Jump to content

Recommended Posts

Hi, Looking for some advice please.

I have a IC Realtime 2TB Breeze HD 4 channel NVR

with 4 x 1.3 Megapixel IP Infa-Red Cameras connected.

each camera is connected to the NVR via a Cat6 cable

and powered via POE from the NVR.

The NVR is connected to my broadband router and i

have broadband speeds of about 4Mbps download.

Just want to confirm the best possible encode settings for the

camera's to get the best possible picture quality without

losing too much recording memory on my Hard drive.

All four camera's are recording 24/7.

 

The thing i'm really not sure about is where to set the Bit Rate

on the Main stream and Extra Stream? Does this setting effect the

quality of the picture locally and when viewing remotely,

can someone explain this to me in simple non-tech terms.

 

My current Encode settings are as follows on all 4 cameras:

Normal Stream - H.264 encode, Resolution 1.3M, frame rate 10fps,

Bit Rate Type - CBR, Bit Rate - 1024kbps

Extra Stream - Resolution D1, frame rate 10fps, bit rate type - CBR

Bit Rate - 1024 kbps

 

Best Regards M.

Share this post


Link to post
Share on other sites

My current Encode settings are as follows on all 4 cameras:

Normal Stream - H.264 encode, Resolution 1.3M, frame rate 10fps,

Bit Rate Type - CBR, Bit Rate - 1024kbps

Extra Stream - Resolution D1, frame rate 10fps, bit rate type - CBR

Bit Rate - 1024 kbps

 

Having both the main stream (1.3Mp) and substream (D1) at 1024kbps doesn't make any sense at all...

 

Also, check your upload bandwidth; the download bandwidth doesn't matter at all, when what you have to do is send data, not receive it.

Share this post


Link to post
Share on other sites

That's why i'm asking for some advice, i'm not a cctv installer.

How do i check my upload bandwidth?

If you mean by speedcheck, i just did one and

my upload speed is 0.838 mb/s

So what should i set both the bit rates too?

If being the same makes no sense. Sorry i don't

understand it.

many thanks

Share this post


Link to post
Share on other sites

I run my two 3MP cams at 8192, eleven 2MP cams at 8192, and another VGA wireless PT camera at something else (who knows - it is a glorified baby cam).

 

The two 3MP cams run at 20FPS (they're used for ANPR with 12mm lenses and a min of 1/1000 shutter on a residential - up to 30 MPH street).

 

Two of the eleven 2MP cameras run at 20 fps while all of the other run at 15 fps.

 

Point is I can run all of that on continuous recording and I get just short of two weeks of footage. I have 10 TB dedicated to my NVR server.

 

I personally think the bit rate should always be maxed.

Share this post


Link to post
Share on other sites

Bit rate is the h.264 way of specifying compression. Back in the old days, like 3 years ago, most cameras were MPEG or MJPEG and you specified compression as a percentage but with h.264 you do it by bitrate and the lower the number, the higher the compression, meaning the camera and NVR work harder to decode the compressed stream, but takes up less network bandwidth and space, but you get lower quality. So it's up to you to experiment and see at what point it bothers you. For me, with 3MP cameras, I found 5K to be the sweet spot with noticeable compression artifacts at 4K and not much difference after 5K but everyone's needs are different.

 

To further confuse you as that's my intent, there's different h.264 profiles and some cameras you can specify them. There's Baseline (lowest), Main and High (highest). Lower like Baseline means lower compression, Main and High are higher compression.

Share this post


Link to post
Share on other sites

Look at your recorded mainstream footage. If your settings satisfy your needs and the footage looks good, keep mainstream where you have it. Try it higher and see if you appreciate the difference. Typically you keep substream lower to make it easier to view your cameras via cell phone or internet remotely. This is especially true if you open all four at once. If the camera feeds aren't bogging down and freezing, you're OK. If you have problems, try D1 at 512 or even a little lower. It will not effect your higher resolution mainstream recordings. If you're recording substream as well that's fine, it'll be lower resolution recordings and can be useful if you want to check playback remotely as well. You'll always have the higher resolution mainstream files for the best quality. That's the benefit of two separate streams.

Share this post


Link to post
Share on other sites

Maybe this helps. I hope so, it took me about 30 minutes to get these screen shots and info put together. Here is the way I have my cameras configured and this is the network utilization result. At this time I have ten 2MP Dahua (the 11th one is not being used right now), two 3MP Hikvision, and one VGA SunEyez Glorified baby monitor:

SunEyezVGAsettings.png.8ee5776b2b925058f1140210c774b8d1.png

Hikvison3MPsettings.png.0e1d707f5835a6b010265b490da4c478.png

Dahua2MPsettings.png.e1abfb9808b28b47bc458b368138ad37.png

Edited by Guest

Share this post


Link to post
Share on other sites

So you can see from the analysis, the total Bytes per second for each camera represent the total throughput from the IP camera's two streams (both Main and Sub Stream). Remember, on the camera configuration screen the streams' bit rates are defined as bits per second (not Bytes).

 

You will note when I had my client set to Visualization Profile for all cameras, that was essentially taking a gauge of the total bandwidth for all cameras' substream settings at once (equaling about 4 MB per second) while the other client, the X60, shows you the demand for seeing four channels of substream at once (about 1 MB per second).

 

Compare that to when I placed my client on Recording Profile for all cameras (this is the Main Stream feed from the cams). You can see the bandwidth demand on all 13 came to a total of about 11.2 MB per second. That is about three times more than the all substream demand. So you might be safe to say that a single 2MP camera shown at 1.3MB per second is comprised of a Mainstream feed of about 1 Megabyte per second and a substream feed of 336 Kilobytes per second (with my configuration settings).

 

To recap: the total B/sec seen on the network activity represents both sub and main stream. Selecting the stream to view from your client will determine which portion of that bandwidth is sent for for remote viewing. You can set your mainstream up for recording high and your substream lower to accommodate remote or mobile viewing if you need to.

 

I am not suggesting you use my substream settings. I have mine set this way specifically because I utilize the substream on my NVR for analytics. I am providing this only as an example to explain how this works. Additionally, I have somewhat different internet speed than you:

InternetSpeedTest.png.b7222f69a089df5839747d32ba63164b.png

Share this post


Link to post
Share on other sites

My experience is that increasing bit rate helps to a certain point, then stops having any effect except bigger file sizes and higher network loads. This point will be different depending on your cam, resolution, frame rate, the level of detail in the scene, and the amount of motion.

 

Don't test bit rate on a scene with little or no motion, as it may look good even at pretty low bit rates. Motion and detail are critical to bit rate setting tests.

 

Hik has a PDF of recommended bit rates for different resolutions, quality, and frame rates that's a pretty good general guideline:

ftp://hikfirmware:Hikvision123@ftp.hikvisionusa.com/Technical%20Bulletin/2014/IP%20Camera%20Recommended%20Bitrate%20v2.0_20140219.pdf

 

For instance, they recommend these settings for the popular 2CD2xxx cams:

1080p, 15 fps, best quality: 3584 kbps

704x480, 15 fps, best quality: 1024 kbps

 

I run my 2MP and 3MP cams at 10 fps, 4096 kbps, and can't see any difference going to higher bit rates.

Share this post


Link to post
Share on other sites

Thank you everyone so much for your input and knowledge,

it's really helped me get to grips with this, and i'm sure it will

help others searching this great forum in the future.

I'll be raising my main stream and doing lots of bit rate tests this week for sure!

Best Regards

Share this post


Link to post
Share on other sites

When adjusting bitrates, keep a stream open to see what it does to the image. Not only bitrate, but other settings, like in the Axis P3301 i took the images from, there is a setting to prioritize framerate OR image quality. These streams are framerate priority at 30fps.

If i choose image quality priority, framerate drops to close to 1-2fps and image quality is equal to high bitrate.

 

1024kbps, 640x480, 30fps, GOV 1, framerate priority

 

 

8120kbps, 640x480, 30fps, GOV 1, framerate priority

 

 

So when adjusting the bitrate, keep a stream open and stop lowering the bitrate when the image starts to pixelate or blur. Also something worth keeping in mind is image with no motion requires WAY less bandwidth/bitrate than image with motion.

8120kbps.jpg.818f19756820bff8ad0d1143ef5aeddd.jpg

1024kbps.jpg.46981416432ebe31724c5ff29ae54c7b.jpg

Share this post


Link to post
Share on other sites

Is GOV the same as i-frame or GOP? I-frame = 1 would be essentially the same as mjpeg, which is a pretty high bandwidth setting and would account for the terrible image at 640x480 and 1024 kbps. Normally I'd expect better image quality from that setting with a typical i-frame of 30 or 60.

Share this post


Link to post
Share on other sites
Is GOV the same as i-frame or GOP? I-frame = 1 would be essentially the same as mjpeg, which is a pretty high bandwidth setting and would account for the terrible image at 640x480 and 1024 kbps. Normally I'd expect better image quality from that setting with a typical i-frame of 30 or 60.

 

Yes, GOV is the amount of P-frames sent before next I-frame, in this example I just lowered it to 1 for maximum effect on what to expect when tweaking the bitrate.

 

For those reading and not sure what I- and P-frames are:

I-image is a complete image

P-image is only differences in the image as compared with the previous image

 

So to save bandwidth, full size images (keyframes) are only sent in the beginning of set of frames, for example:

30frames per second stream, first image of the second is I-image/keyframe, rest 29 are P-images that only send the differences. This might appear as sort of pulsating image though, as if the camera has a heartbeat.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×