU Transcoding: Badaboom 1.1
CPU Transcoding: Xilisoft Video Ultimate COnverter 5
Benchmark 1: Transcoding a 01h 28m 46sec .avi video to Ipod Nano Highest Quality Preset
Benchmark 2: Transcoding a 01h 38m 14sec .avi video to 1920x1080HD Highest Quality
Bencmark 3: Transcoding a 24m 32sec video to 720x576 Highest Quality