CPU Utilization

With any computer application, there will be a portion of the CPU allocated for it depending on the type of task it is performing. The amount of CPU power used while an application is running is known as the CPU overhead, and we were interested in what kind of overhead we would see while using these tuners, since recording a program is something that should ideally be able to be done in the background while performing other tasks on your computer.

We used the windows task manager to get a general measurement of CPU usage during three different states with each of these cards. The first was while simply watching live TV on the system, then while watching TV and recording simultaneously, and lastly, while recording live TV in the background (with WMCE closed) and no other programs running. We also measured the amount of CPU usage while the system was idle just to verify that it was the expected 0%, and it was. Here are our CPU utilization results.

Theater 650 Watching Live TV:


Theater 650 Recording Only (not watching):


Theater 650 Watching and Recording:


Theater 550 Watching Live TV:


Theater 550 Recording Only:


Theater 550 Watching and Recording:


DualTV MCE Watching Live TV:


DualTV MCE Recording Only:


DualTV MCE Watching and Recording:


CPU Utilization Summary
Watching
Live TV
Recording Watching +
Recording
NVIDIA DualTV MCE 25% 7% 35%
ATI MSI Theater 550 Pro 21% 4% 29%
ATI MSI Theater 650 Pro 21% 4% 25%


We can see that the Theater 550 and 650 show similar results, which makes sense as the 650 is based off of the 550. We see significantly higher CPU usage with the DualTV, particularly while watching and recording live TV, but this is because we are able to watch a different channel than the one we are recording. Had we installed a second Theater 550 or Theater 650 card and used it to watch another channel, we have no doubt the CPU usage would have gone up as well. It is possible one of ATI's partners could include two tuners on one board to provide a solution similar to NVIDIA's DualTV, but for those not interested in multitasking, the Theater 650 does offer some benefits over the competition (as you will see in the next section).

TV Tuner Comparisons Image and Video Quality
Comments Locked

78 Comments

View All Comments

  • rqle - Wednesday, June 14, 2006 - link

    Currently own a Theater 550 chipset and i like it much better then my roommate nvidia dualTV one. My question is, can i buy another Theater550 so i can dualTV, would the system go crazy on me? Would like to have dual screen running different sport program at once and would be cheaper for me as well.
  • DerekWilson - Wednesday, June 14, 2006 - link

    that would work fine
  • Woodchuck2000 - Wednesday, June 14, 2006 - link

    When is Anandtech going to publish a decent article on TV tuners? In no particular order:

    Why use ATI promotional material that is designed to prove that the 650 is better than the 550? Use your own material if you want a fair test. More to the point, why transcode it? Im sorry, but file size considerations are not a good justification for degrading the image quality <IN AN IMAGE QUALITY TEST>.

    How do you justify adding throwaway comments about equivalence with Hauppage products in the conclusions if you havn't addressed the cards at any point in the review?

    Why have you taken the instantaneous CPU utilisation from task manager and tabulated it? You need to take an average over a reasonable period of time if it is to have any relevance.

    Seriously, there's not enough depth in any area to make the review really worthwhile - you may as well post a single page saying "Look everyone, ATI have sent us this nice new card they want us to plug!" and it would have the same amount of relevance to 99% of your audience.
  • Cardio - Wednesday, June 14, 2006 - link

    Why is there no mention of how this card preforms on DTV, that is its major difference and only really new feature? Am I missing something here? You didn't test the Hauppauge because you have looked at it before. You've looked at the ATI 550 before also. In the earlier test you campare a dual tuner card to a single tuner card rather to the other available dual tuner card. Will ATI only give you stuff to test if you don't compare it to the competition? As you said, Nvidia's card is only available directly from them and at twice the price. The Hauppauge MCE 500 is available anywhere and is cheap. Oh, you have looked at those cards before.

    I can tell you for sure that I am a lot more interested in Picture Quality and functionality than in power consumption or cpu usage.
    Not worthy of AT and its getting to be a habit.
  • DerekWilson - Wednesday, June 14, 2006 - link

    We transcoded it for your benefit (smaller file to download) -- in our image quality analysis, we used the unmodified output of the cards. Sorry if this was unclear.

    As for using ATI's material, you can't argue with a zone plate. Better is better and worse is worse. Same for the other tests we ran which were designed by SMPTE (http://www.smpte.org/)">http://www.smpte.org/) ... ATI was gracious enough to provide us with these tests we would have otherwise had to purchase ourselves.

    testing the Hauppage card is a reasonable request, but it doesn't change the fact that it was similar in quality and performance to the Theater 550 in previous evaluations. That is perfectly reasonable information to use in drawing conclusions.
  • SHSPVR - Wednesday, June 14, 2006 - link

    One thing I really hate more then anything is the fact that min reviewer are so lazy in the cheating game by using the PrtScn Button and do not do ture original video RAW decode MPEG screenshot in the rigth application software like Womble MPEG-VCR and always seem to be using MCE DVR-MS proprietary video and audio format carp what matter are they just to lazy in using other PVR software or GraphEdit to get real MPEG files but as we all know that the NVIDIA DualTV is not option with GraphEdit which has no exposes advanced features properties.

    2nd problem as today no body turely using low bitrate under 4MBit/sec which is more reasonable testing to see how well it really dose encoding without artifacts.

    So you wondering why I said cheating well each encoder dose there video difference from each other and there for the decoder filter display difference in case depend on the clean up filter like Deinterlace, difference Enhanced Video Processing Technologies which get apply on top the MPEG Video.

    Now as for comparison.
    Hauppauge, ATI and NVIDIA
    Hands down Hauppauge PVR cards wins becuase of some major reason it has support for more OS then any one eles Windows, Linux, Mac (Apple) and yes even OS/2 Warp or eComStation and has more 3rdparty application support.

    As for a real encoding comparison all test they should done with the fellowing device
    VCR, Stardard Antenna TV, CableTV Analogue/Digital and Satellite with fellowing output source S-Video/Composite and Tuner

    CPU Utilization is really useless becuase all of today Video card are still mostly base on 3rdparty SW-Decoder which mean you still need fast CPU unlike a ture HW-Decoder which provides a real decoding performance unlike today video card which only do part HW-accelerated decoding not full HW-decoding.
    Let say for Example you have Intel P3 800MHz runing 2000Pro or XP with 512MB and you can chosse from GeForce"4" MX4000 or GeForce"7" 7800 becuase you want upgarde your old doggie ATI 3D RAGE PRO 4MB what do you think you see in CPU Utilization from the two card newer card.
  • andrep74 - Tuesday, June 20, 2006 - link

    One thing I really hate more [than] anything is the fact that I have to try to read cr*p like your comments, which try to pass off the sh*t in your head for English.
  • SHSPVR - Wednesday, June 14, 2006 - link

    Oh I also for add You need do Cartoon as well as reg video show

Log in

Don't have an account? Sign up now