Videomaker Reviews the NewTek NDI HX-PTZ1 Camera

Not long ago I openly admitted that I wear the rants in the family. This goes along those lines. It was inspired by the Chris Monlux’s review of the NewTek NDI HX-PTZ1 Camera published January 14th by Videomaker Magazine. I found this review to be deeply disappointing, and I’d like to tell you why.

First, take the time to read their review. I’ll wait. And you need the context.

I’ve been writing review since the early 90s. I started when working for a small print magazine. I learned a tremendous amount from the owner and publisher. She insisted that we actually use the products under review. There was no substitute for substantive, hands-on experience. Further, we had find the good and the bad in each product. Nothing is perfect, so present a balanced view.

The Videomaker review of the PTZ1 fails to meet my expectations in a variety of ways. The author spends a lot of time addressing the workflow implications of IP vs traditional SDI. He talks about how great it is to remote control the camera over the same wire that handles power and signaling. He describes its integration with a Newtek Tricaster production switcher.

That said, in a thousand-word review he doesn’t really touch on its performance as a camera. Does it produce high-quality images?  Is it noisy? How does it handle dark scenes? What sort of color setup or camera matching control does it have?

The PTZ1 has 3 outputs: HD-SDI, HDMI and NDI. Are they all simultaneously active? How do they differ?

He mentions “NDI” multiple times, only making passing reference to NDI|HX in a summary table. How do they differ?

Fortunately, I can fill in some of these details. I believe that all the outputs on the the PTZ1 are active at the same time. However, they are a bit different.

The PTZ1 delivers NDI|HX, which is a reduced bandwidth version of NDI. It’s essentially a proprietary form of H.264. It was designed to make things easier for users, but at a price.

What we’ve come to know as full-NDI (not HX) includes an intraframe compression scheme that requires as much as 150 mpbs for a 1080p60 stream. This is why you will hear it said that “NDI requires gigabit Ethernet networking. ” Since each frame is processed in isolation, it adds something less that one frame of delay. There’s effectively no latency. At least, zero frames.

In contrast, NDI|HX squeezes the stream down to just 15 mbps using interframe compression. There’s good reason to do this. In the case of the Newtek Spark, a Wi-Fi capable NDI encoder, it makes it practical to send the resulting stream across a real-world Wi-Fi link.

Convenience always has a price. In this case, its two-fold: image quality and latency. Long-GOP compression exacts a toll on image quality. Some will care. Others will not. It really depends upon your application.

Long-GOP compression also inserts some delay into the stream. It’s not as much as traditional H.264 encoder, but it’s not zero. It’s unclear if the SDI and HDMI outputs are also delayed to stay coincident.

When deployed along side SDI cameras that are not NDI|HX-based, the PTZ1 output will be slightly late. Here again, some will care. Others will not.

Some in the video production community feel that long-GOP compression has no place in production. It’s acceptable in transmission or delivery, where latency doesn’t really matter. I can’t really disagree with that perspective.

I do understand that Newtek needed to find a way to deliver some form of NDI reliably over commodity wireless networks. If they didn’t, it would dramatically increase their support burden, and leave an opening for competing video-over-IP technologies to gain a foothold.

To be powered from the network, the PTZ1 requires 802.3at power-over-Ethernet (aka POE+.) That’s a POE standard that’s not exactly rare, but also not yet commonplace. Switches meeting 802.3af standard are most common, being used to power IP phones, Wi-FI access points and security camera for many years.

If you want to power the camera over the network (and you should!) you’ll need to make sure that your network switch or POE inserter is up to the task.

Finally, the author makes no effort to show any images of the camera configuration menus. Nor does he offer any sample clips.

Since the review was presumably written for a magazine I can somewhat forgive the lack of sample video. There’s no way to deliver it in print. When the review was loaded to their website the lack of it leaves me wanting.

It would have been especially interesting to know how the PTZ1 compares to the Lumens VC-A50PN, since they are apparently based upon the same OEM hardware.

I certainly appreciate that NDI is new and shiny. It’s inclusion in a PTZ camera brings the opportunities to explore new, all-IP workflows. It’s exciting. Even so, the PTZ1 is a camera and needs to be considered for those things that cameras do. Evaluated alongside its peers as a source of imagery. The Videomaker review could have been written from the spec sheet, without ever laying hands on the device. To my mind, it leaves far to many questions unanswered.

I left a list of my concerns about the review in a comment. Three weeks later there’s been no response. Perhaps I should approach Videomaker about writing reviews?