So what cameras under $25,000 actually put out a TRUE 10-bit signal for recording to the upcoming Gemini recorder?

No SMPTE 10-bit signal with an empty 2-bits like the XLH1 effectively making it 8-bit.

Sony PMW-F3 does right out of the box with no firmware update. You only need the update if you want to run S-Log 10bit.

Depends if you need an ENG cam such as the Sony PMW-350 or S35 camera.

Really, the F3 + S-Log and its 10bit 444 RGB output and Gemini for $25,000 (including SSDs & SxS) is an amazing combo which competes with far higher priced cameras including Red. You get 14 stops of Dynamic Range, 10bit and 444 in a very easy to use system. Part of the reason so many people are flocking to the F3 rather than Red (or even from Red) is the simpler workflow.

With the release of Gemini, raid card manufacturers should see an increase in sales; so, people can keep up with the high data rate.

To C-D guys, if you want any assistance/suggestions/advice on raid for your customers, I would be glad to help in any way. I have a 3ware card in my home workstation, an Areca 1680ix in my HP Z800 and another 3ware in my playout server. I have 14 drives in my home workstation, 12 in my Z800 and 5 in the server. Editing uncompressed video entails a lot more than just hooking up a bunch of drives in Raid 0 if you want stability, reliability and a smooth editing experience.

Don't forget the EX1/3 put out a true 10bit signal as well.

....Editing uncompressed video entails a lot more than just hooking up a bunch of drives in Raid 0 if you want stability, reliability and a smooth editing experience.
Being able to online to Prores LT, I wouldn't spend much money in super-fast storage.
rafael

So what cameras under $25,000 actually put out a TRUE 10-bit signal for recording to the upcoming Gemini recorder?

No SMPTE 10-bit signal with an empty 2-bits like the XLH1 effectively making it 8-bit.

You can get Pana 3100 for $19,500.. fantastic camera for the price.. for doco/live action work where you need a zoom.. or drama..

Dear Steve,

Thank you for your offer.

We welcome your suggestions.

I think all SONY XDCAM and all PANA AVC-I put out true 10b Unc.
Modern cameras process at 12 and 14 bits.
8 or 10b out is a manufacturers decision.
rafael

The problem with some lower end cameras like the EX1/3 is that they are so noisy that 10bit adds nothing.

Steve, noise has nothing to do with bit depth.
Noisy or not is not the same having 256 shades grades than 1024.

When camera are not properly set, the best camera can be noisy.
My EX-1 is not noisy at all and the 10b shines when recording Prores with the ioHD.
rafael

Rafael,

I am sorry to inform you but the EX1 & EX3 are very noisy cameras no matter how the PP is set. I have done extensive testing with my EX1 & EX3 and only after I process 280Mb nanoFlash files with Neat Video and export to Cineform 444, I can start to actually push the colors in After Effects without the noise destroying the image. I have some experience with DaVinci Resolve and I can duplicate what Resolve does within AE, albeit at a much slower pace.

Also, noise and sensitivity determine the actual bit depth. Dan Keaton explained that for every 1bit, you need a clean 6db of sensitivity. With the EX1's measured sensitivity of 48db, there is only 8bits of info available.

The easiest way to see the noise is to shoot a greenscreen and then key out the green with a white background. You will then see the large blocky noise.

Steve, you know very well I'm with you on how noisy the EX1/3 cameras are...

That said, I still believe it'd be nice to capture all 10 bits on the nanoFlash (as we're on the CD forum now, right?). I also am using NeatVideo extensively, and very often - with most noise removed - the resultant images start to show a very awful color banding, I believe that should NeatVideo be run on 10 bit 422 files, the result would be even better!

Before de-noising, those extra 2 bits of (mostly) noise help disguise banding a bit as the image is dithered from 10 to 8 bits.

Also, my little theory is that the noise we see on those large, uniform-color, mid-bright areas is partly due to those 2 bits of color information missing. The fluctuations of current from the photosites translate to very slight fluctuations of the color information; did we have all the 1024 shades of it (rather than just 256), those shimmering stains/blots would not be as striking as they are now...

Piotr

Piotr,

Very good point about the noise removal. I haven't run into it yet probably because I 'pre-process' certain clips by using Neat Video in Premiere and exporting to Cineform 444, which I then grade in After Effects.

I got the idea for this pre-processing from someone in the Cineform forum who does something similar with 5D footage. He ran into the same banding problem so he added a very small amount of film grain.

Btw, I did a little test with NV and Cineform where I exported the same clip twice: one with NV and one without. The clip WITHOUT NV was over 3 times as large due to all the noise. The source clip was from an EX3 and 280Mb nanoFlash.

I seriously would love someone to make a program that automates the conversion of clips to Cineform with Neat Video noise removal included. Because I use the multiple XDCAM cameras, I can create a noise profile to remove enough noise to make a considerable difference.

And yes, I also would like to record to 10bit but I will never use a Pro Res recorder due to its limited quality (Pro Res HQ at 1080 23.98p has a max data rate of 176Mb/s and its VBR so you don't usually get the full 176Mb/s). Give me Cineform or Uncompressed for the same price and I would be happy :)

Also, noise and sensitivity determine the actual bit depth. Dan Keaton explained that for every 1bit, you need a clean 6db of sensitivity. With the EX1's measured sensitivity of 48db, there is only 8bits of info available.

No sense.
Bit depth is a format setting.
Is the number of bits you use to sample an absolute value (luma, chroma, sound,..)
Has nothing to do with noise, sensitivity or whatever.


Before de-noising, those extra 2 bits of (mostly) moise help disguise banding a bit as the image is dithered from 10 to 8 bits.

Also, my little theory is that the noise we see on those large, uniform-color, mid-bright areas is partly due to those 2 bits of color information missing. The fluctuations of current from the photosites translate to very slight fluctuations of the color information; did we have all the 1024 shades of it (rather than just 256), those shimmering "stains" would be as striking as they are now...

If you are testing everything with NANO footage (8b), how you expect to have 1024 ranges of shades?
Have you shoot anything in 10b with your camera?
Do it and compare pictures.

BTW, I'm using NeatVideo since a couple of years. I use it in not only for the EX stuff, but for everything. I stated in this forum long ago the need to apply this filter on the raw footage, before any other processing.

This counts for every kind of footage. Color grading, masks, and any other filter will work better in a noiseless picture than on a noisy one. Once you have processed the picture, the NeatVideo Noise Profiles won't work.

If you are testing everything with NANO footage (8b), how you expect to have 1024 ranges of shades?


Dear Rafael,

You didn't get me right - of course I realize I'm only getting 256 shades with both the native EX and nanoFlash recordings, hence I put this speculation with the "my little theory" disclaimer...

Just had a look at the NeatVideo noise reduction plugin. Astounding.

I use the NeatVideo plug-in religiously ... best when used on original material when first put on a timeline before you do anything else. It can be time-consuming but the results are very good.

Dean,

Could you walk us through how you use the neatvideo plug-in (especially if you are using the nanoflash footage)? I'm interested especially to know how many 'generations' you need to render, and whether this has an impact on the final quality at the end?

Dean,

Could you walk us through how you use the neatvideo plug-in (especially if you are using the nanoflash footage)? I'm interested especially to know how many 'generations' you need to render, and whether this has an impact on the final quality at the end?

If you download the plug-in ... I use it in FCP ... you can take a section of a frame within a section with uniform coloring (black for example) and have the plug-in analysis the noise in RGB. It will provide a % of noise present in the frame. If you are satisfied with that % reading say 40% to 50% then proceed to apply that percentage of clean-up in the sequence that was shot within that period. You can do this nested on a whole timeline if the sequence was shot in the same way during the same session and then render. It's really as simple as that. I encourage you to give it a try. There is a numeric value you can apply ... from 1 to 5 adaptive ... that reflects slight or major adjustment if you wish on the fly but I tend to get an analysis of sections and find that is more accurate.
ps ... if this is the first thing you do on material in a time line ... you don't need to adjust for noise again.

Dear Piotr,

The problem is that your "little theory" is built on top of misconceptions.
Sorry to say, but you are not very clear some basic concepts.

"Dan Keaton explained that for every 1bit, you need a clean 6db of sensitivity. With the EX1's measured sensitivity of 48db, there is only 8bits of info available".

Statements like that makes no sense at all. You are mixing there Analog concepts (SIGNAL to NOISE RATIO, BAND WITH) with digital elements (BIT DEPTH).

The "Signal to Noise Ratio" tells the quality of any system (video or audio), and of course will have impact in the picture quality, but is not exactly the "picture noise".

Again Piotr, whatever the camera you have, will ALWAYS be better to record that in 10b than in 8b.

Instead of trying complicated tests:

- Take the VIDEO COMPOSITE out of any cheap MiniDV camcorder.
- Set Color Bars out of the camcorder.

Record that at 8b Unc and at 10b Unc and compare.

Rafael

Thanks for that. So once you have established the correct amount of noise reduction you then render that sequence to the same format as the original (nanoflash mxf/mov codec), or is it changed to something different?

Any further colour correction would then be done on these new files?

It depends on your final output. I find prores produces the best output for the net and TV but you can play with this master. You color correct after you've done the noise cleanup. You can output from the master at H.264 if you want via compressor or Flash via Adobe Flash if you have that software. On my reel on exposure room ... I out put a commercial 'Andy's Fish' to flash for the net. It's very clean.
Andy's Fish By Dean Harrington On ExposureRoom (http://exposureroom.com/members/DeanHarrington/9f5fe50051b14e6eb0de330e1dd411f9/)

ps ... this was shot long GOP 100 ... you might prefer 50 but that's your choice.

Got it - thanks!

Dear Piotr,

"Dan Keaton explained that for every 1bit, you need a clean 6db of sensitivity. With the EX1's measured sensitivity of 48db, there is only 8bits of info available".

Statements like that makes no sense at all.

Dear Rafael,

I guess you're confusing things, and who said what. The above statement from Dan has NOT been a basis for my little theory (in fact, it doesn't convince me at all, either).

As a matter of fact, my feeling is that the 2 of us mean the same, just using different wording...

Piotr

Dear Friends,

Just to be clear, I do not believe that it was me who made that original statement.

Dear Dan & Piotr,
Sorry I didn't intended start a discussion about who said this or who said that.
I'm pointing to statements that needs to be proved and to and to wrong concepts.
We are mixing the "Camera noise" with the concept of Signal to Noise Ratio and the 60dbs implied in 10b sampling.


And statements like this "the EX1's measured sensitivity of 48db" needs to be explained.
Who have measured that and how?

About the 8b/10b-Camera Noise relation, and in short:
When you sample a Video signal with more or less precision (10b vs 8b) you don't only sample the picture with more or less precision: You will be also sampling the noise with more or less precision.

This won't changes the noise, BUT If you sample the noise with more precision, you can clean the noise with more precision.
That's what matters.
rafael

No sense.
Bit depth is a format setting.
Is the number of bits you use to sample an absolute value (luma, chroma, sound,..)
Has nothing to do with noise, sensitivity or whatever.
No, fundamentally Steve is right. Bit depth does indeed define the number of variations in the shade of an image - 256 for 8 bit, 1024 for 10 bit - but if the noise is greater than the differences in level represented by the least significant bit, resolving to the 10 bit level becomes a bit pointless. And that's the case for all but the most expensive cameras.

If you have a set of scales accurate to + or - 2 grams, what would be the point of having a set of weights that would allow you to measure to 0.1 of a gram?

If a perfect mid grey would resolve to a level of 128 in an 8 bit system, then noise may mean in the non perfect world it will be between 125 and 131. Go to 10 bit, and the same scene will give values between 500 and 524. All the 10 bit recording is really doing is resolving the noise more accurately. The fine gradations in the image tone become masked by the noise.

And 10 bit will mean a 25% increase in data rate to keep all else equal. That data may be better used in reducing overall compression, rather than increasing bit-depth. Not just bit depth, but also compression can give banding, especially if less data is allocated to chrominance wrt luminance. (Try forming a gradient in Photoshop and making a JPEG on maximum compression if you don't believe me.)

Just as a fun bit of history, the Quantel Domino system scanned 35mm motion picture film and saved oversampled images at approx 3K, and those files were 8 bit.

If you think that's severe, at the time, Quantel claimed that they could actually reproduce the picture perfectly in 6 bits, and the last two bits were just for grain/noise!

Kodak took a different, 10bit log approach with Cineon, but many, many feature films have shots and sequences run through totally in the 8bit domain via Domino. Movies like Stargate, First Night, Lost in Space, etc.

Regards,

Hi Dan,

I tried to find the post about 1bit per 6db of S/N but I haven't had any luck. I honestly thought that it was you who posted this. If it was you, can you provide an explanation again about 1-bit and 6db?

I do know that many others including Alister have said that recording 10bit from a noisy camera does not add much quality if any, and I must defer to him as he is far more knowledgeable due to his engineering background.

Thanks

Dear Friends,

I take no offense whatsoever about the 1 bit quote.

I do not know who originally said this and while it makes sense to me, I do not know if it is technically accurate or not. My best guess is that is probably accurate, I have just not studied this concept in detail.

I do agree completely with Alister and others that some camera's that technically have a 10-Bit HD-SDI output do not have a low enough noise floor to make the 10-Bit truely useful.

On the plus side, the Sony PMW-F3 is one of the great cameras, especially in its price range, as it is an extremely low noise camera. Thus the 10-Bit's of the F3 are very useful.

The funny part to me about all of this is that since the F3 is such a low noise camera the 8-Bit nanoFlash works extremely well with this camera.

This is a direct side benefit of it being such a low noise camera. The nanoFlash does not have to waste any of its compression power on compressing the noise, thus it can produce just stunning images.

One of the reasons that the nanoFlash shines in this application is that the Sony XDCam 422 codec that we use is an especially low noise codec.

When our Gemini 4:4:4 ships, it will be ideally suited for the F3.

The Sony F3 + S-Log Firmware Upgrade + Gemini 4:4:4 is going to be just magical. Especially when one considers the price of this combination.

What are the opinions/facts on the Sony FS-100? does it have a true usable 10bit output that would make sense with the Gemini?

(Unfortunately the FS-100 does not have HD-SDI and relies on HDMI as the output - grrr)

Unfortunately (from the viewpoint of Gemini potential buyers), the FS100 only outputs 8 bits via HDMI. So, investing in the Gemini for the FS100 would be an overkill, IMHO. It will make a perfect companion for the F3, though (I wish I could afford them both...).

Fortunately for the current nanoFlash users (myself included), after implementing TC through HDMI, our little marvels will serve this camera perfectly :)

Piotr

David,
I mostly agree with you.
Simplifying, you lose "1bit". Is like you end up working in 9b.
512 levels instead of 1024.
BUT that's still much better than 256.
Better picture sampling, and as you points too, easier to clean the noise.

What I do not agree with you when you say:
"If you have a set of scales accurate to + or - 2 grams, what would be the point of having a set of weights that would allow you to measure to 0.1 of a gram?"
Hey, those cameras processes at 12/14 bits before going 10b.
Our original scale is not set in grams or even decigrams, is milligrams.


Dan wrote:
"I do agree completely with Alister and others that some camera's that technically have a 10-Bit HD-SDI output do not have a low enough noise floor to make the 10-Bit truly useful".
Dear, Dan:
You don't get the full advantage of the 10b, but you go way much better than at 8b.


Hi Dan,

I tried to find the post about 1bit per 6db of S/N but I haven't had any luck. I honestly thought that it was you who posted this. If it was you, can you provide an explanation again about 1-bit and 6db?

Thanks
Steve: that's not wrong. I more bit means duplicating the number of sampling levels, so duplicating the sensibility. That means 6dbs of dinamic range.

About the "EX-1 proved 48db sensitivity", up to the EX-1 brochure "the CAMCORDER" (In SxS recording) has a S/N of 54db.
The SDI signal doesn't fallow the same process.
We still talking about the EX-noise and the 10b issue without having yet shot any test at 10b Unc from that SDI.
We are mixing the concept of system noise with camera noise (that can be generated for a wrong setting), and compression noise and .
rafael

What I do not agree with you when you say:
"If you have a set of scales accurate to + or - 2 grams, what would be the point of having a set of weights that would allow you to measure to 0.1 of a gram?"
Hey, those cameras processes at 12/14 bits before going 10b.
Our original scale is not set in grams or even decigrams, is milligrams.
Ah, but the original scale is for the raw output from the chip, and before ANY of the necessary camera processing - knee compression, gamma compensation, gain, anything. That will inevitably scale many of the quantisation steps up a great deal, so if (say) a 12 bit system, a single bit difference pre-processing may equate to many bits post-processing. (And in other parts of the waveform many bits will be compressed into one.)

That's the whole reason for the necessity of the front end needing to be at a far better bit depth than the output.

To carry on the weight analogy, it's a bit like a "system" whereby you measure in hundreths of a gram, then give an "output" of 100x that weight. So if a "customer" comes to you with .04g, you give out 4grams. You may need milligram accuracy for the input - you only need gram accuracy for the output.

Maybe a better analogy would be if you are making something from drawings, and the drawings are to 1:10 scale. If you need an accuracy of at least a centimetre in the finished product, then obviously the drawings should be accurate to the nearest millimetre.

The 1 bit=6dB is true in principle, but in real life there are all sorts of complications. The idea of a camera having a defined s/n ratio is great - but in practice the figure will vary depending on whether it's measured in the highlights, lowlights, or mid-greys. I suspect manufacturers will want to publish the figures that look the best, and gloss over on what part of the curve the measurements are made. That's why I'd rather stick to statements like "if the noise is greater than the differences in level represented by the least significant bit, resolving to the 10 bit level becomes a bit pointless" and not get deeply into figures.

The science is that the human eye can only resolve something like 100 different shades of grey at a single time - and they can be *just about* represented within a 7 bit binary number. I have a book (not online - sorry) which shows a 6 bit greyscale with obvious banding, then the same greyscale with noise added, and the banding is totally masked. Practically, 8 bits becomes desirable - gives a little leeway - for straightforward viewing. The depth resolution is better than the human eye. Processing is a different matter, and 10 bit processing may be a very good thing even if the original acquisition has been at 8 bit - 10 bit acquisition only really becoming worthwhile if the camera noise level is low, very low.

David,
I agree with all your points. Is pure maths.
What I do not agree is with your starting point:

"Ah, but the original scale is for the raw output from the chip, and before ANY of the necessary camera processing - knee compression, gamma compensation, gain, anything".

When a camera claims to process in 12 or 14 bits, that counts for all the process from digitizing the RGB on the CCD/CMOS all the way (RGB matrix, knee, detail,..) to a 12/14b Y'C'bC'r/4.4.4.
Then the signal just need to be down-sampled (in bit depth and Chroma) to 10b YUV 422 for SDI out for to whatever codec the camera may uses.

See what the AF-100 brochure claims:
"The AG-AF100 series is equipped with a high-performance 18-bit digital signal processor (DSP) for image processing. Optimized for HD video recording, the DSP handles various image rendering processes, such as dynamic range stretch (DRS), gamma, 12-axis independent color correction* and detail enhancement, as well as conversion to HD/SD video formats - all with exceptional precision and high image quality.

A camera claiming "12/14b digital processing" and that 12/14b stoping at the raw output of the chip,that would be a scam.
I haven't read no where the supposed bit depth processing of the EX-1/3.
rafael

OK, to try to make it clearer, yes, it's true that the entire early processing chain will be 12/14 bit - but that's not to say that the DATA within that will have that sort of resolution. And the processing will affect which of the range is relevant, which is not. So you may start with a nominal 12 bit number (say 000000101010) multiply it by (say) 16, and you get 001010100000. Now whatever number you start with, after a multiplication by 16 the last four significant bits will always be zero, and hence can be discarded. The analogy is that after "processing" all the information will be in the most significant 8 bits and it's pointless to keep the others.

The above is obviously highly simplistic - processing involves far more than multiplication by 16 :-) - but the principle does hold true. You DO need a high bit depth for initial processing, to ensure the calculations don't get rounded or overflow, but the likelihood is that at output the least significant bits won't be meaningful and it's sensible to just lose them. (Which is true even if 10 bit final output.)

The question is only where to draw the line. And one of the main factors in that will be the noise level of the camera after processing. Which brings us back to where this started - a noisier camera can have the line drawn at keeping a higher bit depth than one with an inherently lower s/n ratio. In practical terms, the most 8 significant bits are all that is relevant for the majority of cameras, there may be something to be gained for 10 bits for the most expensive.

Hi David,

Your maths are perfect, but for a completely different scenery.
You are talking about "multiplying x 16"; you are talking about about AMPLIFYING.

In the camera video signal processing there is not any AMPLIFYING involved.
Well, we have the "Gain" for desperate situations.
Out of the Gain, the multipliers you will find in a camera processor are values between "0" and "1".
The product of any multiplication can't ever be bigger than the multipliers, and can't ever be bigger than "1", so in fact what you are always doing is DIVIDING, reducing values.

- After the raw RGB is Gamma corrected (R'G'B'), Luma is calculates with this formula (709):
Y'= 0.2126 R' + 0.7152 G' + 0.0722 B').

- When you matrix R'G'B' to get C'b and C'r you go to similar calculations.

No one of those operations will generate "0's" at the right of your values.

YOU NEVER GET A BUNCH OF 000000 in the LESS IMPORTANT BITS.
All the range is full of detail.
When reducing bit-depth you won't ever be trashing a bunch of empty bits.

You also you forget the substance of what we are talking about.
You start from "OK, to try to make it clearer, yes, it's true that the entire early processing chain will be 12/14 bit - but that's not to say that the DATA within that will have that sort of resolution"
Well if you start with something as poor as"xxxxxxxxxx00", you are right, but that's not what you should be getting from a CMOS/CCD claiming 12/14 bits.
A camera just sample electric charge that may range all the way from "0 to 1,099 Volts".
Here things doesn't works in "steps" yet. Here we have infinite intermediate values to represent.
The more bit depth you use to sample that charge, the more definition will have. Noise included.
rafael

You are talking about "multiplying x 16"; you are talking about about AMPLIFYING.

In the camera video signal processing there is not any AMPLIFYING involved.
Yes there is, though it may not be obvious as such at first sight. The obvious examples are knee compression and gamma adjustment. Referenced to peak value, the former acts by division for the highest values and hence multiplication of the values corresponding to darker levels. Gamma correction is effectively multiplying the lowlight values by a far greater factor than highlights.
YOU NEVER GET A BUNCH OF 000000 in the LESS IMPORTANT BITS.
All the range is full of detail.
When reducing bit-depth you won't ever be trashing a bunch of empty bits.
No, practically I doubt you ever would get all 000's - I did say my analogy was "highly simplistic". In pactice, any manipulation is unlikely to be simple multiplication by a power of 2. But the principle remains true. You may need the high bit depth to be able to perform the calculations - after calculation the least significant bits will be random if not 0000. The fluctuations in them will ONLY represent noise and are not worth wasting bandwidth on preserving.
A camera just sample electric charge that may range all the way from "0 to 1,099 Volts".
Here things doesn't works in "steps" yet. Here we have infinite intermediate values to represent.
The more bit depth you use to sample that charge, the more definition will have. Noise included.
rafael
But just what is the point of more and more accurately defining the level any instantaneous value of noise is at? This is equivalent to trying to state a weight to a hundreth of a gram when your scales are only accurate to +/- a single gram.

Let me try one more example. Let's say a businessman wants to sell a product at the average price that 1,000 other stores charge, plus 1 penny. All the prices are in the range 15p to 31p - so can be defined with a two digit number. The first step is obviously to add all the 1,000 prices together, and let's say the result is 23,034 - obviously we need at least FIVE digit processing to start processing the data. Next, divide by 1,000, which may give 00,023.034 - which you may say means EIGHT digit processing, and finally add a penny to give 00,024.034 pence.

But do you think that is the price that will go on the ticket? Obviously not, and the practical answer will be "24 pence" - it makes sense to go back to a two digit number for output. The three most significant bits have served their purpose in processing, the last three likewise. An improved analogy would be to realise that his 1,000 competitors change their prices slightly day by day and the next day the average result may be "00.023.893 pence". The digits after the decimal point here represent noise, and it's pointless to worry about them.

So in the analogy a two digit number is quite adequate to define input and output, but you need five/eight digit numbers to process the calculation. And the principle for video is exactly the same. 12/14 digit processing does not necessarily mean 10 bit output will offer any advantage over 8 bit output - that will depend on the noise level.