MSRP: £1,799.00
Introduction
Philips’ 9000 series sits at the top of a fairly large selection of LCD TV ranges. Although that top-end range does include LED backlit displays, today’s 47PFL9664 is a 47” LCD HDTV which uses standard CCFL backlighting. Philips’ AmbiLight Spectra 2 is on board and claims to provide more immersive viewing, and there are a host of image processing technologies afforded by Philips’ own “Perfect Pixel HD Engine” chip. Let’s see if the entire package can come together to provide a top quality viewing experience.
Design and Connections
The 47PFL9664 is easily the best designed HDTV I’ve ever had here to review. It’s thin, has subtly rounded edges, a slim bezel, and best of all, has a fantastic, professional brushed steel look. This is a major leap forward when compared to the slab-like glossy black appearance of most similar products.
The back panel features 4 HDMI inputs, 2 SCART terminals for standard-def sources, Component video inputs, a VGA (D-Sub) PC input, and an antenna input for connecting your aerial to. A USB input, a 5th HDMI Input, and Composite and S-Video inputs can be found on the side of the display.
The back panel features 4 HDMI inputs, 2 SCART terminals for standard-def sources, Component video inputs, a VGA (D-Sub) PC input, and an antenna input for connecting your aerial to. A USB input, a 5th HDMI Input, and Composite and S-Video inputs can be found on the side of the display.
Features
The 47PFL9664 features Ambilight Spectra 2. For those not familiar with Ambilight, it’s a Philips-exclusive feature whereby strips of coloured LEDs mounted on either side of the TV illuminate the wall or other surrounding environment. The idea is to psychologically expand the size of the screen for a more immersive feel. By default, the coloured LEDs output an average of the on-screen colours, but they can also be configured to output white light, which makes Ambilight a useful bias lighting system, useful for avoiding eyestrain (and distracting from LCD constrast limitations!) in darkened environments. Naturally, you can turn Ambilight off entirely if you find the effect distracting. The TV also features Net TV, which is one of the more useful “Internet on your TV” implementations, for two reasons: first of all, the TV is Wi-fi enabled so can connect to your wireless home network without any LAN cables, and secondly, it’s not a walled-off system (the user can access any web page they wish).
Test Results
After leaving the Settings Assistant, I changed the “Smart Settings” to the Cinema preset, as this is usually the viewing mode that gives the most correct picture on consumer TVs. Immediately, I noticed that the image was somewhat dark and looked a little too yellow-ish, which made it off-putting. The reason for the darkness was the TV’s Light Sensor, which was enabled and was overcompensating for my slightly dim viewing environment. The reason for the yellow tint became apparent later.
I made sure that the basic controls such as Brightness and Sharpness were correctly set, and then noticed that there was subtle high frequency disturbance in the most detailed portions of one of the test cards I had input to the TV. I also noticed that there was a a small amount of colour bleed, where the colours in the image were noticeably offset from the black and white portion of the picture. Fortunately, the TV has a “PC Mode”, which is designed for use with a computer (hence the name). This skips certain video processing steps in order to reduce input lag, and also completely cured these two issues - what a surprise(!)
Up to this point, I had set the Philips 47PFL9664 up as best I could with only my eyes (that is, without the precision of a colour measuring device and associated software). Although I could have experimented with adjusting colour temperature by eye to neutralise the green tint, this would by no means be exact, and would likely have just introduced another colour bias into the image. I ran a set of measurements for a scientific assessment of the display’s output:
I made sure that the basic controls such as Brightness and Sharpness were correctly set, and then noticed that there was subtle high frequency disturbance in the most detailed portions of one of the test cards I had input to the TV. I also noticed that there was a a small amount of colour bleed, where the colours in the image were noticeably offset from the black and white portion of the picture. Fortunately, the TV has a “PC Mode”, which is designed for use with a computer (hence the name). This skips certain video processing steps in order to reduce input lag, and also completely cured these two issues - what a surprise(!)
Up to this point, I had set the Philips 47PFL9664 up as best I could with only my eyes (that is, without the precision of a colour measuring device and associated software). Although I could have experimented with adjusting colour temperature by eye to neutralise the green tint, this would by no means be exact, and would likely have just introduced another colour bias into the image. I ran a set of measurements for a scientific assessment of the display’s output:


As we can see on the charts, there is an excess of green in the TV’s Greyscale mix. However, Gamma is, with this error aside, very accurate indeed, which bodes well for the TV’s post-calibration performance. Measuring the primary and secondary colours made it evident that although absolute accuracy hasn’t been obtained, the 47PFL9664 comes very close already and thankfully doesn’t attempt to produce a wide colour gamut. There is no colour management system on this display, but there are no stand-out errors in this department already, and the colour reproduction is highly watchable.


As is so often the case with modern displays, calibrating Greyscale is possible via the user menu controls and doing so reduces errors to negligible levels. Like the rest of Philips’ current HDTV lineup, the “Custom Tint” controls are slightly limited. The high end Red control is already at its maximum position, so errors in this region have to be corrected by adjusting Green and Blue. Furthermore, at the low end, there is no control over Blue, so adjustments have to be made with Red and Green. Fortunately, this unusual layout doesn’t appear to have had any ill effects in this case. Gamma tracking is also outstanding, although you’ll be able to see the slight excess of Blue greyscale at 90-100 IRE reflected in the chart. Had Philips had a dedicated "Contrast" control (rather than the shared Backlight/Contrast control present here), I could perhaps have fixed this, but in any case, it's not really noticeable. There is no Colour Management System on this display, but I was able to improve colour accuracy by reducing the single “Colour” control by a few clicks (from 58 to 54). This brought most errors down to very small levels, at the expense of desaturating red slightly. A CMS would, in theory, allow a “no compromises” correction.
Earlier, I described how the 47PFL9664’s built-in video processor introduces subtle flaws in the image, which can thankfully be side-stepped by turning on the “PC Mode”. I moved on to test how well the display processes different types of Standard Definition video.
First, I tested the TV’s Diagonal Interpolation capabilities. Diagonal Interpolation is the process whereby steep edges in interlaced video (such as televised sports, soap operas, etc.) are smoothed in order to avoid appearing obviously jagged. The TV did a good (but not amazing) job here. If you’re familiar with the HQV test disc deinterlacing pattern which features the three rotating bars, then it’ll be helpful to know that the 47PFL9664 managed to smooth the top two bars (except for some flickering at their corners), and made the bottom bar look somewhat smooth. This is a decent result.
The film cadence detection tests were less remarkable. Unfortunately, the TV failed to detect the presence of Film content in the HQV test clip (PAL 2:2 test), meaning that vertical resolution was sacrificed when playing film content from an SD source. I should note that the decent diagonal interpolation (mentioned above) suppressed the resulting flickering quite well, but this is of little consolation when you consider that there should be no flickering at all. American NTSC style content, as usual, fared better, with the 3-2 cadence being successfully detected and compensated for. (There are technical reasons for this that make sense, but I always remark about the irony of a TV from Holland handling American content in a superior way to the types of signals broadcast in its native land).
Finally, the quality of the actual scaling (interpolation of pixels to fill the HD panel) was good. There was a small amount of ringing around fine details, which is common for in-TV scaling. Most SD content does not really reveal this limitation, so the scaling performance is perfectly serviceable.
The Philips 47PFL9664 lagged at around 30ms in all modes. This is a decent result and is considerably better than previous Philips displays. Interestingly, the “PC Mode” didn’t appear to make any difference to input lag, only to Chroma resolution. Additionally, using sources that output full resolution chroma (such as games consoles and computers in RGB mode), I was delighted to see that the colour stayed at full resolution on this TV. The video processors in many TVs downsample the chroma components, which means that tiny pixel-coloured details become smudged. This has a mild to non-existent effect on the performance of Blu-ray Discs (if your player has fancy chroma upsampling, the benefits may be lost if your TV doesn’t treat the signal well), but becomes blatantly obvious if a computer is ever hooked up to the display, since we’re so used to looking at laptop and PC monitors that reproduce full chroma resolution. I’s excellent to see that Philips’ TV reproduces full chroma, too, and the picture looks more naturally vibrant as a result. Hopefully we see more and more displays in the future following this example.
Since the 47PFL9664 uses CCFL tubes as a light source, the energy consumption is constant and depends on the intensity of these tubes. On the Philips TVs, this is controlled by the “Contrast” setting in the menu: from levels 63-100, this setting is affecting the Backlight intensity (below that, it becomes a digital adjustment and starts to adjust white level). Using a Contrast setting of 90, the TV consumed 225 watts at all times. Obviously, this level would fluctuate if I had enabled the Dynamic Backlight system, but I chose not to because of the noticeable light fluctuation.
Earlier, I described how the 47PFL9664’s built-in video processor introduces subtle flaws in the image, which can thankfully be side-stepped by turning on the “PC Mode”. I moved on to test how well the display processes different types of Standard Definition video.
First, I tested the TV’s Diagonal Interpolation capabilities. Diagonal Interpolation is the process whereby steep edges in interlaced video (such as televised sports, soap operas, etc.) are smoothed in order to avoid appearing obviously jagged. The TV did a good (but not amazing) job here. If you’re familiar with the HQV test disc deinterlacing pattern which features the three rotating bars, then it’ll be helpful to know that the 47PFL9664 managed to smooth the top two bars (except for some flickering at their corners), and made the bottom bar look somewhat smooth. This is a decent result.
The film cadence detection tests were less remarkable. Unfortunately, the TV failed to detect the presence of Film content in the HQV test clip (PAL 2:2 test), meaning that vertical resolution was sacrificed when playing film content from an SD source. I should note that the decent diagonal interpolation (mentioned above) suppressed the resulting flickering quite well, but this is of little consolation when you consider that there should be no flickering at all. American NTSC style content, as usual, fared better, with the 3-2 cadence being successfully detected and compensated for. (There are technical reasons for this that make sense, but I always remark about the irony of a TV from Holland handling American content in a superior way to the types of signals broadcast in its native land).
Finally, the quality of the actual scaling (interpolation of pixels to fill the HD panel) was good. There was a small amount of ringing around fine details, which is common for in-TV scaling. Most SD content does not really reveal this limitation, so the scaling performance is perfectly serviceable.
The Philips 47PFL9664 lagged at around 30ms in all modes. This is a decent result and is considerably better than previous Philips displays. Interestingly, the “PC Mode” didn’t appear to make any difference to input lag, only to Chroma resolution. Additionally, using sources that output full resolution chroma (such as games consoles and computers in RGB mode), I was delighted to see that the colour stayed at full resolution on this TV. The video processors in many TVs downsample the chroma components, which means that tiny pixel-coloured details become smudged. This has a mild to non-existent effect on the performance of Blu-ray Discs (if your player has fancy chroma upsampling, the benefits may be lost if your TV doesn’t treat the signal well), but becomes blatantly obvious if a computer is ever hooked up to the display, since we’re so used to looking at laptop and PC monitors that reproduce full chroma resolution. I’s excellent to see that Philips’ TV reproduces full chroma, too, and the picture looks more naturally vibrant as a result. Hopefully we see more and more displays in the future following this example.
Since the 47PFL9664 uses CCFL tubes as a light source, the energy consumption is constant and depends on the intensity of these tubes. On the Philips TVs, this is controlled by the “Contrast” setting in the menu: from levels 63-100, this setting is affecting the Backlight intensity (below that, it becomes a digital adjustment and starts to adjust white level). Using a Contrast setting of 90, the TV consumed 225 watts at all times. Obviously, this level would fluctuate if I had enabled the Dynamic Backlight system, but I chose not to because of the noticeable light fluctuation.
Picture Quality
Even with solid Greyscale and Colour performance, the LCD panel itself is often the Achilles’ heel of flat panel displays. Manufacturers often have agreements with different LCD manufacturing facilities who tend to produce panels with recurring strengths and weaknesses, meaning that the most expensive LCD TVs do not always have the best overall panel performance. The panel fitted to the Philips 47PFL9664 is at the better end of the LCD spectrum, but is not the best-performing panel I’ve seen.
First, the good: it does not have any stand-out problems with motion, beyond the typical low motion resolution that LCDs display. By this, I mean that the usual LCD blur is uniform and not obviously confined to certain colours or shades (for example, the Philips Cinema 21:9 widescreen display, reviewed earlier, had issues where black objects on the screen had a higher amount of motion blur than others).
I checked out some scenes from the Region A Blu-ray Disc of “Ponyo”, which is the sort of content that is kind to LCD’s limitations (it has low overall motion and almost entirely fills a 16:9 TV, so there are no black borders to draw attention to the panel’s black level). After calibration, the picture was mesmerisingly clear when viewing the panel face-on, and was free of the small amount of panel-generated noise that a Plasma display introduces. The total lack of noise gave the hand-drawn animation a delightfully silky look. Additionally, it was a very good example of the problems introduced by the TV when not running in “PC Mode”: the slight colour bleed was especially visible during the film’s more highly saturated moments and most viewers would probably distinguish the resulting lack of absolute clarity in the image, even if they weren’t aware of the reason. Fortunately, there is no reason to tolerate this error when it can be removed by enabling PC Mode. I also made sure to confirm that when the PC Mode is on, 24p content is still displayed clearly without judder. (Some “PC Mode” implementations assume you are sending 60hz video and introduce judder into everything else).
Live action content could look similarly great provided the image was viewed face-on, and provided the scene wasn’t overly dark. Darker shots began to reveal the limitations of the LCD panel’s ability to produce deep blacks.
There is also a moderate amount of viewing angle fall-off. The picture characteristics change quite noticeably when you aren’t viewing the screen face-on (although this might not be a problem for you, depending on your tolerance for the issue and your viewing position).
The screen uniformity of the 47PFL9664 I received for review was largely good, although there was a noticeable mura defect (a “stain” of light) at the top left of the LCD panel, visible on dark scenes. The unit I reviewed was supplied by one of our sponsors rather than being hand-picked from Philips’ review stock, but remember that uniformity is highly dependant on the individual TV. Blacks were not bad by LCD standards, but they are behind several significantly cheaper Plasma TVs.
First, the good: it does not have any stand-out problems with motion, beyond the typical low motion resolution that LCDs display. By this, I mean that the usual LCD blur is uniform and not obviously confined to certain colours or shades (for example, the Philips Cinema 21:9 widescreen display, reviewed earlier, had issues where black objects on the screen had a higher amount of motion blur than others).
I checked out some scenes from the Region A Blu-ray Disc of “Ponyo”, which is the sort of content that is kind to LCD’s limitations (it has low overall motion and almost entirely fills a 16:9 TV, so there are no black borders to draw attention to the panel’s black level). After calibration, the picture was mesmerisingly clear when viewing the panel face-on, and was free of the small amount of panel-generated noise that a Plasma display introduces. The total lack of noise gave the hand-drawn animation a delightfully silky look. Additionally, it was a very good example of the problems introduced by the TV when not running in “PC Mode”: the slight colour bleed was especially visible during the film’s more highly saturated moments and most viewers would probably distinguish the resulting lack of absolute clarity in the image, even if they weren’t aware of the reason. Fortunately, there is no reason to tolerate this error when it can be removed by enabling PC Mode. I also made sure to confirm that when the PC Mode is on, 24p content is still displayed clearly without judder. (Some “PC Mode” implementations assume you are sending 60hz video and introduce judder into everything else).
Live action content could look similarly great provided the image was viewed face-on, and provided the scene wasn’t overly dark. Darker shots began to reveal the limitations of the LCD panel’s ability to produce deep blacks.
There is also a moderate amount of viewing angle fall-off. The picture characteristics change quite noticeably when you aren’t viewing the screen face-on (although this might not be a problem for you, depending on your tolerance for the issue and your viewing position).
The screen uniformity of the 47PFL9664 I received for review was largely good, although there was a noticeable mura defect (a “stain” of light) at the top left of the LCD panel, visible on dark scenes. The unit I reviewed was supplied by one of our sponsors rather than being hand-picked from Philips’ review stock, but remember that uniformity is highly dependant on the individual TV. Blacks were not bad by LCD standards, but they are behind several significantly cheaper Plasma TVs.
Conclusions
With the 47PFL9664, Philips have produced a gorgeous looking, large screen HDTV with a few unique, but inessential features. Certain aspects of its picture quality are very accurate (especially so after calibration), but the image quality, whilst generally fairly good, is never remarkable unless the content you’re watching happens to be especially suited to avoiding the limitations of the LCD panel.
Sadly, I have to end this review as I so often do with LCD televisions. Although there are many likeable attributes of the Philips 47PFL9664, the sad reality is that LCD still imposes a number of performance limitations which make its place in premium TV products questionable. There are cheaper competing LCD and Plasma displays which produce deeper blacks with less viewing angle limitation, meaning that the 47PFL9664’s price tag will be a tough pill to swallow unless you’re absolutely sold on AmbiLight or Net TV.
Sadly, I have to end this review as I so often do with LCD televisions. Although there are many likeable attributes of the Philips 47PFL9664, the sad reality is that LCD still imposes a number of performance limitations which make its place in premium TV products questionable. There are cheaper competing LCD and Plasma displays which produce deeper blacks with less viewing angle limitation, meaning that the 47PFL9664’s price tag will be a tough pill to swallow unless you’re absolutely sold on AmbiLight or Net TV.
Scores
Contrast/Dynamic Range/Black Level
•
•
•
•
•
•
.
.
.
.
6 Screen Uniformity
•
•
•
•
•
•
.
.
.
.
6 Colour Accuracy
•
•
•
•
•
•
•
.
.
.
7 Greyscale Accuracy
•
•
•
•
•
•
•
•
.
.
8 Video Processing
•
•
•
•
•
•
•
.
.
.
7 Picture Quality
•
•
•
•
•
•
•
.
.
.
7 Sound Quality
•
•
•
•
•
•
•
•
.
.
8 Smart Features
•
•
•
•
•
•
•
•
.
.
8 Build Quality
•
•
•
•
•
•
•
•
.
.
8 Ease Of Use
•
•
•
•
•
•
•
.
.
.
7 Value for Money
•
•
•
•
•
.
.
.
.
.
5 Verdict
•
•
•
•
•
•
•
.
.
.
7
7
OUT OF
10
Our Review Ethos
Read about our review ethos and the meaning of our review badges.
To comment on what you've read here, click the Discussion tab and post a reply.