Any words of wisdom re the above..

Thanks

Yes,
This has been discussed at length in a few threads on this forum Robin. So you can get a quick answer to this and any other queries you have - and you should ask, try the search utility and type in the key words.

You will have pretty much all the answers you want in an instant.

Cheers

Dear Robin,

Here are some guidelines:

At lower bit rates, Long-GOP has a definite advantage in quality, as it is approximately 2 to 2.5 times more efficient.

Putting it differently, at 100 Mbps, Long-GOP is much better than I-Frame.

50 Mbps Long-GOP is much better than 100 Mbps I-Frame.

We currently do not offer any I-Frame Only bit-rates less than 100 Mbps due to quality concerns.

Sony XDCam EX (35 Mbps 4:2:0) has proven how nice 35 Mbps Long-GOP can be, as it has proven how great XDCam 50 Mbps 4:2:2 Long-GOP can be.

Our 100 Mbps 4:2:2 Long-GOP is still the sweet spot and is far superior to DVCPro HD 100 especially in 1080 modes.


From a practical standpoint, for the highest bit rates:

280 Mbps I-Frame is just ever so slighly better than 180 Mbps Long-GOP

I do not have enough data, at this moment, to comment on 180 Mbps Long-GOP versus 180 Mbps I-Frame.


Our Long-GOP's ease of editing must not be confused with the ease of editing (or lack thereof) of editing HDV.

Ours is always full raster (for 1080 this means full 1920 x 1080) and this means that our Long-GOP is much easier to edit than 1440 x 1080 Long-GOP, such as HDV.

Our Long-GOP's ease of editing must not be confused with the ease of editing (or lack thereof) of editing HDV.
Ours is always full raster (for 1080 this means full 1920 x 1080) and this means that our Long-GOP is much easier to edit than 1440 x 1080 Long-GOP, such as HDV.

Just to add to Dan's post... Of course the Nano does do HDV (25 Mbps, 1440x1080, 4:2:0) if you choose to use that low data rate for a long run time...but most people would use the Nano for a gain in image quality, which means you're going to be in full raster 4:2:2.

Even HDV has gotten easier to edit in the last couple years. MPEG as a codec was developed to be used on both the encode and decode side with proprietary hardware...dedicated to simply working with MPEG. A computer CPU was comparatively clumsy at executing the necessary instructions to handle MPEG quickly and efficiently. Frankly, MPEG probably still isn't suited wonderfully for a computer CPU, it's just that the brute force available in the CPUs has increased so heavily that it simply out-muscles it. Digging a hole with a hammer is incredibly inefficient, but if you need a relatively small hole and you have a massive hammer...you can still get the job done.

Long GOP material will continue to be more taxing to handle because the frames have to be encoded and decoded out of linear order. Certainly the edit systems that tried to take HDV's non-square pixels at 1440 and interpolate them on the fly on playback had their hands full with that operation alone, but just as FCP decided to simply run DVCPro at 960x540 and edit it that way rather than waste CPU cycles on scaling up and down...HDV is typically handled pixel-native on most NLEs these days.

Long GOP remains, as Dan mentions, far more efficient than I-frame codecs. It's why 35 Mbps XDcam and 100 Mbps AVC Intra (an I-frame codec) can run in the same circles in every sense other than bit depth (MPEG in the common applications we all use including XDcam and the Nano are limited to 8 bit while AVC Intra can go to 10 bit color precision...however further pressuring the compression scheme) provided the cameras used on the front end are relatively even matched.

I ask because I got a request to shoot 1080i 50i at 100 Mbps i frame.Iam shooting on my HDX900.Ive only used long GOP before.

I wont be doing the editing..

Your camera only records I frame to tape...DVCProHD.

Are you saying you typically only record to the Nano?

Only done one job with the Nano... recorded tape and nano.. but end of last day.. on not so critical stuff.. we just recorded straight to nano..

program is finished .. without ever using the tapes.. but was good peace of mind..

I wonder also.. are there people only recording to nano.. or is there not that confidence level yet?

Dear Robin,

It is your choice.

Even in a perfect world, with a 100% reliable nanoFlash, having a backup is almost always desirable.

Just one thing I would add re 8 bit v 10 bit. You will only get an advantage out of a 10 bit codec over 8 bit if the cameras noise level is significantly lower than the codecs sample size. In practice this means you need a really quiet (noise free) camera to make the extra data generated by a 10 bit codec useful. Most of the current HD cameras are not clean enough to bring any significant advantage by recording 10 bit.

I wonder also.. are there people only recording to nano.. or is there not that confidence level yet?

Record to the nano and whatever format your camera uses normally (dual recording). However, use the nano recordings as your primary footage if you have set it to record at a better rate than the camera's recording format.

I ask because I got a request to shoot 1080i 50i at 100 Mbps i frame.Iam shooting on my HDX900.Ive only used long GOP before.

I wont be doing the editing..

Why don't you do some sample test recordings in that way and post your observations? Maybe test between I & LGOP and then show the producer the results.

Its the production companies standard settings for Nano use.. but thats usually with an EX3
I,ll be shooting with my HDX900.. not a big deal.. just wondered why some people like long GOP and some want i frame..

Dear Robin,

It is your choice.

Even in a perfect world, with a 100% reliable nanoFlash, having a backup is almost always desirable.

Hi Dan

Yes not casting aspersions.. just wondering if there are lots of people recording just to the nano

When I shoot 3D I just use the Nano's as they start and stop in sync via the cable that CD made for me.

For me it's comes down to the fact that I can shoot at 100Mb/s long GoP and get quality that is as good as if not better than 220Mb/s I frame, but my media lasts twice as long, long term storage takes less space and copying files takes half as long. In the edit I hardly notice any performance difference. So there are genuine cost and time savings by using long GoP without any image quality implications.

I would always record to the internal recorder even with a Nano hooked up. You could much more easily have someone accidentally pull out the wires, especially the consumer grade HDMI connectors from the Nano if you aren't careful. Now lots of times you will have the Nano mounted securely, but because it is outboard, especially in the field I would always run the backup.

Remember that once you verify the integrity of the recorded Nano files, you can always delete the backup recordings, although I keep them anyway.

robin,
I record through the camera and also through the NanoFlash. I primarily use the NanoFlash as my primary footage. There was only one time when I needed backup and that was when I hadn't seeded my SDHC card properly in the EX3 ... the Nano save the day.

When I shoot 3D I just use the Nano's as they start and stop in sync via the cable that CD made for me.

For me it's comes down to the fact that I can shoot at 100Mb/s long GoP and get quality that is as good as if not better than 220Mb/s I frame, but my media lasts twice as long, long term storage takes less space and copying files takes half as long. In the edit I hardly notice any performance difference. So there are genuine cost and time savings by using long GoP without any image quality implications.


So you get less time on your CF card recording iframe? or are you referring to 100Mbps as opposed to 220Mbps..

Thanks

Dear Robin,

I-Frame Only is far less efficient than Long-GOP (up to a certain bit rate).

To be clear, at 100 Mbps, Long-GOP is far better in terms of quality than I-Frame Only.

Please remember that we are impartial here, we support both options in one device.

There is a very nice, quantifiable, increase in quality when going from lower bit-rate Long GOP (say 35 Mbps 4:2:0) to higher bit rates (say 100 Mbps 4:2:2). Of course some of the increase in quality is due to going to 4:2:2.

But as one goes from 100 Mbps to 140 Mbps to 180 Mbps the quality increase is less dramatic with each step. This is an example of diminishing returns.

Stating the obvious: As the Bit-Rate goes up the amount of time one can record on a CompactFlash card of a given size goes down.


Our 100 Mbps, 140 Mbps and 180 Mbps Long GOP 4:2:2 recordings are very high in quality.

Technically, a very slight increase in quality can be obtained by recording 280 Mbps I-Frame Only as opposed to 180 Mbps Long-GOP. This very slight increase in quality can be detected with the right instruments but would be hard to detect with one's eyes, or in Post.


With this background, if you are considering 100 Mbps, always choose Long-GOP, it will be much higher in quality.

If you have to deliver I-Frame Only (Intraframe), then choose a higher quality, higher bit-rate option. Choose at least 140 Mbps or 180 Mbps; 220 Mbps or 280 Mbps would be better, but of course your recording time per card will be reduced.

To be specific, Alister stated that 100 Mbps Long GOP is as good or better than 220 Mbps I-Frame Only. I agree.

And he gets approximately 2 times as much recording time with 100 Mbps Long-GOP.

Note: I edited an obviously incorrect statement above as pointed out by Adam.

Thanks for the info.. I,ll question the prod co about why they want iframe at 100Mbps

To be specific, Alister stated that 100 Mbps Long GOP is as good or better than 220 Mbps I-Frame Only. I agree.



Hi Dan,

The only caveat I'd add here is that when many (Final Cut) users think about 220 Mbps, they're thinking about ProRes. 220 Mbps ProRes is a different codec than 220 Mbps MPEG2. 220 Mbps ProRes has 10 bit color precision, whereas MPEG2 on the Nano is 8 bit at all bitrates...and ProRes is variable bitrate compared to the Nano's constant bitrate (at whatever setting)...

The basic idea here that Long GOP is better (more quality for the bitrate) than I-frame is one that I agree with...no question about it. As long as all other factors are equal.

I just think that our industry has many cases of over-generalization with the intent of trying to simplify concepts for users or customers, and while it probably adds confusion for some, I think attempts to present some of the shades of gray can be more enlightening for users in the long run.

Bitrate in general, I-frame vs Long GOP, 8 bit vs 10 bit vs ?, 4:4:4 vs 4:2:2 vs 4:2:0, DCT vs Wavelet, Constant bitrate (CBR) vs Variable bitrate (VBR), Log vs Linear, RGB vs Y'CbCr... They all combine to create different cocktails of image processing.

Then, there is also the question of 'runnability' in editing...

I'm not trying to start an argument so much as wishing to make sure that the idea of multiple factors contributing to a codec's image quality is out there.

(Disclaimer: I'm a Nano user...and a KiPro user. They each have their sweet spot. I also think that CineForm is an exceptional codec... I've also acquired to CineForm. If you would ask if I had a horse in this race, I guess I'd have to say "all of them.")

Dear Tim,

I agree with all of your comments.

220 Mbps ProRes is an entirely different codec.

One nice thing about the nanoFlash, and in the spirit of this discussion, as long as we are discussing 4:2:2 nanoFlash footage (all of our 50 Mbps and above options are all 4:2:2), then all things are equal, except for the items that we are discussing, either the bit-rate or I-Frame versus Long-GOP.

And your point is valid.

So, when I said 220 Mbps some could have incorrectly inferred Apple ProRes 220 Mbps.

Also, when we discuss 100 MBps I-Frame Only (Interframe), some could incorrectly infer DVCPro HD (which is also 100 Mbps (in some modes), but an entirely different codec.)


You are in an elite group in using both a Ki Pro and a nanoFlash.

"Horses for Courses" comes to mind here.

Our strengths are offerring a very versatile product that produces very low-noise footage.

And many like our compact size and low power requirements.

Stating the obvious: As the Bit-Rate goes up, the size of the files goes down

You may want to rethink that statement, Dan. :)

Dear Tim,

You are in an elite group in using both a Ki Pro and a nanoFlash.

"Horses for Courses" comes to mind here.

Our strengths are offerring a very versatile product that produces very low-noise footage.

And many like our compact size and low power requirements.


I like the Nano for exactly those reasons, and have been quite happy with the footage it produces.

Since I (as many of us) do this for a check, I use whatever the client requests, and the KiPro is a very popular device as well...

"Horses for Courses" (anybody know where the heck that saying came from?) is right.

Dear Adam,

Yes, my mistake. I need to take a little longer in proofreading my posts.

Thank you for pointing it out.

I have corrected the original post.

You're quite welcome, Dan.

I,ll be shooting with my HDX900.. not a big deal.. just wondered why some people like long GOP and some want i frame......Hi Robin: The reason I like I-Frame recording over Long GOP for very high end (Read Digital Cinema Origination and *Not* television), is two fold.

1. I can obtain very high image quality @ 280 Mbps in i-Frame recording with my XDR, which I consider to be slightly superior to Long GOP 180 Mbps.

2. Recording in I-Frame produces video files which are pretty much universally accepted by most major NLE post systems, such as Avid Media Composer and Final Cut Studio 7.x

* My standard approach to paying TV job shoots is to shoot in Long GOP MXF 50 Mbps setting. This produces Avid compatible Sony XDCAM HD 4:2:2 Full Raster video files equal to what the TV networks record on disc or tape. (Sony HDCAM SR is the only other exception).
** May I suggest as a basic rule of thumb, that if you're shooting for Network Television, and you don't know what their specs are, then always assume Long GOP 50 Mbps in MXF standard and you shall not go wrong ! Assume Avid Media Composer in their post and not Final Cut Pro (Including Discovery Channel).

Hi Mark

Thanks for the advise.. I,ll remember it.

In this case they sent the settings(for EX3 use) I,ll be shooting HDX900.. they want QT so I guess FCP.. like you say, before I have shot Long GOP 50Mbps.. but it seems they want 100 Mbps QT iframe.. anyway have emailed to confirm iframe.. just now seems strange as everyone here says that at 100Mbps long GOP is a far better image..and fitting any NLE system doesnt seem to be a problem as they state QT.. ,

Thanks again all

Hi Mark

Thanks for the advise.. I,ll remember it... just now seems strange as everyone here says that at 100Mbps long GOP is a far better image..and fitting any NLE system doesnt seem to be a problem as they state QT.. ,

Thanks again all.....If they say I-Frame 100 Mbps, then they're probably posting on Avid Media Composer. Please be advised anything in Long GOP MXF which is over 50 Mbps *Will NOT work in Avid !*

Robin:

The HDX-900 has a Video Bit Rate of 100mbs with 4:2:2 Sampling and Intra-Frame Compression. I wouldn't see a reason to duplicate this on a CF card unless a tape deck is not available for transfer or you just want to record at the same time on the NanoFlash in order to save footage for your reel. Once you pass off that tape to a client it's usually gone. If it's for my reel then I would record at 100mbs/Long GOP as many have suggested on this forum. You can't go wrong with this setup.

With that said, if my client wanted footage on a CF card while one is shooting with the HDX-900 then I would still roll tape at all times especially if one is not familiar with the process of transfer and backup of digital media. Why would one not want to roll tape at the same time if it's available? It's one sure way to be responsible in this business and save a major problem from developing.

I don't question what my clients want as far as their record formats. They get locked into a particular standard and don't want to venture in other directions. I'll let them deal with that. I feel my responsiblity is to get the best images possible for the editor. I try to keep in the back of my mind, "it's not the format, but it's what you do with the format" attitude. I will offer suggestions and better ways of doing things from time to time if I think it will help the process. Introducing the Nano to a client can be a tough sale especially if one is still using tape. I'd rather not have a client turned off by a bad experience with the Nanoflash from the beginning.

Yeah I,ll record to tape for sure.. so do you charge for the tapes? if they dont want tapes in the first place? They wanted P2.. but I asked if I could use HDX900 with nano.. luckily seems they are ok with the nano as they have their own set details for it.. (for use with a EX3)
Yes they dont have a deck for the Pana tapes,so didnt want to just shoot on the HDX900.Fair enough..

What is your procedure for backing up CF cards..? besides the tape backup..

Thanks for the advise..

Robin:

I always charge for the tapes. It's a cheap cost for backup when using the HDX-900 and NanoFlash. I use Shotput Pro to dump the CF cards to hard drives. If I'm in the field and I'm shooting more footage than I have CF cards for then I dump to the Nexto as I go along.

Ok thanks for the info

Hi there,

I've heard that some boffins have figured out that 75Mbps is the 'sweet spot' for Long GOP MPGEG2 HD, at which point it becomes indistinguishable through the Terrestrial HD broadcast chain against 100Mbps AVC-I for example.

Perhaps the new 80Mbps datarate in the next firmware rev might be the perfect 'step-up' for those concerned about heavy grading on 50Mbps material.
As I understand it, the NLE's will still see this as 50, but the datarate will in fact be 80.

On redundant recording, it depends - for second camera stuff we shoot only NanoFlash and haven't lost anything yet.
For news and primary camera greenscreen shoots we shoot redundant tape/Nano as a safety measure for 'unrepeatable' events.

Dave

Ok thanks Dave..

Hope Spring lets to Glasgow soon!

Robin:

I always charge for the tapes. It's a cheap cost for backup when using the HDX-900 and NanoFlash. I use Shotput Pro to dump the CF cards to hard drives. If I'm in the field and I'm shooting more footage than I have CF cards for then I dump to the Nexto as I go along.


How do you get your files out of the Nexto.. USB?

Thanks

Dear Robin,

The Nexto's support connecting to the your computer via USB and eSATA.

Hi Dan

Ok thanks.. what is eSATA

While I have you!

How must the camera be set for time lapse..? if I set HDX900 to TL.. will this automatically go to Nano via SDI.. if I set the nano for TL.. what about recording to tape same time?

Thanks

Dear Robin,

eSATA is "External Serial ATA", a very fast computer interface, one of the fastest currently available.

SATA is typically used as a internal, computer to disk drive interface. eSATA is used to connect to external devices.

If you setup your camera for timelapse, then you record on tape in timelapse mode.
But, HD-SDI will be non-timelapse mode.

If you set the nanoFlash to record in timelapse mode, then this has no effect in how the camera operates or records.

You can record timelapse in the nanoFlash, while recording normally in your camera.

Hi Dan

Ok great thanks.. all good info..

Remove pull down... is this restoring footage that has pull down applied to it... as the wording implies..?? eh 23.98p back to true 24P.. or is it applying pull down..???

My HDX900 although set to 24p will give 23.98... as with 30p.. will give 29.98 .. in 59.94 system mode..

Would I need to tick remove pull down shooting 24p or 30p.. ?

Thanks again

You can record timelapse in the nanoFlash, while recording normally in your camera.

Or vice versa - is that right?

Dear Piotr,

Yes, either device can record normal or timelapse while the other records normal or timelapse.

Hi Dan

Ok great thanks.. all good info..

Remove pull down... is this restoring footage that has pull down applied to it... as the wording implies..?? eh 23.98p back to true 24P.. or is it applying pull down..???

My HDX900 although set to 24p will give 23.98... as with 30p.. will give 29.98 .. in 59.94 system mode..

Would I need to tick remove pull down shooting 24p or 30p.. ?

Thanks again


Dear Robin,

On most cameras, when recording 24p, or 23.98, or 23.96, extra, duplicated frames are added to create 59.94 or 60 PSF (Progressive Segmented Frames). This process is known as adding Pulldown.

When you check Pulldown Removal, we remove these duplicated frames prior to recording.

As far as I know, you would only check the Pulldown Removal box for 24p, 23.98, or 23.976 footage.

30p or 29.97p will not have duplicated frames, thus you do not need to check Pulldown Removal.

I welcome comments and corrections in case I have any of this wrong, or if anyone can explain this better.