Atomos ninja 1080 50p

[Jeff Allen]

Dale,

I shot the video with a Sony HVR-V1U camera.

I am going to try and do some more test shots this weekend, trying out a few different modes and methods. Also going to try some longer duration shots as well since I have a clinic coming up and one of the reasons I got the Ninja was so that I would not be limited by the 60min HDV tape. I will try to post some side by side/split screen comparisons of the HDV versus the ProRes Ninja. However, I am not sure that there will be that much noticeable visual difference especially when compressed for web delivery, although I will try to zoom in a little as well to try and demonstrate anything I can find. Where I did notice a difference was in the latitude and quality when I was color correcting. Seemed like there was just more there to work with, and the color correction looked very clean and smooth when I was finished.

Of course, not having to log and capture hours of HDV tape, and not being limited to an hour per tape was worth the price of entry for me alone. The additional benefits are icing on the cake.

It may be a week before I can get anything posted, but I will try and come back with a post here once I get something up, or note that there was not enough visual difference to post.

[Dale Guthormsen]

Jeff,

What I am most interested in is whether the pro res image is significantly sharper than the hdv image.

I have used cineform for ages which is a 422 color space and helps with the coror grading big time.

Look forward to what you will post!!!


dale

[Jeff Allen]

Dale,

I always use ProRes-422 as well for color grading, but I used to have to transcode the HDV content to ProRes, then be able to use the 422 color space within Apple Color. However, since the HDV master is from a 4:2:0 color space, then transcoded to a 4:2:2 ProRes then graded, it didn't seem to have the same latitude as I do now. Everything just appeared to grade much smoother and looked better when originated in 4:2:2 from the Ninja (1920X1080 rather than 1440X1080 probably helps as well), which is something I expected, but also really enjoy.

Another general observation, purely subjective, is that the encode for web delivery also appeared to look better. I assume that since it is coming from a better quality source, with less transcoding, it is a better quality output, but maybe it was just my imagination.

BTW, I completed my long shot test today, getting ready to do some other shots now, and tomorrow. The long shot was broken into segments by the ninja (due to the disk format type I assume) in 4.29GB sections. However, when I drop the folder onto the timeline that contains the clips, Final Cut Pro does a nice job of automatically putting them all on the timeline, and in the right order. Not sure what the behavior other NLE's will exhibit with the clips. I shot in ProResLT, which gave me around 5.5 min per clip.

I will do a "side by side" of Ninja-to-HDV with my shorter shots planned for today and tomorrow, will post any observations and video if relevant.

[Jeff Allen]

All, I have uploaded a new test video on Vimeo in 720P of footage that I shot over the weekend to demonstrate the depth and quality of footage that I was able to capture using the combination of the Ninja and my Sony V1U. Hope folks find this useful. Hope to look to the "Bake-off" next of HDV to Ninja in a side by side, but will take me a couple of days to get that done.

Here is the vimeo link:

Atomos Ninja Test Video on Vimeo

Edit: BTW, you can also download the x.264 file from the Vimeo page, which looks even nicer, but a larger file.

[Dale Guthormsen]

Jeff,


finally got back to this, A couple things, does the ninja always throw 4.6 gb clips (like its fat 32) or have you gotten by that??

Also, If you could muster the time and effort!!! I would very much like to see a split screen of hdv and pro res hq footage, both raw with no correction.

I was looking for video on line and I can't really find any that gives the info I need before I spend a grand for a better image.

Now, do the clips not have time code on them? If so how does that go in the edit sweet? Or should I say how do you tackle that?


thank you so much for the help!!!

[David Stuart Shapton]

We currently use FAT 32 so we do split video into 4GB chunks, there are ways to handle this though and FCP does it automatically.

A couple of things: We are going to change the filesystem or make other filesystems optional, so this isn't a permanent feature. And we are very shortly going to implement Time Of Day timecode in the files.

Dave Shapton
President
Atomos EMEA

[Jeff Allen]

Dale,

Looks like Dave did a good job at addressing the technical questions that you raised (Thanks Dave! Always best to get the information straight from the source).

As to the quality, I have looked at the following scenario:
ProResHQ from the Ninja - Matched with HDV captured as ProRes during capture into the Mac from my V1U.

When zoomed in, you can see a major difference between the images. The HDV to ProRes version has visible "blocks" that look rectangular. The Ninja does not exhibit this. Depending upon the content, it is more or less noticeable. I found that on "black" content, it was pretty noticeable. Also, on larger portions of a semi consistent color/smooth surface texture, it also appears blocky/smeary as well. I assume that this is due to the HDV 4:2:0 and 1440X1080 compression compared to the ProRes native (4:2:2, 1920X1080). ProRes is great, but if it is being built from the HDV source, it will inherit its limitations.

I don't see a huge difference in the overall quality when viewing in normal conditions (viewing after compression for the web or delivery on DVD, etc.), but I do see the difference during and after color correction. I think you get more latitude during color correction due to the lack of compression artifacts if working with the ninja footage. Meaning, that if you are doing color correction, or other compositing work, then I think the Ninja is worth it based on image quality. If you are not doing these actions, then you may not see much if any difference. Purely from an image quality perspective, and depending upon your delivery method.

Of course, eliminating the tape base workflow elements can also be very compelling as well. These benefits were some of the major reasons that I purchased.

I am not sure if I output my test, and then compress for the web, if it will really demonstrate the difference. However, I will try to take a few screen shots of the side by side footage and post them so that you can see what I describe above.

One other item, yet to be confirmed, is that my low light performance on my V1U appears to have been improved by the ninja as well. The V1U is known for its poor low light performance, and it appears to me to be improved by bypassing the HDV compression process. But, I have yet to confirm this in specific tests.

Hope this information is useful. Will try to get the screen shots up either late today or tomorrow.

Regards,
Jeff

[David Stuart Shapton]

Jeff,

Thanks for all this feedback! You're also benefiting from the extra 2 bits of resolution that you get when you record in ProRes. Of course, this doesn't improve the quality of the original material, but it does give you more levels to play with when you are grading and will tend to give a smoother result.

Dave Shapton
Atomos

[Jeff Allen]

Excellent point Dave. Forgot about that when mentioning the differences. Therefore, we have the 4:2:2 to 4:2:0 difference, the 1920X1080 to 1440X10180 difference, the bit depth difference, and the overall codec difference (ProRes Intra-frame compared to HDV).

I think all of these stacking up can make quite a difference when performing compositing or color correcting tasks on the footage.

I wish I had a good green screen setup, as I would imagine that keying would be a very different experience using original Ninja footage when compared to HDV original footage.

Regards,
Jeff

[Jeff Allen]

Okay, here is a still image taken from my Ninja-HDV compare video. I think this is probably the best way to demonstrate the differences that I mention above.

This is a still frame grab (TIFF) from my split video. The left side is Ninja footage captured in ProRes from my Sony V1U. The right side is HDV video, as captured onto tape. I then converted to ProRes during the import. Since this is my normal work flow, it was what I used to see if my perception of color correction differences, etc were actually real based on image definition and quality using my existing work flow. No color correction was done on the image. Pure Ninja/ProRes and HDV/ProRes stacked on the timeline.

I did zoom in, so that the differences were more obvious. This is an action shot, taken outdoors in good lighting conditions.

As you can see in the image, there is a big difference between the left hand side and right hand side. I did compress to jpg for posting here, but I don't think it changed the overall difference too much when comparing left to right. (be sure to click on the image to see a larger version)

As I have mentioned previously, I don't see a "major" difference when playing back the video at speed and normal size, although I do believe that the image is better. However, when taking the footage into Apple Color, I believe I see a major difference in the output and the quality. I imagine compositing/keying would also show a major difference as well based on the artifacting that can be seen. I also believe that my encodes look better when compressing for the web in h.264 or x.264.

[Dale Guthormsen]

Jeff,

First, thank you for taking the time to put together this comparison!!

The real sign of the difference is the fence and the trees in the back ground!! They are significantyl better.

My work flow will be down load to Vegas, transcode to m2v for blu ray output.

Color grading is a real important facite of my work so it will be a worhty investment if for no other reason than that!!

I only have two concerns:

1. How do I deal with no timecode (can I put timecode on in post?)

2. How will Vegas (which will work with pro res) deal with the small files!



Thank you again, I know others investigating this product will have the same questions.

[Adam Palomer]

Does the Ninja have the ability to start recording when the user presses the connected camera's record button?

I mean, it would be a plus if the Ninja could be controlled by the camera's own record button.

[Jeff Allen]

Dale, you can use the LANC control pass through if you camera supports it. That is how I usually capture with my V1U. I use a LANC remote for my focus, standby, and record functions. Works pretty well with the ninja. Especially helpful if you are capturing to both tape and the ninja at the same time. I press record on my remote, and the camera and the ninja begin recording simultaneously.

[Charlton Chars]

Without the tape in the V1U it gives you a "No Tape Installed" type of message on the LCD, but that has no impact on the record function or the video going into the Ninja.