Member Discussion

Using HD IP Camera In SD Mode?

Dear IPVM members,

From your experience on testing new IP cameras, do you notice an improvement in using new high resolution cameras (2MP, 5MP…) in term of quality/bandwidth/low light performances when using these cameras in standard definition mode?

In other words, if I have no specific needs in high resolution cameras do I need to still use SD camera or it is a better choice to use brand-new high res cam as they are better even in SD mode (better image quality, better encoding efficiency…).


That's a fascinating question. I understand what you are saying, but I have not heard that suggested before.

Do you really have no specific needs for higher resolution? If so, what are you monitoring?

As a general rule of thumb, if you are covering an area wider than 10' / 3 meters, you are going to get material benefits from using HD over SD. And since almost everyone needs monitoring more than that, the benefit is clear. See: Shootout: Megapixel vs Analog Cameras


The matter is about the global quality improvement of products. With HD products we get an improvement in resolution but do we keep the same quality even in SD mode.

In case of limited storage capacity, I will preferably store the video in SD with a lower bit rate. But would the size of the file be lower (for the same quality) if I use newly designed HD camera.

In practice, do I get a better image in SD mode using a large size image sensor. This question is may be link do the quality of the down-sizing (HD to SD down-sampling) algorithm implemented in the IP camera.

I honestly don't understand your approach at all.

I don't understand why you want to pay more for an HD camera and then not use HD. There are newer analog cameras (700TVL, etc.) that improve SD quality that are much less expensive than IP HD. There's also really low cost analog HD offerings.

Also, I don't see why you are so concerned about storage cost. Either HD provides more meaningful details for the application and therefore is easily justified or it does not and should not be used at all. Furthermore, there's a strong consensus that if you want to save bandwidth, look to reduce frame rate or record only on motion first.

Ultimately, if you go down this route you should check the specific manufacturers you are interested in and see what the cost and performance tradeoffs are. They likely will vary.

First off, HD cameras do not neccessarily have larger-sized image sensors, and sensor size does not always correllate to any form of "better". (See previous IPVM articles on this).

Also, calling the bit of processing that downsamples an image an "algorithm" is probably being a bit generous. Most camera manufacturers want to output the best image possible, but the basic assumption when a camera is capable of [X]MP and you're using a 1/4[X] stream is that you want an "overview" shot with reduced bandwidth, not that you're trying to get more detail out of the image.

I'd be surprised if your theory had any merit in any mainstream camera.

Often higher megapixel cameras perform worse in low light compared too lower resolution models of the same generation., so going with a lower resolution camera will perform better.

Yes, higher resolution cameras are often newer generations, but come with drawback. Choose a camera that is a current generation, and the lowest resolution for your needs. This will keep costs and storage/bandwidth down.

The only advantage is being able to later on up the resolution as needed (such as infrastructure or recorder upgrades).

The high res / low res low light issue is a tough one.

First, lots of cameras (analog and IP) now have integrated IR, so low light sensitivity is less crucial when the camera is providing their own light.

Many HD cameras are developing super low light models that actually beat SD models (similar light sensitivity with more pixels delivering more details overall), so it's hard to make a general rule. (Axis Q1602 x Q1604 is one example - Q1604 is overall better in low light even though HD)