Jump to content

My response to Alister Chapman's response to "all the FS5 complainers"!


Andrew Reid

Recommended Posts

  • Administrators

DSC02861.thumb.JPG.c9beba91666a707ca564c

Sony Independent Certified Expert Alister Chapman is a freelance videographer and runs workshops on occasion for users of Sony's pro video cameras.

You can check out his response to the FS5 problems widely reported by users. Its 's his blog here. Now you can read my post about why he's wrong!

Read the article here

Link to post
Share on other sites

Alister has been of great help to many sony users. 

I was surprised to read his post trying to defend the undefendable. 

As you say, if everyone pushes Sony on this one, the fs5 can only become better and all users benefit. 

Hell, even Canon is helping too :p 

http://nofilmschool.com/2016/01/canon-c100-mark-i-ii-price-drop-cost

Link to post
Share on other sites

As a huge Sony fan and someone that is about to buy the FS5 soon, I must say THANK YOU for getting this issue out into the spotlight. In order to get this fixed, it will take some degree of customer pressure to get Sony take this problem seriously and develop a fix. This sites and others are doing a great job at explaining the problem and making users aware of what is really happening to their footage.

As a shooter, there is nothing worse than thinking you are recording OK in the field only to come home and getting shocked at what you see afterwords.

Good job and thanks again!  

Link to post
Share on other sites

I recorded out of the HDMI port with the new bm video assist a couple of weeks ago and noticed banding on the footage. I thought it was user error on my part, being new to both the video assist and the FS5. Good to know that that wasn't the case. I'm wondering if I should return the FS5 and wait and see if Sony will fix the problem before committing to it. 

Link to post
Share on other sites

It looks like he's saying that 3200 isn't native ISO, which people have stated here that it is. If that's the case then I can see why there would be problems at 3200. Do the same issues occur at 800?

I believe Chapman is stating that the native ISO is 3200 in S-Log, where as in Cinegamma or Standard Gamma the native ISO is 800 / 1000

Native ISOs don't result in noise free images, and not all of the problems that he's responding to (or dismissing, depending on your point of view) are not only caused by the ISO, but compression and codec, bit depth, data rate, sensor, processor etc., all contribute to the issues. 

If anyone has shot in Cinegamma at ISO 800 in both HD and UHD in various shooing conditions I'd be interested to see the results

 

 

Link to post
Share on other sites
  • Administrators

The thing which is most annoying about Chapman's headline is that it's basically covering up for the ripped edges bug by suggesting that our expectations are too lofty for 8bit and compression.

The ripped edges bug as Alister even admits himself has nothing to do with compression or 8bit and everything to do with flawed temporal noise reduction specific to the FS5.

You can't shoot a serious project at high ISOs with this being the case.

You'd have to swap the camera out during a shoot for the A7R II or A7S II. Hardly an ideal situation especially when you have paid close to $6k for your A-cam.

Cine gamma at base ISO 800 or 1000 doesn't have the edge ripping bug. Presumably because at low ISOs the camera uses a different NR technique or turns it off altogether?

Anyway no matter how many excuses Chapman comes up with or how much advice he gives the shooter on how to get the best image out of the camera, the fact remains it's not acceptable for Sony to put out a worse performing codec and a buggy noise reduction technique on a pro camera, when their much cheaper consumer cameras work fine.

I appreciate everything you guys are doing to bring this issue to Sony's attention. This is what I've experienced so far with the FS5... https://www.youtube.com/watch?v=CEh7oxQ89cs

I see it a bit in your 1080p clip too, though not as bad as in 4K!

As a huge Sony fan and someone that is about to buy the FS5 soon, I must say THANK YOU for getting this issue out into the spotlight. In order to get this fixed, it will take some degree of customer pressure to get Sony take this problem seriously and develop a fix. This sites and others are doing a great job at explaining the problem and making users aware of what is really happening to their footage.

As a shooter, there is nothing worse than thinking you are recording OK in the field only to come home and getting shocked at what you see afterwords.

Good job and thanks again!  

Thanks Cliff. Sony are aware of the problem, let's hope they respond with a solution soon.

I wish they had a version of S-LOG with minimal in-camera processing for noise, etc. Then we could apply noise reduction in post instead for superior results.

Link to post
Share on other sites

Other than the edge tear issue (presumably caused by NR which remains present regardless of where the signal is output) is it safe to say that the other issues, such as banding, compression artefacts such as macro blocking etc., are improved greatly if used with an external recorder in both HD and UHD, s-log or otherwise? 

There are many comments about the contradictory practice of used something like a Shogun or 7Q with the Fs5 as it is no longer fulfils its run & gun / grab & shoot (or whatever branding Sony and its paid cronies want to push) function, but the fact is I don't want the expense or hassle of replacing tripods, sliders, jibs, Movi-10 etc if the recorder can easily be attached to a bar or magic arm to allow the Fs5 to work to a reasonable standard (and by reasonable I mean in comparison to the performance of similar priced or cheaper cameras). 

 

Plus this will avoid the issue of the Premiere Pro unfriendly codec when recorded as a ProRes file.

Link to post
Share on other sites

Well, seems like my experience on the FS7 with XAVC-L is indeed the same with the FS5! It's definitely the quality of the codec which is the offending element, as XAVC-L is pretty terrible on both cameras. 

I think Alister is a fantastic resource for Sony cameras, and I don't think he should be protecting the FS5 when the A7 cameras offer better codec performance for less money. 

I'm not a scientist or technician but.... I do think he is right about Slog and low light though. It's true you can get a smoother curve with Slog from shadows to highlights, however I picked up that Alister is reflecting on how much data is being thrown away in Slog in low light. You are stretching information in scenes where there is actually little information to record (such as very dark shadows) thus amplifying noise and artefacts.  Squashed into an 8-bit codec. So it makes more sense to put this information on a gamma curve where more data can be recorded. 

A huge lack of understanding for many A7S enthusiasts is they think that scenes can be lit by just bumping up the ISO and shooting in Slog for higher dynamic range. They are just amplifying information that isn't there, which greatly affects colour rendition, resolution, noise etc. This is the reason for bucketloads of very poor footage from the camera on Vimeo.

Instead of focusing attention on camera tech, more people should focus on how to use light. This alone is the single, biggest quality difference between one image to another. Many of us are forever chasing the resolution/bit rate/HFR race with ignorance towards more important pieces to the craft. 

Buy one body, learn it, stick with it and spend your other money and time on lighting/composition and how to use it. 

Link to post
Share on other sites
  • Administrators

Well, seems like my experience on the FS7 with XAVC-L is indeed the same with the FS5! It's definitely the quality of the codec which is the offending element, as XAVC-L is pretty terrible on both cameras.

Someone really needs to prove the difference between 8bit and 10bit with images and not theories. I'm tired of putting 8bit and 10bit from the same camera, say the GH4, on the timeline in Resolve or Premiere, grading the hell out of it but having them both respond exactly the same to the grade! It's not what'd you'd expect.

In Joe Simon's recent test of the FS7 it did keep the sky intact with no banding in 10bit, so XAVC-L can't suck that much.

We need more examples like this.

I speak highly of 8bit most of the time. There's a reason Canon's JPEGs are 8bit not 10bit... they look superb in 8bit. We keep getting promised smooth gradation with 10bit and a lack of macro blocking at high bitrates but it never really materialises in the footage.

FS5 10bit and 4K 8bit, even 8bit AVCHD 1080p, they all seem to grade similarly.

Same with GH4 10bit over external recorder.

All very disappointing.

Blackmagic Pocket Cinema Camera in 10bit ProRes even, in most situations grades the same as the GH4 in 4K 8bit. Most cameras have more noise than smooth roll off in their gradients.

I think Alister is a fantastic resource for Sony cameras, and I don't think he should be protecting the FS5 when the A7 cameras offer better codec performance for less money.

I am confident Sony will fix this in a firmware update and because Alister has suddenly decided to get nasty and personal in his blog post comments regarding my article, when that firmware update materialises and the edge fuzziness is no longer there, I will do my damnedest to remind both him and Sony of just what an unnecessarily unhelpful situation he created with THAT cover-up post - blaming the users.

He should be quite frankly ashamed of himself for his behaviour and wilful ignorance to the evidence online.

And some of the stuff he comes out with is just fundamentally wrong.

I'm not a scientist or technician but.... I do think he is right about Slog and low light though. It's true you can get a smoother curve with Slog from shadows to highlights, however I picked up that Alister is reflecting on how much data is being thrown away in Slog in low light. You are stretching information in scenes where there is actually little information to record (such as very dark shadows) thus amplifying noise and artefacts.  Squashed into an 8-bit codec. So it makes more sense to put this information on a gamma curve where more data can be recorded.

It depends. It's not as simple as *always* going to a Cine profile in low light rather than using SLOG.

It depends on the light distribution and density in that given situation.

It depends on the LUT you need to use and matching shots across different cameras on the same low light scene.

It depends how the Cine profile you might use respond in the grade and whether it will look consistent with your other SLOG scenes in better light.

Most of all it depends on if the audience will even notice if you have 5 shades of grey instead of 3, or whether they will notice the window covering half the frame completely blown out from not having enough dynamic range in your Cine profile to render it properly!!

A huge lack of understanding for many A7S enthusiasts is they think that scenes can be lit by just bumping up the ISO and shooting in Slog for higher dynamic range.

There are no invalid techniques in filmmaking. If that's what someone wants to do and the look they want to create, just let them get on with it. It's not right or wrong creatively.

They are just amplifying information that isn't there, which greatly affects colour rendition, resolution, noise etc. This is the reason for bucketloads of very poor footage from the camera on Vimeo.

Poor grading can also amplify the wrong information in the image.

Yes good lighting will help you capture more higher quality information.

But poor lighting is not the only reason why there's a bucketload of bad footage from SLOG cameras on Vimeo.

Instead of focusing attention on camera tech, more people should focus on how to use light. This alone is the single, biggest quality difference between one image to another. Many of us are forever chasing the resolution/bit rate/HFR race with ignorance towards more important pieces to the craft.

It's all important. The whole lot.

Buy one body, learn it, stick with it and spend your other money and time on lighting/composition and how to use it. 

Not always practical. The most important thing is to train your eye to see the best existing lighting and composition given the limits of the available natural light you have in any given situation.

Other than the edge tear issue (presumably caused by NR which remains present regardless of where the signal is output) is it safe to say that the other issues, such as banding, compression artefacts such as macro blocking etc., are improved greatly if used with an external recorder in both HD and UHD, s-log or otherwise? 

There are many comments about the contradictory practice of used something like a Shogun or 7Q with the Fs5 as it is no longer fulfils its run & gun / grab & shoot (or whatever branding Sony and its paid cronies want to push) function, but the fact is I don't want the expense or hassle of replacing tripods, sliders, jibs, Movi-10 etc if the recorder can easily be attached to a bar or magic arm to allow the Fs5 to work to a reasonable standard (and by reasonable I mean in comparison to the performance of similar priced or cheaper cameras).

For me the 10bit internal 1080p 4:2:2 on the FS5 was a nice selling point, one of the reasons I bought it.

I expected that 10bit 4:2:2 to clear up the image, but instead I see no advantage over the 8bit 4K and instead a ton of macro blocking. It simply isn't good enough for Sony to market this mode as having "nuanced gradation" when it actually does not.

Yes you're right the external recorder is not an ideal solution as it makes it more of a tripod based controlled set camera rather than a grab and shoot / run and gun / whatever the current politically correct term is for a handheld camera!

The edge tearing doesn't go away on the HDMI or SDI via uncompressed output. The macro blocking should do in theory, let's see if that's the case. To me, uncompressed HDMI never looks uncompressed. The only truly uncompressed image is a raw one.

Link to post
Share on other sites

Someone really needs to prove the difference between 8bit and 10bit with images and not theories. I'm tired of putting 8bit and 10bit from the same camera, say the GH4, on the timeline in Resolve or Premiere, grading the hell out of it but having them both respond exactly the same to the grade! It's not what'd you'd expect.

In Joe Simon's recent test of the FS7 it did keep the sky intact with no banding in 10bit, so XAVC-L can't suck that much.

We need more examples like this.

I speak highly of 8bit most of the time. There's a reason Canon's JPEGs are 8bit not 10bit... they look superb in 8bit. We keep getting promised smooth gradation with 10bit and a lack of macro blocking at high bitrates but it never really materialises in the footage.

FS5 10bit and 4K 8bit, even 8bit AVCHD 1080p, they all seem to grade similarly.

Same with GH4 10bit over external recorder.

All very disappointing.

Blackmagic Pocket Cinema Camera in 10bit ProRes even, in most situations grades the same as the GH4 in 4K 8bit. Most cameras have more noise than smooth roll off in their gradients.

I am confident Sony will fix this in a firmware update and because Alister has suddenly decided to get nasty and personal in his blog post comments regarding my article, when that firmware update materialises and the edge fuzziness is no longer there, I will do my damnedest to remind both him and Sony of just what an unnecessarily unhelpful situation he created with THAT cover-up post - blaming the users.

He should be quite frankly ashamed of himself for his behaviour and wilful ignorance to the evidence online.

And some of the stuff he comes out with is just fundamentally wrong.

It depends. It's not as simple as *always* going to a Cine profile in low light rather than using SLOG.

It depends on the light distribution and density in that given situation.

It depends on the LUT you need to use and matching shots across different cameras on the same low light scene.

It depends how the Cine profile you might use respond in the grade and whether it will look consistent with your other SLOG scenes in better light.

Most of all it depends on if the audience will even notice if you have 5 shades of grey instead of 3, or whether they will notice the window covering half the frame completely blown out from not having enough dynamic range in your Cine profile to render it properly!!

There are no invalid techniques in filmmaking. If that's what someone wants to do and the look they want to create, just let them get on with it. It's not right or wrong creatively.

Poor grading can also amplify the wrong information in the image.

Yes good lighting will help you capture more higher quality information.

But poor lighting is not the only reason why there's a bucketload of bad footage from SLOG cameras on Vimeo.

It's all important. The whole lot.

Not always practical. The most important thing is to train your eye to see the best existing lighting and composition given the limits of the available natural light you have in any given situation.

I'm not going to comment every reply there, I'l just say they were informative and good to read. 

The whole thing is very subjective, and what technically doesn't work for one user is another user's creative holy grail. I've left shots in my work that are full of technical errors, but eh ... I got the shot yeah? ;)

I think my point with lighting is that many times people point their finger at the camera (ugly colour, loads of noise!!) when they haven't actually taken the steps to understand how to work with their tools. I can see what Alister is getting at by blaming the users with a tech-savvy explanation (regarding the FS5), however in this circumstance I know from experience using Sony cameras and their codecs that the fault is with Sony (poor XAVC-L codec). 

On the FS7, a perfectly exposed shot in XAVC-L will show hideous macro blocking on a magnolia painted walk, for example. I personally don't want to worry about things like that while shooting when I've done everything I can to expose the image and maximise technical quality.  For that reason, the FS5 (currently) would drive me nuts. 

 

Link to post
Share on other sites
  • Administrators

On the FS7, a perfectly exposed shot in XAVC-L will show hideous macro blocking on a magnolia painted walk, for example. I personally don't want to worry about things like that while shooting when I've done everything I can to expose the image and maximise technical quality.  For that reason, the FS5 (currently) would drive me nuts. 

Is that in 10bit?

Compression is the enemy of skies and walls isn't it! They need a WALL MODE on the camera :)

It's really bad if the FS7 does this in 10bit 4K. I didn't think it did.

They need to prioritise less detailed areas in the image, give them more of bit rate allocation.

Good point about having a camera that you don't have to worry about undoing your good work in terms of lighting and setting up the shot. A camera with a limitation I can work around with good lighting (i.e. a noisy high ISO camera) is preferable in that situation to one with a severe codec retardation.

Do you have any 'broken' FS7 original files you can share to give me an idea of how severe the problem is?

Canon's codec expertise seems better. Doubt the C300 II has such issues.

To be fair to XAVC-L though, I am able to bring the shadows / blacks up A HELL OF A LOT on my FS5, more than I remember being able to do with XAVC-S. Very little noise in them.

Link to post
Share on other sites

 

On the FS7, a perfectly exposed shot in XAVC-L will show hideous macro blocking on a magnolia painted walk, for example. I personally don't want to worry about things like that while shooting when I've done everything I can to expose the image and maximise technical quality.  For that reason, the FS5 (currently) would drive me nuts. 

 

I don't understand why you wouldn't shoot in XAVC-I then?

Link to post
Share on other sites

I don't understand why you wouldn't shoot in XAVC-I then?

Argh not this again.... I do shoot in XAVC-I, but like I've said before sometimes I've had to switch due to a lack of card space or because I'm comparing it to an XAVC-I shot for laughs. 

I've attached a shot here which I recorded at the end of a shoot. Cards were nearly out of space so just recorded some stuff for giggles - still properly lit and all that. Look at the wallpaper though. Slog3 with a bit of grading. XAVC-L. (XAVC-I doesn't do this). 

Is that in 10bit?

Compression is the enemy of skies and walls isn't it! They need a WALL MODE on the camera :)

It's really bad if the FS7 does this in 10bit 4K. I didn't think it did.

They need to prioritise less detailed areas in the image, give them more of bit rate allocation.

Good point about having a camera that you don't have to worry about undoing your good work in terms of lighting and setting up the shot. A camera with a limitation I can work around with good lighting (i.e. a noisy high ISO camera) is preferable in that situation to one with a severe codec retardation.

Do you have any 'broken' FS7 original files you can share to give me an idea of how severe the problem is?

Canon's codec expertise seems better. Doubt the C300 II has such issues.

To be fair to XAVC-L though, I am able to bring the shadows / blacks up A HELL OF A LOT on my FS5, more than I remember being able to do with XAVC-S. Very little noise in them.

 

It can wrok in a lot of situations. Some of my shots look great and it grades pretty well. But it breaks. A lot. Sadly i can't share actual files because of the contracts. 

macro block 2.jpg

Link to post
Share on other sites
  • Administrators

That is in 10bit on the FS7?

Oh dear!

Maybe FS5 not so far off then.

XAVC-L is such a disappointment.

If they had some noise to dither the banding or to keep low detail areas of the image intact from too much compression, it would really help.

Turn the noise reduction off Sony!

Link to post
Share on other sites

That is in 10bit on the FS7?

Oh dear!

Maybe FS5 not so far off then.

XAVC-L is such a disappointment.

If they had some noise to dither the banding or to keep low detail areas of the image intact from too much compression, it would really help.

Turn the noise reduction off Sony!

Yes indeed!

I've said it before and I'll say it again. XAVC-L is a bag of crap (sometimes, not always, but enough to annoy the hell out of you). 

Link to post
Share on other sites

Argh not this again.... I do shoot in XAVC-I, but like I've said before sometimes I've had to switch due to a lack of card space or because I'm comparing it to an XAVC-I shot for laughs. 

 

 

Everyone has to get their kicks somewhere I suppose....

If anything your test shot has reminded me to take plenty of cards with me to paid shoots.

Link to post
Share on other sites

To be fair, this is not a camera I need to own.

However, if it were, I don't think I'd have the stomach to gamble on all of these issues being fixed -- and fixed in a timely fashion.

This isn't a single "black hole where the sun should be" problem.

And if enough of these issues don't get fixed -- what are you left with -- and how are you going to sell it?

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...