DV Info Net

DV Info Net (https://www.dvinfo.net/forum/)
-   Sony XDCAM EX Pro Handhelds (https://www.dvinfo.net/forum/sony-xdcam-ex-pro-handhelds/)
-   -   XDCAM EX 35Mbit/s codec comparison (https://www.dvinfo.net/forum/sony-xdcam-ex-pro-handhelds/482600-xdcam-ex-35mbit-s-codec-comparison.html)

Mike Marriage July 29th, 2010 09:35 AM

XDCAM EX 35Mbit/s codec comparison
 
I did some testing of the XDCAM EX codec against ProRes as a benchmark and I was pretty impressed by the results. I was going to test it against the 50Mbit/s XDCAM variant as well but I may not bother now as I'm pretty sure that the difference will be even less significant.

Mike Marriage - Blog

Olof Ekbergh July 29th, 2010 10:13 AM

The XD codec is great for acquisition but not so good for editing.

The difference from Prores and XD is, intra frame (individual frame compression) vs long gop MPEG.

XD is much more processor intensive to play back and transitions etc don't play back in real time, this depends a bit on your edit setup. And recompressing back to mpeg is not so good in the editing process, if jumping to AE to Color and then back to your NLE.

So I view the XD codec as an excellent "tape" format. But I always edit using a ProRes or similar codec. I use M100 for most of my editing and I can play back 4 or 5 layers of video and graphics in full 1080 at the real framerate w/o any dropped frames in a timeline. No rendering at all. You can just play out to tape or whatever. I use AJA KONA, so it plays out realtime both HD and SD.

Even FCP will behave much better if you use Prores codecs while editing.

I always pick the shots I want and then transcode those for editing. It is very fast depending on your drive and cpu. If it is a long project I do the transcoding overnight. This is done automatically in M100 when dropping clips in a bin.

Mike Marriage July 29th, 2010 10:44 AM

Hi Olof,

I was only really using ProRes as a benchmark for this test. It was more to see whether it was worth using an external recorder like the Nano Flash or AJA Ki Pro. For me, I don't think it is for 99% of what I shoot. I have performed filmouts from PMW350 35Mbit/s footage and it looked first class. I did also play around with grading the footage and the XDCAM held up pretty well - there wasn't that much more there in the ProRes recording. Having said that, in certain situations there would probably be a more pronounced difference.

Olof Ekbergh July 29th, 2010 10:58 AM

Mike,

The NanoFlash actually uses the XD codec for its long gop codec. It also can do intra frame up to 280mb/s.

I use the Nanoflash often and it is really great, there is a difference, it is subtle but it is definitely there. I usually use it at 100mb/sec mpeg in the NF.

One nice thing about the NanoFlash is that it can record at the same time as the SxS cards, so it is great for redundancy. Although I have never had a problem with SxS or the CF's in the NanoFlash.

The NanoFlash is really small and hardly uses any power. It is a great way to get 422 out of the EXcams. The BBC accepts long gop footage shot at 50mb/s and up only.

But I must say that the 35mb/s XD codec is fantastic and very compact for archiving, a huge improvement over 25mb/s HDV.

Mike Marriage July 29th, 2010 11:53 AM

Sorry, I should have been clearer, I wasn't suggesting that the Nano Flash encodes to ProRes, I was just using ProRes to compare the 35Mbit/s codec too. Sometimes a side-by-side can help to highlight flaws. I am very familiar with the NF and its codecs as I have been thinking about buying one for a while. However, this test discouraged me.

If a producer demands 50Mbit/s I will hire a NF but I want to be able to make informed recommendations about kit. I still want to test the onboard recording on the 3350 against a NF and establish how great/subtle the difference is at 35, 50 & 100Mb/s MPEG.

I think "hassle" is a bad thing in real world shooting and adding a NF certainly adds hassle, if only a little. I was trying to decide when that hassle will be worthwhile for my shooting. I'm not prepared to spend time and money on something that doesn't add value for my client. IMO, if the client and their audience can not differentiate between the onboard recording and an external recorder, I don't want to use one.

Simon Denny July 29th, 2010 01:55 PM

Hi Mike,
Thanks for the test. One question, how did you acquire the ProRez footage?

The XDCAM 1090p 25 held up really well with just a bit of softness coming from the ProRes @ 1080p 25.


Thanks

Mike Marriage July 29th, 2010 02:13 PM

"The camera recorded directly to SXS cards whilst simultaneously outputting HD SDI through a Matrox MXO2 in a Mac Pro where the ProRes was encoded. This allowed for comparison of identical frames from the video. I tested both 1080i25 and 1080p25. Shutter was 1/50."

:)

Simon Denny July 29th, 2010 02:17 PM

Thanks Mike,

I thought there was an extra step. I initially thought you were re-compresiing the XDCAM footage into ProRes via FCP or compressor.

Cheers

Mike Marriage July 29th, 2010 02:37 PM

No, just one stage of compression on each, I'll add a clarification because I think everyone associates ProRes with being just an intermediate codec. Of course it can be a original recording codec and now even on board on the new Arri Alexa.

I would have used uncompressed as the benchmark but my hard drives can't cope with it.

Alister Chapman July 29th, 2010 04:05 PM

You'll see some of the biggest differences between 35Mb/s EX and ProRes HQ or 100Mb/s NanoFlash on things like fine patterns such as brickwork or gentle gradients like blue sky where where the 35Mb/s will be a touch blocky compared to the others. The larger the viewing screen the more obvious the difference of course. The differences are not always large, but they are there and they can get worse over multiple generations even if using a high quality intermediate. As well as the codec camera noise is just as important as a low noise camera will stress any codec a lot less than a noisy one.


All times are GMT -6. The time now is 01:26 PM.

DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network