K-5 jpeg one * burst

falconeye · 11-26-2010, 03:12 AM

Originally posted by RonHendriks1966

Same test with the K-5, 16 megapixel one * quality. This brings in one minute 238 Jpeg's, that's just under 4 fps and a datastream of 8,0 MB/s (2,0 MB average).

I've made numerous comments on the issue, starting Sep. 20 for the K-5.

My latest comment was that I only get 44 images whatever I do. The smallest possible JPGs are from lens cap shots (2 MP, 1 star) resulting in only 0.8 MByte/s datastream. All settings manual, incl. exposure and focus. Sandisk Extreme III card. As you say, K-7 streams forever.

Your 4 fps forever result most likely is from having AF.C enabled. Which is a good trick showing that the K-5 can maintain 4fps but not 7fps.

Firmware 1.01. With 1.00 it was even less.

Originally posted by Adam

FYI I tried shooting at 2MP and I still got about 34 JPEGs - so I can't help but think that the camera just throws unprocessed data into the buffer.

I confirm the number for regular JPGs (it can go up for 1 star quality).
I don't think it is unprocessed data. But some buffer is still limiting things. The raw buffer size doesn't shrink with RAW+JPG so it's not final SD card data written to the buffer indeed.

Originally posted by WerTicus

1080p video is 2mp, so may as well just film at 25fps

Video is working with completely different data. Argument not applicable here.

I notice different differences between K-7 and K-5 buffer handling too which is more cumbersome for me.

If I use the remote trigger, the K-7 remains responsive after the first shot for about 8 more shots. The K-5 starts to become a bit unresponsive after the first or second shot but doesn't have the break-down after about 9 shots. That's a problem for people using a remote wireless trigger in the studio. Fortunately, the K-5 behaves just as it should (remains responsive) with direct shutter presses. So, part of the buffer processing must stall the remote receiver and I consider it a bug.

Last edited by falconeye; 11-26-2010 at 04:08 AM.

dosdan · 11-26-2010, 04:55 AM

Originally posted by falconeye

f I use the remote trigger, the K-7 remains responsive after the first shot for about 8 more shots. The K-5 starts to become a bit unresponsive after the first or second shot but doesn't have the break-down after about 9 shots. That's a problem for people using a remote wireless trigger in the studio.

I presume a wired remote would be fine though.

Dan

falconeye · 11-26-2010, 06:56 AM

Originally posted by dosdan

I presume a wired remote would be fine though.

Dan, I think so too. But the regression bug with the K-5 remote trigger surprised me.

A wired trigger would have the additional benefit of an AF button. Then, there is a new IR trigger from Pentax with an AF button too. I'll see if the new Pentax remote trigger works better. Otherwise, I'll report it as bug to Pentax.

RonHendriks1966 · 11-26-2010, 07:51 AM

Originally posted by nanok

(16*(2^14))/(14.6*(2^12))
4.38356164383561643835

perhaps where it all starts?

(hint: the buffer is not holding (only) jpegs..)

I understand your calculation. Just wondering if this could be true? If the K-5 handles the Jpeg's like RAW then it is amazing it even get's to the total number of 238. So I don't think that can be the case.

Originally posted by Adam

FYI I tried shooting at 2MP and I still got about 34 JPEGs - so I can't help but think that the camera just throws unprocessed data into the buffer.

With the K-7 is the same. Smaller files make you think they are more easier to handle, but the camera makes a full size picture and starts then calculating them back to a 2 megapixel version wich will take more processing time.

Originally posted by falconeye

Your 4 fps forever result most likely is from having AF.C enabled. Which is a good trick showing that the K-5 can maintain 4fps but not 7fps.

Well it is just a little slower. I forgot the first burst until full buffer, so the speed is some 3,6 fps after full buffer. The buffer is very quickly empty, it takes only 5 seconds to write all data after the last shot is taken.

So just one more test:
In Liveview one star jpg's just a fast burst until buffer is full and slowing down, or just stopping a little before. A 5 second burst and then flushing the datastream. Taking between 31 and 36 shots and flushing time including taking pictures never exceded total time of 10 seconds (some 9,5 seconds not very accurate).
So the total data amound was between 55 and 63 MB and thus making the SD-card never to be the problem. Datastream is rather small.

**ManuH** · 11-26-2010, 08:16 AM

I was surprised too by the back LCD being completely unresponsive during a short burst. If you're using the VF, you can still work and shoot. If using LV you're out of luck and have to wait, the camera doesn't even respond to the LV button (to get out of LV).

My K-x could also burst unlimited (AF-S, 4.7fps) on one-star jpegs.

falconeye · 11-26-2010, 09:46 AM

Originally posted by ManuH

I was surprised too by the back LCD being completely unresponsive during a short burst. If you're using the VF, you can still work and shoot. If using LV you're out of luck and have to wait, the camera doesn't even respond to the LV button (to get out of LV).

My K-x could also burst unlimited (AF-S, 4.7fps) on one-star jpegs.

Wait a second. I may be mistaken.

Actually, when using the remote trigger I was in LV and when using the shutter button I was in optical VF. So, it may rather be the buffer handling in LV mode which is the problem. Of course, using LV and remote trigger in combination is not a combination unheard of

Regarding fps:

The initial result (3.6 fps) was most likely with AF.C. I guess there is a barrier between 5 and 7 fps for infinite bursts.

No, 2MP is not more computational work. Downsampling is cheaper than JPG compression. And JPG compression is always the same computational work independent on compression settings. It depends on bit depth and #pixels only.

I just checked:

Yes, it is LV.
In optical VF, pressing shutter 5x in fast succession: no problem.
In LV, pressing shutter 5x in fast succession: only 1 or 2 images!

The K-7 stays a bit more responsive in LV, but not by a large margin.

RonHendriks1966 · 11-26-2010, 09:57 AM

Originally posted by falconeye

Regarding fps:

The initial result (3.6 fps) was most likely with AF.C. I guess there is a barrier between 5 and 7 fps for infinite bursts.

No it was in AF-s so no limitation there, should/could run on 7 fps but doesn't.

falconeye · 11-26-2010, 10:46 AM

Originally posted by RonHendriks1966

No it was in AF-s so no limitation there, should/could run on 7 fps but doesn't.

Do you mean then the sustained rate after the initial burst of about 40 images?

Yes, this rate is slower then 7fps and slower than K-7's 5fps.

But it has nothing to do with SD card speed. Some internal processing step which sits between an early processing and the SD card write buffer isn't faster than this. We don't know what it is. It's not the JPG compressor which is known to be faster than this (at least 4.7fps at 16MP).

It certainly would be a good idea if Pentax added a mid speed mode which sustains the card write speed.

**jpzk** · 11-26-2010, 10:57 AM

Originally posted by vievetrick

Do you have k5 set to include your copywrite info?

Just saw this post: does it really affect the write speed?
I have mine set that way (on the K7), so maybe I should have it off?

JP

**ManuH** · 11-26-2010, 11:23 AM

Originally posted by RonHendriks1966

No it was in AF-s so no limitation there, should/could run on 7 fps but doesn't.

Was it in low light? I've noticed that in modes other than M the fps slow down if you're in relatively low light. So if maximum fps is what you want, put everything in manual (Focus, meter, WB, NR off).

(BTW you can also use AE-lock, it will allow you to get faster fps too).

**ManuH** · 11-26-2010, 11:32 AM

Originally posted by falconeye

I just checked:

Yes, it is LV.
In LV, pressing shutter 5x in fast succession: only 1 or 2 images!

Were you in single shot mode? I see the lag in that mode but strangely in hi-fps mode, it seems a bit more responsive. Possibly a firmware bug?

**ChipB** · 11-26-2010, 11:56 AM

One thing to make sure of is that both cameras are shooting at the same shutter speed - slow(er) shutter speed will obviously affect the number of shots in a period of time.

falconeye · 11-26-2010, 12:24 PM

Originally posted by ManuH

Were you in single shot mode? [...] Possibly a firmware bug?

Yes.
Maybe not a bug but the firmware has potential for improvement in that area.

**ChipB** · 11-26-2010, 12:35 PM

Another thing to consider is the file size - just because the 1* option for JPEG was selected doesn't mean the file sizes between different shots are the same. I just took a photo of a relatively simple scene - 1.61MB - and an outdoor, very busy scene - 1.91MB.

Manel Brand · 11-26-2010, 12:38 PM

Originally posted by RonHendriks1966

So what is holding the K-5 back?

Just guessing (I don't have it, yet). Lateral chromatic aberration is set to ON or maybe that Lens Correction thing?

11-26-2010, 09:46 AM	#21
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871	Originally posted by ManuH I was surprised too by the back LCD being completely unresponsive during a short burst. If you're using the VF, you can still work and shoot. If using LV you're out of luck and have to wait, the camera doesn't even respond to the LV button (to get out of LV). My K-x could also burst unlimited (AF-S, 4.7fps) on one-star jpegs. Wait a second. I may be mistaken. Actually, when using the remote trigger I was in LV and when using the shutter button I was in optical VF. So, it may rather be the buffer handling in LV mode which is the problem. Of course, using LV and remote trigger in combination is not a combination unheard of Regarding fps: The initial result (3.6 fps) was most likely with AF.C. I guess there is a barrier between 5 and 7 fps for infinite bursts. No, 2MP is not more computational work. Downsampling is cheaper than JPG compression. And JPG compression is always the same computational work independent on compression settings. It depends on bit depth and #pixels only. I just checked: Yes, it is LV. In optical VF, pressing shutter 5x in fast succession: no problem. In LV, pressing shutter 5x in fast succession: only 1 or 2 images! The K-7 stays a bit more responsive in LV, but not by a large margin. Last edited by falconeye; 11-26-2010 at 09:57 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Nature About to Burst............	eaglem	Post Your Photos!	3	10-07-2010 10:34 PM
Burst mode	Orli	Troubleshooting and Beginner Help	3	02-22-2009 06:20 PM
RAW + JPEG with JPEG on One Star quality	laissezfaire	Pentax DSLR Discussion	58	12-10-2008 02:42 PM
Burst Mode	shefaet	Pentax News and Rumors	42	01-25-2008 08:50 AM

11-26-2010, 03:12 AM	#16
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871	Originally posted by RonHendriks1966 Same test with the K-5, 16 megapixel one * quality. This brings in one minute 238 Jpeg's, that's just under 4 fps and a datastream of 8,0 MB/s (2,0 MB average). I've made numerous comments on the issue, starting Sep. 20 for the K-5. My latest comment was that I only get 44 images whatever I do. The smallest possible JPGs are from lens cap shots (2 MP, 1 star) resulting in only 0.8 MByte/s datastream. All settings manual, incl. exposure and focus. Sandisk Extreme III card. As you say, K-7 streams forever. Your 4 fps forever result most likely is from having AF.C enabled. Which is a good trick showing that the K-5 can maintain 4fps but not 7fps. Firmware 1.01. With 1.00 it was even less. Originally posted by Adam FYI I tried shooting at 2MP and I still got about 34 JPEGs - so I can't help but think that the camera just throws unprocessed data into the buffer. I confirm the number for regular JPGs (it can go up for 1 star quality). I don't think it is unprocessed data. But some buffer is still limiting things. The raw buffer size doesn't shrink with RAW+JPG so it's not final SD card data written to the buffer indeed. Originally posted by WerTicus 1080p video is 2mp, so may as well just film at 25fps Video is working with completely different data. Argument not applicable here. I notice different differences between K-7 and K-5 buffer handling too which is more cumbersome for me. If I use the remote trigger, the K-7 remains responsive after the first shot for about 8 more shots. The K-5 starts to become a bit unresponsive after the first or second shot but doesn't have the break-down after about 9 shots. That's a problem for people using a remote wireless trigger in the studio. Fortunately, the K-5 behaves just as it should (remains responsive) with direct shutter presses. So, part of the buffer processing must stall the remote receiver and I consider it a bug. Last edited by falconeye; 11-26-2010 at 04:08 AM.

11-26-2010, 04:55 AM	#17
dosdan Pentaxian Join Date: Dec 2007 Location: Brisbane, Australia Photos: Gallery \| Albums Posts: 1,741	Originally posted by falconeye f I use the remote trigger, the K-7 remains responsive after the first shot for about 8 more shots. The K-5 starts to become a bit unresponsive after the first or second shot but doesn't have the break-down after about 9 shots. That's a problem for people using a remote wireless trigger in the studio. I presume a wired remote would be fine though. Dan

11-26-2010, 06:56 AM	#18
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871	Originally posted by dosdan I presume a wired remote would be fine though. Dan, I think so too. But the regression bug with the K-5 remote trigger surprised me. A wired trigger would have the additional benefit of an AF button. Then, there is a new IR trigger from Pentax with an AF button too. I'll see if the new Pentax remote trigger works better. Otherwise, I'll report it as bug to Pentax.

11-26-2010, 07:51 AM	#19
RonHendriks1966 Banned Join Date: Jan 2009 Photos: Gallery Posts: 9,675 Original Poster	Originally posted by nanok (16(2^14))/(14.6(2^12)) 4.38356164383561643835 perhaps where it all starts? (hint: the buffer is not holding (only) jpegs..) I understand your calculation. Just wondering if this could be true? If the K-5 handles the Jpeg's like RAW then it is amazing it even get's to the total number of 238. So I don't think that can be the case. Originally posted by Adam FYI I tried shooting at 2MP and I still got about 34 JPEGs - so I can't help but think that the camera just throws unprocessed data into the buffer. With the K-7 is the same. Smaller files make you think they are more easier to handle, but the camera makes a full size picture and starts then calculating them back to a 2 megapixel version wich will take more processing time. Originally posted by falconeye Your 4 fps forever result most likely is from having AF.C enabled. Which is a good trick showing that the K-5 can maintain 4fps but not 7fps. Well it is just a little slower. I forgot the first burst until full buffer, so the speed is some 3,6 fps after full buffer. The buffer is very quickly empty, it takes only 5 seconds to write all data after the last shot is taken. So just one more test: In Liveview one star jpg's just a fast burst until buffer is full and slowing down, or just stopping a little before. A 5 second burst and then flushing the datastream. Taking between 31 and 36 shots and flushing time including taking pictures never exceded total time of 10 seconds (some 9,5 seconds not very accurate). So the total data amound was between 55 and 63 MB and thus making the SD-card never to be the problem. Datastream is rather small.

11-26-2010, 08:16 AM	#20
ManuH Site Supporter Join Date: Dec 2007 Location: Montreal Photos: Gallery \| Albums Posts: 1,249	I was surprised too by the back LCD being completely unresponsive during a short burst. If you're using the VF, you can still work and shoot. If using LV you're out of luck and have to wait, the camera doesn't even respond to the LV button (to get out of LV). My K-x could also burst unlimited (AF-S, 4.7fps) on one-star jpegs.

11-26-2010, 09:57 AM	#22
RonHendriks1966 Banned Join Date: Jan 2009 Photos: Gallery Posts: 9,675 Original Poster	Originally posted by falconeye Regarding fps: The initial result (3.6 fps) was most likely with AF.C. I guess there is a barrier between 5 and 7 fps for infinite bursts. No it was in AF-s so no limitation there, should/could run on 7 fps but doesn't.

11-26-2010, 10:46 AM	#23
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871	Originally posted by RonHendriks1966 No it was in AF-s so no limitation there, should/could run on 7 fps but doesn't. Do you mean then the sustained rate after the initial burst of about 40 images? Yes, this rate is slower then 7fps and slower than K-7's 5fps. But it has nothing to do with SD card speed. Some internal processing step which sits between an early processing and the SD card write buffer isn't faster than this. We don't know what it is. It's not the JPG compressor which is known to be faster than this (at least 4.7fps at 16MP). It certainly would be a good idea if Pentax added a mid speed mode which sustains the card write speed.

11-26-2010, 10:57 AM	#24
jpzk Site Supporter Join Date: Jan 2009 Location: Québec Photos: Gallery \| Albums Posts: 8,251	Originally posted by vievetrick Do you have k5 set to include your copywrite info? Just saw this post: does it really affect the write speed? I have mine set that way (on the K7), so maybe I should have it off? JP

11-26-2010, 11:23 AM	#25
ManuH Site Supporter Join Date: Dec 2007 Location: Montreal Photos: Gallery \| Albums Posts: 1,249	Originally posted by RonHendriks1966 No it was in AF-s so no limitation there, should/could run on 7 fps but doesn't. Was it in low light? I've noticed that in modes other than M the fps slow down if you're in relatively low light. So if maximum fps is what you want, put everything in manual (Focus, meter, WB, NR off). (BTW you can also use AE-lock, it will allow you to get faster fps too).

11-26-2010, 11:32 AM	#26
ManuH Site Supporter Join Date: Dec 2007 Location: Montreal Photos: Gallery \| Albums Posts: 1,249	Originally posted by falconeye I just checked: Yes, it is LV. In LV, pressing shutter 5x in fast succession: only 1 or 2 images! Were you in single shot mode? I see the lag in that mode but strangely in hi-fps mode, it seems a bit more responsive. Possibly a firmware bug?

Thread Tools	Search this Thread
	Search this Thread: Advanced Search

11-26-2010, 11:56 AM	#27
ChipB Site Supporter Join Date: Nov 2007 Location: Eau Claire, WI Photos: Gallery \| Albums Posts: 2,701	One thing to make sure of is that both cameras are shooting at the same shutter speed - slow(er) shutter speed will obviously affect the number of shots in a period of time.

11-26-2010, 12:24 PM	#28
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871	Originally posted by ManuH Were you in single shot mode? [...] Possibly a firmware bug? Yes. Maybe not a bug but the firmware has potential for improvement in that area.

11-26-2010, 12:35 PM	#29
ChipB Site Supporter Join Date: Nov 2007 Location: Eau Claire, WI Photos: Gallery \| Albums Posts: 2,701	Another thing to consider is the file size - just because the 1* option for JPEG was selected doesn't mean the file sizes between different shots are the same. I just took a photo of a relatively simple scene - 1.61MB - and an outdoor, very busy scene - 1.91MB.

11-26-2010, 12:38 PM	#30
Manel Brand Veteran Member Join Date: Sep 2008 Location: Porto Photos: Gallery \| Albums Posts: 853	Originally posted by RonHendriks1966 So what is holding the K-5 back? Just guessing (I don't have it, yet). Lateral chromatic aberration is set to ON or maybe that Lens Correction thing?