Testing 'Megapixel' Analog CameraBy: Derek Ward, Published on Mar 26, 2014
Megapixel analog sounds like a contradiction in terms. 'Analog' by definition, or at least in common use, is constrained by 60+ year old NTSC / PAL specifications and cannot be 'megapixel.'
However, now manufacturers are starting to incorporate megapixel sensors into analog cameras. While the cameras must output NTSC / PAL analog, the imagers are capturing megapixel.
Many argue that this is nonsense, that there can be no image improvements and that it is simply manipulative marketing.
They are wrong.
We bought a 900TVL camera and tested it alongside our analog resolution setup (including 450TVL, 600TVL, 700TVL, 960H, etc.) matched up against true 1.3MP and 2MP IP cameras.
900TVL cameras delivered a clear improvement in image quality compared to all of the other analog cameras.
Questions we examine inside, with image comparisons:
- How much better quality was 900TVL compared to 960H, 700TVL and 600TVL cameras?
- What was the impact in bandwidth consumption?
- How did quality vary between night and day?
- How did 900TVL match up against HD / MP IP cameras?
- How did performance vary between moderate (~20') and wide (~40') Field of Views?
- What should you use for specific applications?
- The "mega pixel" analog camera (900 TVL) beat all other analog cameras tested (450,600,700,960H/700 TVL).
- 900 TVL reduced bandwidth, improved image quality, and low light performance compared to the 960H.
- IR performance of the 900 TVL was strongest amongst the analog cameras tested, with the lowest amount of artifacting and gain in the scene.
- A "mega pixel" analog camera is unable to produce the same level of detail that a megapixel IP camera in full light, dark, and outside scenes (especially in wide scenes).
- A 900 TVL camera connected to a D1 DVR produces significantly lower image quality than connected to a 960H DVR.
- Bandwidth consumption of the 900 TVL was ~2x lower than the 720p IP camera, and ~8x lower than 1080p. Also, 900 TVL consumed less bandwidth than the 960H/700TVL analog camera, our theory being that the 900 TVL image is cleaner than 960H one, resulting in less demanding encoding for the same pixel count.
- Q-See QM6006B 600 TVL analog bullet
- Q-See QM9901B 900 TVL analog bullet
- Q-See QM9702B 960H/700TVL analog bullet
- Q-See QCN7001B 720p IP bullet
- Avigilon 2.0W-H3-B01 1080p IP bullet camera
- QT524 4-channel 960H DVR
- QT534 4-channel D1 DVR
The big challenge in deciding what to do is that 'megapixel analog' requires both new cameras (i.e., 900 TVL / 1.3MP sensor ones) and new recorders (960H) to deliver the potential quality improvements.
The main attraction is that these cameras and recorders are relatively inexpensive (the MP analog one is ~$60 each, a 4 channel 960H recorder is ~$150, plus HDD, total ~$500 compared to ~$700 for a 4 channel MP IP kit).
If each camera is only going to cover a narrow FoV of 10 feet or less, with a low camera count and simple needs, 900TVL looks to be a better choice than IP MP and a meaningful upgrade over moderately less expensive traditional camera counts.
"Mega Pixel" Analog vs. Analog
For reference, this image illustrates the approximate 18' FOV for all cameras.
First we shot out the 900 TVL analog camera to the 960H/700TVL, and 600 TVL analog cameras.
900 TVL vs. 600 TVL and 960H/700TVL Analog Cameras
Starting with 600 TVL vs. 960H/700TVL vs. 900 TVL, we can see that the 900 TVL analog camera outperforms both the 600 TVL and 960H/700TVL cameras. Line 5 of our test chart is legible, while both the 600 TVL and 960H/700TVL cameras are able to capture lines 3 and 4. Also, our subject's face in the 900 TVL image is clearly the most detailed, with more detail visible on his eyes, hair, and mouth.
In a dark scene, with IR on, the 900 TVL analog camera is top performer against the 600 TVL and 960H/700TVL, with low amounts of visible gain and artifacting present, while being able to discern line 4 of our test chart. Both the 600 TVL and 960H/700TVL cameras have high amounts of artifacting present, and the 960H/TVL appears to wash out our test chart with IR illumination.
"Mega Pixel" Analog vs. Mega Pixel IP
Next, we comparing the "mega pixel" (900 TVL) analog camera to a 720p and 1080p IP camera, and take note of the difference between them, if present.
900 TVL vs. 720p and 1080p
We moved onto testing the top analog performer (900 TVL) to a 720p and 1080p IP camera. Although performance is good on the 900 TVL analog camera, it does not surpass the 720p or 1080p cameras, which can distinguish characters on lines 6-7 of our test chart.
The same is true in low light with IR on, with both HD cameras providing more detail:
960H vs. D1 Encoding
While there are no claimed DVR's that are or are not compatible with the 900 TVL analog camera, users should note that connecting a "mega pixel" analog camera to a lower resolution DVR will create gaps in image potential and clarity. In the image below, this is the same 900 TVL analog camera, but connected to both a 960H DVR and D1 DVR.
In the comparison below, past line 3 in the D1 DVR image is illegible, and our subjects face is clearly more pixelated.
Outside Parking Lot
Finally, we took the cameras outside, allowing for a larger testing FOV. Below is the FOV we used for testing.
900 TVL vs. 960H/700TVL
With a 40' horizontal field of view, images from both analog cameras are pixelated and noisy. However, the 900 TVL camera is able to distinguish our subject face and a few characters on our test chart more than the 960H/700TVL camera.
900 TVL vs. Mega Pixel IP
Similar to our indoor scene, the 720p and 1080p clearly perform stronger than the 900 TVL analog camera, with our subject's face and lines 2-3 on our test chart more visible.
Walking across the scene, our subject's profile in the 900 TVL image has strong artifacting, and it is difficult to obtain any useful details, such as clothing or hairstyle.
With our subject walking towards the cameras, the 900 TVL camera still has difficulty producing a clear image of our subject, while the 720p and 1080p cameras can identify few facial features and apparel detail (such as the grey stripe on our subjects jacket).
We tested bandwidth in both full light and dark scenes with IR on. The chart below compares bandwidth consumption on all cameras. We gauged bandwidth for the analog cameras based off the RTSP stream bit rate from the QT524 4-channel 960H DVR to Exacq. A Q of 28 was used for all cameras, with FPS set to 10. Measurements are based off Mb/s.
Overall Resolution Comparisons
For those interested, we have compiled the entirety of the analog and megapixel shootouts into one. In full light:
In the dark:
All cameras were tested using default settings.
These are the camera models used in this test: