Author Topic: D850 frame rate vs buffer  (Read 9489 times)

chambeshi

  • Guest
D850 frame rate vs buffer
« on: October 15, 2017, 14:37:28 »
For those who have the MB-18 grip - excerpt of trials by Brad Hill

http://www.naturalart.ca/voice/blog.html

14 October 2017: Nikon D850 Burst Depth - It's ALL About Frame Rate!

What did I find? The results are VERY simple to explain - 4 of 6 variables I tested had absolutely NO affect on burst depth. The 4 variables having NO AFFECT on burst depth were:
• Auto ISO vs. Fixed ISO
• ISO Value
• XQD Card Type
• Secondary Card Slot status
Which two variables had any effect on burst depth? Scene complexity had a measurable - but almost trivial - effect: bursts of the simple scene (blue sky) were - on average - 1 frame longer than those of the moderately complex and complex scenes.
HOWEVER, Frame Rate had a MAJOR impact on burst depth. Here are my results:
• At 7 fps (Continuous High with NO battery grip): Burst depth = 40 frames
• At 8 fps (Continuous Low highest rate with Battery Grip ON): Burst depth = 30 frames
• At 9 fps (Continuous High with Battery Grip): Burst depth = 25 frames

So...at least over the range of 7-9 fps, the higher the frame rate the lower the burst depth.

Ilkka Nissilä

  • NG Member
  • *
  • Posts: 1714
  • You ARE NikonGear
Re: D850 frame rate vs buffer
« Reply #1 on: October 15, 2017, 15:49:54 »
The menu guide on p. 269 have a table of the burst depth at 9fps. User manual p. 362 shows the corresponding specs for gripless camera at 7fps.

Nikon’s values are a bit higher than Hill’s. I wonder if Hill has some in-camera setting on which could affect the burst depth. I would try with low ISO and all post processing (distortion and vignetting correction as well as noise reduction and ADL off) and a fast shutter speed, manual exposure to see if Nikon’s results can be reproduced.

Anyway the result is quite good. 12-bit compressed is probably the setting to use for most action where you expect to use high fps.

Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 12640
  • Bonn, Germany
Re: D850 frame rate vs buffer
« Reply #2 on: October 15, 2017, 18:17:42 »
burst depth is longer with raw medium size says Nikon

I did not yet receive my grip, although I had it preordered and received the camera on day one
You are out there. You and your camera. You can shoot or not shoot as you please. Discover the world, Your world. Show it to us. Or we might never see it.

Me: https://youpic.com/photographer/frankfremerey/

arthurking83

  • NG Member
  • *
  • Posts: 240
  • Good to be back on NikonGear
Re: D850 frame rate vs buffer
« Reply #3 on: October 15, 2017, 21:40:32 »
...
Nikon’s values are a bit higher than Hill’s. I wonder if Hill has some in-camera setting on which could affect the burst depth. I would try with low ISO and all post processing (distortion and vignetting correction as well as noise reduction and ADL off) and a fast shutter speed, manual exposure to see if Nikon’s results can be reproduced..

Nikon's buffer values are marked with an addendum(2) which states those values are for ISO100, which may drop in some situations.
it then continues on that the number may drop for example images marked with a star * (but none are marked as such) .. and that Auto distortion control can affect.
Arthur

Ilkka Nissilä

  • NG Member
  • *
  • Posts: 1714
  • You ARE NikonGear
Re: D850 frame rate vs buffer
« Reply #4 on: October 15, 2017, 23:23:27 »
I guess at higher ISO the noise makes the NEF images compress less effectively (since in lossless compression also the noise has to be included as "information") so that may be why the burst depth is reduced at high ISO.

According to the manual and the menu guide, use of medium or small NEF reduce significantly the burst depth so they're of no help.

Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 12640
  • Bonn, Germany
Re: D850 frame rate vs buffer
« Reply #5 on: October 16, 2017, 00:02:59 »
I was talking to a NPS spokesperson at release and they said they ran their own tests. small raw is not helpful, medium raw is. but. they did not specify if they tested 7fps only or 9fps too.
You are out there. You and your camera. You can shoot or not shoot as you please. Discover the world, Your world. Show it to us. Or we might never see it.

Me: https://youpic.com/photographer/frankfremerey/

Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 12640
  • Bonn, Germany
Re: D850 frame rate vs buffer
« Reply #6 on: October 16, 2017, 00:05:40 »
question to those who have the grip: which batteries work?

original 18; 18a; 18b; replacements: which?
You are out there. You and your camera. You can shoot or not shoot as you please. Discover the world, Your world. Show it to us. Or we might never see it.

Me: https://youpic.com/photographer/frankfremerey/

Ilkka Nissilä

  • NG Member
  • *
  • Posts: 1714
  • You ARE NikonGear
Re: D850 frame rate vs buffer
« Reply #7 on: October 16, 2017, 00:14:32 »
I don't have the MB-D18 grip but I downloaded the manual. It says

"One EN-EL15a or EN-EL18b rechargeable Li-ion battery, eight alkaline (1.5 V)
or lithium (1.5 V) AA batteries, eight Ni-MH (1.2 V) rechargeable AA batteries,
or an EH-5c AC adapter (requires EP-5B power connector); EN-EL15, EN-EL18,
EN-EL18a batteries and EH-5b AC adapters are also supported"

chambeshi

  • Guest
Re: D850 frame rate vs buffer
« Reply #8 on: October 16, 2017, 08:33:19 »

Update by Brad Hill confirms dropping file size to boost buffer performance

15 October 2017: Nikon D850 Burst Depth II - STRETCHING It Out!
http://www.naturalart.ca/voice/blog.html

Ethan

  • NG Supporter
  • **
  • Posts: 208
  • You ARE NikonGear
Re: D850 frame rate vs buffer
« Reply #9 on: October 16, 2017, 10:55:12 »
1- What was the mode used and what was the shutter speed?
1.1- Did he try the result in LV?
1.2- If the shoot is not fixed and using AF plus Auto ISO, wouldn't that affect the results?

I could not find mention of either in his reviews.
Did he used a fixed focus with the camera on a tripod?

2- Why the buffer should be of any concern for the D850? If the buffer is a factor in the shoot then use a D5 or similar.
I don't do testing. If it works for my needs then it is fine. If not, then it's not.
The D850 has a reasonable buffer to get you by. It is not what I would use for high shutter speed and/or filling of buffer.

3- I received my MB-D18 plus battery at the beg of the month and only tried it on to check the balance v the D5. I still prefer the D5 for balance.
The D850 plus a heavy lens plus BGrip is rather cumbersome. Add an on camera flash and you will need a Sherpa. If a flash has to be mounted, the D5 is just about the limit of weight and balance.
The D5 is more practical in such cases and I am looking forward next week trying the new PF A1 to go with it.

Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 12640
  • Bonn, Germany
Re: D850 frame rate vs buffer
« Reply #10 on: October 16, 2017, 10:57:18 »
Update by Brad Hill confirms dropping file size to boost buffer performance

Yes, DX instead of FX is clearly having an effect on burst depth, but does reducing to "Medium RAW"  has that effect also?
You are out there. You and your camera. You can shoot or not shoot as you please. Discover the world, Your world. Show it to us. Or we might never see it.

Me: https://youpic.com/photographer/frankfremerey/

Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 12640
  • Bonn, Germany
Re: D850 frame rate vs buffer
« Reply #11 on: October 16, 2017, 11:01:19 »
1- What was the mode used and what was the shutter speed?
1.1- Did he try the result in LV?
1.2- If the shoot is not fixed and using AF plus Auto ISO, wouldn't that affect the results?

I could not find mention of either in his reviews.
Did he used a fixed focus with the camera on a tripod?

2- Why the buffer should be of any concern for the D850? If the buffer is a factor in the shoot then use a D5 or similar.
I don't do testing. If it works for my needs then it is fine. If not, then it's not.
The D850 has a reasonable buffer to get you by. It is not what I would use for high shutter speed and/or filling of buffer.

3- I received my MB-D18 plus battery at the beg of the month and only tried it on to check the balance v the D5. I still prefer the D5 for balance.
The D850 plus a heavy lens plus BGrip is rather cumbersome. Add an on camera flash and you will need a Sherpa. If a flash has to be mounted, the D5 is just about the limit of weight and balance.
The D5 is more practical in such cases and I am looking forward next week trying the new PF A1 to go with it.


1) Regarding the buffer depth it is clear that a huge one is only needed in rare cases, 10 Frames is enough for next to any real world situation. I am with you here.

2) What is a PF A1???
You are out there. You and your camera. You can shoot or not shoot as you please. Discover the world, Your world. Show it to us. Or we might never see it.

Me: https://youpic.com/photographer/frankfremerey/

Ilkka Nissilä

  • NG Member
  • *
  • Posts: 1714
  • You ARE NikonGear
Re: D850 frame rate vs buffer
« Reply #12 on: October 16, 2017, 12:23:23 »
2- Why the buffer should be of any concern for the D850? If the buffer is a factor in the shoot then use a D5 or similar.

A high resolution camera with high fps can be useful for sports photography as well as wildlife.  If the buffer is small, this dampens the practical value of the concept of having high resolution and high speed in one camera. Thankfully it seems with 12-bit NEFs it is sufficient for most purposes.

I give an example. I often shoot figure skating. To keep ISO moderate, shutter speed high (the skaters can spin very fast), my first choice is the 200/2 provided that I'm shooting from a position close to the ice. At f/2, especially if the skater is close to filling the frame, I get a nice ethereal blur of the background and avoid excessive disturbance from backgrounds but still it is recognizable that there are spectators, judges etc. in the background. Sometimes the skater's arms and legs are extended like this

https://www.flickr.com/photos/ilkka_nissila/36953629383/in/dateposted-public/

and sometimes the skater is small

https://www.flickr.com/photos/ilkka_nissila/37575916146/in/dateposted-public/

I need to crop the image to accommodate for these variations (and different distances). I don't have enough space to use a supertelephoto zoom like the 200-400 (nor do I own one) and that would push me from ISO 1600 to 6400.  So the option offered by a high resolution camera such as the D850 is attractive. It would allow me to crop from the images from the prime and have a bit more detailed images than if cropping from D5 images.

Is high fps for extended times necessary? No, it is not, I mostly shoot single shots but if I want to get jump sequences of a particular skater

https://www.flickr.com/photos/ilkka_nissila/30275677831/in/dateposted-public/

then yes, high fps can be very useful and the sequence can last several seconds. In fact there are quite many occasions where one might use high fps, though I'm not a big fan of such an approach I recognize that sometimes it's necessary.

As for how much gain there is from shooting with a high resolution camera vs. intermediate resolution in such applications, well, it's to be determined, I feel the details in the dress seem more crisply defined with smoother transitions when using the D810 than when using the D5. However, on the other hand the D5 images are less noisy at high ISO.

In many action situations when working with a prime lens you may find the subject changing distance and the photographer is unable to compensate, so there is cropping in post-processing to get the best final compositions.  And even with a zoom lens you still want some space around the subject to give some margin of error. The fastest focusing and highest quality supertelephoto lenses are still primes. 

Of course there is the disadvantage of large quantity of data that results from high fps capture of high resolution images. However, one might use such technique only for a particular athlete of special interest and then the quantity of data would probably not be an issue. The other issue is balancing between SNR and resolution; at low to medium ISO there is no question high resolution cameras can give more detailed images. At high ISO the extra details may or may not show from the noise, and this is something interesting to see from D850 images. Focus accuracy can also play an important role in the outcome.

Generally I find 20-24MP is well suited to my photography, even though I recognize more detailed images can be obtained using higher resolution sensors; I spend too much time editing and want to reduce that time. However, if the goal is the best possible quality in a specific task then I wouldn't be surprised if the D850 came out ahead in many cases.

Quote
3- I received my MB-D18 plus battery at the beg of the month and only tried it on to check the balance v the D5. I still prefer the D5 for balance.

Without doubt. However, I believe the majority of photographers prefer to have the option of a smaller camerawhen the grip is not needed. I personally use the D810 without grip when I'm shooting landscape etc. but mount the grip usually when photographing people as I typically shoot a lot of verticals with telephoto lenses. However, if I am tight for bag space and especially if it is the second camera I might leave it without the grip.

Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 12640
  • Bonn, Germany
Re: D850 frame rate vs buffer
« Reply #13 on: October 16, 2017, 12:40:42 »
A high resolution camera with high fps can be useful for sports photography as well as wildlife.  If the buffer is small, this dampens the practical value of the concept of having high resolution and high speed in one camera. Thankfully it seems with 12-bit NEFs it is sufficient for most purposes.

I give an example. I often shoot figure skating. To keep ISO moderate, shutter speed high (the skaters can spin very fast), my first choice is the 200/2 provided that I'm shooting from a position close to the ice. At f/2, especially if the skater is close to filling the frame, I get a nice ethereal blur of the background and avoid excessive disturbance from backgrounds but still it is recognizable that there are spectators, judges etc. in the background. Sometimes the skater's arms and legs are extended like this

https://www.flickr.com/photos/ilkka_nissila/36953629383/in/dateposted-public/

and sometimes the skater is small

https://www.flickr.com/photos/ilkka_nissila/37575916146/in/dateposted-public/

I need to crop the image to accommodate for these variations (and different distances). I don't have enough space to use a supertelephoto zoom like the 200-400 (nor do I own one) and that would push me from ISO 1600 to 6400.  So the option offered by a high resolution camera such as the D850 is attractive. It would allow me to crop from the images from the prime and have a bit more detailed images than if cropping from D5 images.

Is high fps for extended times necessary? No, it is not, I mostly shoot single shots but if I want to get jump sequences of a particular skater

https://www.flickr.com/photos/ilkka_nissila/30275677831/in/dateposted-public/

then yes, high fps can be very useful and the sequence can last several seconds. In fact there are quite many occasions where one might use high fps, though I'm not a big fan of such an approach I recognize that sometimes it's necessary.

As for how much gain there is from shooting with a high resolution camera vs. intermediate resolution in such applications, well, it's to be determined, I feel the details in the dress seem more crisply defined with smoother transitions when using the D810 than when using the D5. However, on the other hand the D5 images are less noisy at high ISO.

In many action situations when working with a prime lens you may find the subject changing distance and the photographer is unable to compensate, so there is cropping in post-processing to get the best final compositions.  And even with a zoom lens you still want some space around the subject to give some margin of error. The fastest focusing and highest quality supertelephoto lenses are still primes. 

Without doubt. However, I believe the majority of photographers prefer to have the option of a smaller camerawhen the grip is not needed. I personally use the D810 without grip when I'm shooting landscape etc. but mount the grip usually when photographing people as I typically shoot a lot of verticals with telephoto lenses. However, if I am tight for bag space and especially if it is the second camera I might leave it without the grip.


@Ilkka: I love your work, shown on the linked images! When I see these exceptional moments I ask: What is the burn rate? Do you need 30 shots to get one of those or is it more in the 10 shots arera?
You are out there. You and your camera. You can shoot or not shoot as you please. Discover the world, Your world. Show it to us. Or we might never see it.

Me: https://youpic.com/photographer/frankfremerey/

chambeshi

  • Guest
Re: D850 frame rate vs buffer
« Reply #14 on: October 16, 2017, 12:53:04 »
Yes, DX instead of FX is clearly having an effect on burst depth, but does reducing to "Medium RAW"  has that effect also?
Yes, I just tested - twice - panning against moderate clutter (vegetation). 86 frames in Medium RAW (25.6 M) before a noticeable pause. But I did not change the file depth so I presume this is at 14bit (?)

I also do not have the grip yet. But I plan to. With wildlife there are those rare behavioural reportoires that demand sustained bursts, which the D500 is very good at capturing.