220mbps
 

Hi,

I don't understand, why everyone only uses/talks about 160mbps long gop with nano flash. As I read on different pages the nano can do 220 mbps in i-frame only mode.
Isn't 220mbps better than 160mbps and isn't i-frame only better than a long-gop mode?

Please help me to understand that.

best
Axel

Dear Alex,

Many people associate Long-GOP with HDV.

HDV is harder to edit, in that the footage is 1440 x 1080 and not 1920 x 1080. So the computer has a great deal of work to do just to convert HDV footage to 1920 x 1080 for editing.

We record the full 1920 x 1080, always, even with 1440 x 1080 cameras, so this major computer workload is eliminated.


Our Long-GOP is 2 to 2.5 times as efficient as I-Frame Only. Thus, for a given quality, Long-GOP produces smaller file sizes and longer recording time.

No, 220 Mbps I-Frame only is not superior to 100 Mbps Long-GOP. But this is subjective. I do not have any definitive tests performed by independent experts. At best, they are roughly equal.

Our 140 Mbps and 160 Mbps Long-GOP are superior to 220 Mbps I-Frame Only.

In general, for almost all shoots where high quality is a must, we recommend 100 Mbps Long-GOP.

If one's editing system does not work well with high bit-rate Long-GOP footage, then we recommend I-Frame Only.

As always, we recommend that everyone run their own tests with their own hardware and workflow. We support both I-Frame Only and Long-GOP, so it is just a matter of selecting the codec type in the menu.

I am even more confused now.

We are using a EX3 where I thought we have long-gop with 1920x1080.

The EX3 records internally using long GoP Mpeg 2 at 35 Mb/s. The camera also outputs the raw pictures from the camera head via the HD-SDi connector. This output bypasses any compression, it is uncompressed.

You can take the uncompressed output and record it with an external device, such as the NanoFlash. The external device then determines the type of codec you use.

The 100 Mb/s NanoFlash files are of extremely high quality and as Dan says almost certainly at least equal too or better than the 220 Mb/s I frame only files. The beauty of the 100 Mb/s files is that they take up less space on the cards and are quicker to copy or backup than the 220 Mb/s files. You can also edit 100Mb/s footage off USB hard drive, although I wouldn't really recommend it. You can do that with the 220Mb/s files and if you are dealing with several streams or layers of 220Mb/s footage you will need a hard drive system with very good performance.
As CPU's get more powerful decoding long GoP gets easier and easier. It's really very easy these days to edit HD 1920x1080 Mpeg 2 material, even with a laptop so the need to use I frame only is less than it used to be.

For me the advantages of the smaller files, longer record time, faster transfers and simpler hard drive requirements far outweigh any very small decoding performance advantage that I frame may have. For that reason I prefer using Long GoP.

Alister, why don't you use 160mbps long gop?

I know the internal long-gop codec the EX3 has, but I was confused by what Dan said about long-gop and 1440x1080.

I have always assumed that the I frame codecs will handle high-motion situations better though.
Steve

Dear Axel,

Some cameras, not the EX1/EX3, have 1440 x 1080 sensors. This includes some of the CineAlta cameras. But, the HD-SDI output is always 1920 x 1080.

We always record 1920 x 1080, which makes the editing experience more responsive.

Your EX1/EX3 cameras have a 1920 x 1080 sensor, and outputs 1920 x 1080 and we record 1920 x 1080, which is all good.

I-Frame Only 220 Mbps is not as efficient, not as sophisticated, as Long-GOP.

Your camera records internally at 4:2:0 at 35 Mbps. If you record using a nanoFlash at 100 Mbps, you get 4:2:2 which is much better.

To learn the difference between 4:2:0 and 4:2:2, you can enter "Chroma Subsampling" in Wikipedia.

Ya, 4:2:0 vs 4:2:2 is no prob, know about that.
It was just about the bandwidth, but I think you answer my question, at least I feel better informed now.

Dear Steve,

I feel that is a myth.

All of the detail and motion is preserved with our high bit-rate Long-GOP modes.

While I have run tests to prove this, it would be best for others to post their individual experiences.

Remember, we can be neutral in these format discussions since we support both equally well. But, I do have an opinion as to which I feel is more efficient.

100 Mbps Long-GOP creates excellent images, even with an extreme amount of detail and motion. 140 Mbps and 160 Mbps are even better in the most extreme of cases.

220 Mbps I-Frame only also creates excellent images, but I still prefer 100 Mbps for all of the reasons Alister stated. I also feel that 100 Mbps Long-GOP has an edge in image quality.

100 Mb/s is the sweet spot. I don't see any advantage at going to 160 Mb/s. At 100 Mb/s I feel the pictures are already so close to uncompressed quality that going to a higher bit rate brings no real advantage. I have no issues with motion, even 50Mb/s handles motion very well, at 100 Mb/s I don't see any motion issues.

Dear Friends,

We agree with Alister.

While we offer 140 Mbps and 160 Mbps, the 100 Mbps Long-GOP is the "sweet spot".

If one was shooting a feature, and one had an extreme amount of detail in a scene, and one had lots of motion in the same scene, and the camera was hand-held, I would recommend 140 Mbps for this scene. 100 Mbps would work just fine, but 140 gives just a little added insurance in this very specifc case.

In almost all cases, 100 Mbps, 140 Mbps, and 160 Mbps would yield identical results, in my personal opinion. One may choose any one of these and be confident in the image quality.

Once a codec, at a given bit-rate, is "Good Enough", then going to a higher bit-rate offers very little, if any advantage.

By "Good Enough", I mean perfect, or very close to it.

By "Good Enough", I mean that 100 Mbps footage, when examined frame by frame, will reveal almost no artifacts for hours of very detailed, motion intense images. At normal playback speed, no artifacts will be detected.