Is "sample variation" another reason to consider primes?

tbirdas · 11-30-2010, 09:00 AM

Reading through the excellent threads in this forum, many users report significant sample variation for certain zooms. My most recent concern is the sigma 10-20 vs the Tamron 10-24. I tried them both (one copy of each) - no glaring decentering issues etc and the picture is quite comparable but not vary sharp (although the sigma may be a little sharper). But how do I know if these were good copies? Even "pro" reviews acknowledge this is an issue and their recommendations should be taken with a grain of salt. How should one decide then if a copy is good or not? Exchanging 2-3 copies, while feasible and frequently cost-neutral, is certainly a bit of a pain. So, with all these in mind, is that another reason to go for a prime (in this case the DA 15mm Ltd) and forget about all the sample variation issues (presumably primes are immune to that)?
I'd love to hear thoughts and suggestions.

northcoastgreg · 11-30-2010, 11:05 AM

If the sample copies of the Sigma and the Tamron were not as sharp as you'd like, I wouldn't assume the problem is sample variation (it's unlikely that both would be bad copies, in any case, and if one of them were bad, it would significantly worse than the other). 3rd party lenses, however decent, are not likley to match the quality of a Pentax limited. As far as QC issues, it may not be merely a problem with zooms; it's often assumed that Sigma lenses (and perhaps Tamron as well, though I've heard more complaints about Sigma) do suffer from more QC issues than camera brand lenses. While it's true that Pentax had some problems with some of their SDM zooms a few years back, I haven't heard about any problems with the limiteds.

Another option, assuming you can afford it, is the DA 12-24. I can't imagine anyone complaining about it's sharpness at the wide end of the lens.

timo · 12-02-2010, 08:07 AM

Originally posted by tbirdas

Reading through the excellent threads in this forum, many users report significant sample variation for certain zooms. ...So, with all these in mind, is that another reason to go for a prime (in this case the DA 15mm Ltd) and forget about all the sample variation issues (presumably primes are immune to that)?

Doubt if primes are immune, but they are simpler in construction so it seems likely they are less prone to quirks in the assembly process.

Based on my experience of 3 x Sigma 17-70, 2 x Pentax 55-300, and 3 x 18-55, I think the answer to your question is yes, without a doubt.

Ron Kruger · 12-02-2010, 09:58 AM

Fewer moving parts, stronger construction, more emphasis on IQ, smoother MF...YES, while zooms are much better these days, primes (especially Limited) still offer many advantages for serious photographers.
The DA 15mm Limited, by the way, is the best lens I've used in about 4 decades of this, and that includes some Zeiss glass.

Todd Adamson · 12-02-2010, 01:42 PM

Primes are definitely not immune to sample variation; Lloyd Chambers even has a section on his Zeiss lens page devoted to sample variation. So I would say no to your question. For me, the best reason to shoot primes is IQ.

bobmaxja · 12-02-2010, 01:55 PM

I have the Tamron and that lense is very sharp

**Lowell Goudge** · 12-02-2010, 02:57 PM - **1 Like**

I personally think the biggest issue with sample variation is immediately behind the viewfinder.

While I don;t claim to have perfect luck, clearly I have not won the lottery yet, I have never had what I consider a bad copy of a lens. I have generally looked at the lenses I have bought, and the images they produce, with the methods I use the lenses under, and accepted the results for what they are.

Today, with computers and huge monitors we have the ability to look at images microscopically, and many of the Sample issues, aside from decentering which fortunately I have not had, are the result of technique not product.

To properly consider sharpness and image quality generally requires use of a tripod.

The old rule of thumb 1/focal length, for film or 1/focal length/1.5 for digital really only applies to looking at an image in 8 x 10 format, and not looking at it zoomed in to the pixel level, the definition of acceptable sharpness was based upon the same circles of confusion that Depth of field was based upon, although applied a little differently, and therefore does not apply to pixel peeping at what would equate to a 10 foot high print,

We need to get serious about this, if it was not shot with a tripod or flash, the person behind the viewfinder has much more to do with lens sharpness than the lens.

Todd Adamson · 12-02-2010, 03:23 PM

I agree with everything you say here, for sure. People are far too quick to claim "bad copy" without appropriate testing. But there's no doubt there are bad copies out there, of most lenses, and that certain lenses or manufacturers have more than others.

I have bought many Nikon lenses and never had a bad copy. But I know people who have definitely had some bad ones. On two occasions I've had a bad copy of a Tamron lens, and most recently I got a bum copy of the Zeiss 100MP. I proved that it was bad and it was replaced. If I suspect a bad copy, I do very throrough testing, with focus charts, on a good tripod, using a remote trigger and mirror lock-up.

Originally posted by Lowell Goudge

I personally think the biggest issue with sample variation is immediately behind the viewfinder.

While I don;t claim to have perfect luck, clearly I have not won the lottery yet, I have never had what I consider a bad copy of a lens. I have generally looked at the lenses I have bought, and the images they produce, with the methods I use the lenses under, and accepted the results for what they are.

Today, with computers and huge monitors we have the ability to look at images microscopically, and many of the Sample issues, aside from decentering which fortunately I have not had, are the result of technique not product.

To properly consider sharpness and image quality generally requires use of a tripod.

The old rule of thumb 1/focal length, for film or 1/focal length/1.5 for digital really only applies to looking at an image in 8 x 10 format, and not looking at it zoomed in to the pixel level, the definition of acceptable sharpness was based upon the same circles of confusion that Depth of field was based upon, although applied a little differently, and therefore does not apply to pixel peeping at what would equate to a 10 foot high print,

We need to get serious about this, if it was not shot with a tripod or flash, the person behind the viewfinder has much more to do with lens sharpness than the lens.

Ron Kruger · 12-02-2010, 06:49 PM

I think I once had a bad copy of a lens, but I was so drunk during that period, I can't remember.

RioRico · 12-05-2010, 05:47 PM

Originally posted by Lowell Goudge

The old rule of thumb 1/focal length, for film or 1/focal length/1.5 for digital really only applies to looking at an image in 8 x 10 format...

Just a note: The 1/FL ROT (rule of thumb) never really worked for 35mm film either. Ansel Adams wrote in THE CAMERA of shooting a 35mm cam with 50mm lens handheld, and did not get acceptably sharp photos at 1/50 sec. Not until he set the shutter to 1/250 sec were the images sharp. That is a 2+ stops difference. SR on my K20D might swamp-out that difference. But for handheld sharpness, a better ROT would be 1/(2*FL). Fire that 50mm lens at 1/100 sec, eh? And no coffee...

11-30-2010, 09:00 AM	#1
tbirdas Forum Member Join Date: Jun 2010 Location: Indianapolis Posts: 83	Is "sample variation" another reason to consider primes? Reading through the excellent threads in this forum, many users report significant sample variation for certain zooms. My most recent concern is the sigma 10-20 vs the Tamron 10-24. I tried them both (one copy of each) - no glaring decentering issues etc and the picture is quite comparable but not vary sharp (although the sigma may be a little sharper). But how do I know if these were good copies? Even "pro" reviews acknowledge this is an issue and their recommendations should be taken with a grain of salt. How should one decide then if a copy is good or not? Exchanging 2-3 copies, while feasible and frequently cost-neutral, is certainly a bit of a pain. So, with all these in mind, is that another reason to go for a prime (in this case the DA 15mm Ltd) and forget about all the sample variation issues (presumably primes are immune to that)? I'd love to hear thoughts and suggestions.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Any good reason to "upgrade" from K2000 to K-x?	wedge	Pentax DSLR Discussion	26	09-30-2010 11:55 AM
"normal" primes with A setting ?	Steve Beswick	Pentax SLR Lens Discussion	4	08-08-2009 07:35 AM
AF Microadjustment, for primes only? (OR more: "Better for primes?")	morfic	Pentax DSLR Discussion	4	10-02-2008 10:36 AM
Sample NON retouched from "In the Mood"	benjikan	Photographic Technique	14	01-28-2008 04:34 PM
DA 70mm f/2.4 "pancake" sample images	Gio645	Pentax SLR Lens Discussion	3	10-22-2006 11:37 PM

11-30-2010, 11:05 AM	#2
northcoastgreg Pentaxian Join Date: Feb 2010 Location: Northern Michigan Photos: Gallery \| Albums Posts: 6,173	If the sample copies of the Sigma and the Tamron were not as sharp as you'd like, I wouldn't assume the problem is sample variation (it's unlikely that both would be bad copies, in any case, and if one of them were bad, it would significantly worse than the other). 3rd party lenses, however decent, are not likley to match the quality of a Pentax limited. As far as QC issues, it may not be merely a problem with zooms; it's often assumed that Sigma lenses (and perhaps Tamron as well, though I've heard more complaints about Sigma) do suffer from more QC issues than camera brand lenses. While it's true that Pentax had some problems with some of their SDM zooms a few years back, I haven't heard about any problems with the limiteds. Another option, assuming you can afford it, is the DA 12-24. I can't imagine anyone complaining about it's sharpness at the wide end of the lens.

12-02-2010, 08:07 AM	#3
timo Veteran Member Join Date: Nov 2006 Location: former Arsenal football stadium Photos: Gallery Posts: 431	Originally posted by tbirdas Reading through the excellent threads in this forum, many users report significant sample variation for certain zooms. ...So, with all these in mind, is that another reason to go for a prime (in this case the DA 15mm Ltd) and forget about all the sample variation issues (presumably primes are immune to that)? Doubt if primes are immune, but they are simpler in construction so it seems likely they are less prone to quirks in the assembly process. Based on my experience of 3 x Sigma 17-70, 2 x Pentax 55-300, and 3 x 18-55, I think the answer to your question is yes, without a doubt.

12-02-2010, 09:58 AM	#4
Ron Kruger Veteran Member Join Date: Aug 2010 Location: Missouri Posts: 534	Fewer moving parts, stronger construction, more emphasis on IQ, smoother MF...YES, while zooms are much better these days, primes (especially Limited) still offer many advantages for serious photographers. The DA 15mm Limited, by the way, is the best lens I've used in about 4 decades of this, and that includes some Zeiss glass.

12-02-2010, 01:42 PM	#5
Todd Adamson Veteran Member Join Date: Oct 2010 Location: Iowa Posts: 722	Primes are definitely not immune to sample variation; Lloyd Chambers even has a section on his Zeiss lens page devoted to sample variation. So I would say no to your question. For me, the best reason to shoot primes is IQ.

12-02-2010, 01:55 PM	#6
bobmaxja Veteran Member Join Date: Mar 2010 Location: Laval, Quebec Canada Posts: 2,171	I have the Tamron and that lense is very sharp

12-02-2010, 02:57 PM - 1 Like	#7
Lowell Goudge Site Supporter Join Date: Jan 2007 Location: Toronto Photos: Gallery \| Albums Posts: 17,888	I personally think the biggest issue with sample variation is immediately behind the viewfinder. While I don;t claim to have perfect luck, clearly I have not won the lottery yet, I have never had what I consider a bad copy of a lens. I have generally looked at the lenses I have bought, and the images they produce, with the methods I use the lenses under, and accepted the results for what they are. Today, with computers and huge monitors we have the ability to look at images microscopically, and many of the Sample issues, aside from decentering which fortunately I have not had, are the result of technique not product. To properly consider sharpness and image quality generally requires use of a tripod. The old rule of thumb 1/focal length, for film or 1/focal length/1.5 for digital really only applies to looking at an image in 8 x 10 format, and not looking at it zoomed in to the pixel level, the definition of acceptable sharpness was based upon the same circles of confusion that Depth of field was based upon, although applied a little differently, and therefore does not apply to pixel peeping at what would equate to a 10 foot high print, We need to get serious about this, if it was not shot with a tripod or flash, the person behind the viewfinder has much more to do with lens sharpness than the lens.

12-02-2010, 03:23 PM	#8
Todd Adamson Veteran Member Join Date: Oct 2010 Location: Iowa Posts: 722	I agree with everything you say here, for sure. People are far too quick to claim "bad copy" without appropriate testing. But there's no doubt there are bad copies out there, of most lenses, and that certain lenses or manufacturers have more than others. I have bought many Nikon lenses and never had a bad copy. But I know people who have definitely had some bad ones. On two occasions I've had a bad copy of a Tamron lens, and most recently I got a bum copy of the Zeiss 100MP. I proved that it was bad and it was replaced. If I suspect a bad copy, I do very throrough testing, with focus charts, on a good tripod, using a remote trigger and mirror lock-up. Originally posted by Lowell Goudge I personally think the biggest issue with sample variation is immediately behind the viewfinder. While I don;t claim to have perfect luck, clearly I have not won the lottery yet, I have never had what I consider a bad copy of a lens. I have generally looked at the lenses I have bought, and the images they produce, with the methods I use the lenses under, and accepted the results for what they are. Today, with computers and huge monitors we have the ability to look at images microscopically, and many of the Sample issues, aside from decentering which fortunately I have not had, are the result of technique not product. To properly consider sharpness and image quality generally requires use of a tripod. The old rule of thumb 1/focal length, for film or 1/focal length/1.5 for digital really only applies to looking at an image in 8 x 10 format, and not looking at it zoomed in to the pixel level, the definition of acceptable sharpness was based upon the same circles of confusion that Depth of field was based upon, although applied a little differently, and therefore does not apply to pixel peeping at what would equate to a 10 foot high print, We need to get serious about this, if it was not shot with a tripod or flash, the person behind the viewfinder has much more to do with lens sharpness than the lens.

12-02-2010, 06:49 PM	#9
Ron Kruger Veteran Member Join Date: Aug 2010 Location: Missouri Posts: 534	I think I once had a bad copy of a lens, but I was so drunk during that period, I can't remember.

12-05-2010, 05:47 PM	#10
RioRico Veteran Member Join Date: Jun 2008 Location: Limbo, California Posts: 11,263	Originally posted by Lowell Goudge The old rule of thumb 1/focal length, for film or 1/focal length/1.5 for digital really only applies to looking at an image in 8 x 10 format... Just a note: The 1/FL ROT (rule of thumb) never really worked for 35mm film either. Ansel Adams wrote in THE CAMERA of shooting a 35mm cam with 50mm lens handheld, and did not get acceptably sharp photos at 1/50 sec. Not until he set the shutter to 1/250 sec were the images sharp. That is a 2+ stops difference. SR on my K20D might swamp-out that difference. But for handheld sharpness, a better ROT would be 1/(2*FL). Fire that 50mm lens at 1/100 sec, eh? And no coffee...