Jump to content

To compress or not to compress


bouic49

Recommended Posts

Advertisement (gone after registration)

But what if you use Lightrooms compression on an already compressed (by the M9) DNG?

In this case you get files with about 8-10MB.

 

Does anybody now for sure, that these files in the end are not worse than the uncompressed DNGs (if we originate that the loss of dynamic range from the camera compression is negligible)?

 

If so, that would mean we can compress the 36MB DNG's to 8-10MB without significant loss in image quality.

 

 

A real-world picture with much detail from a CD rack shot in both compressed and uncompressed in the camera and both stored lossless from LightRoom gave 11 Mbyte and 22.7 Mbyte sized files respectively. So yes, that saves more room on the hard disk.

 

As I said before, it is very hard to see the difference between compressed and uncompressed, but it can be measured in terms of digitization noise. It's up to you if that tiny difference is worth the extra space on the hard disk, on the SD card and the writing times to the SD in camera and hard disk in computer.

 

Other factors are more important, like the choice of ISO: even 1/3 of a stop upward along the ISO scale completely cancels the effect of shooting uncompressed. :D

Link to post
Share on other sites

  • Replies 50
  • Created
  • Last Reply

Well, yes, although the difference is more clear if one (over)processes the image strongly. But it may well be that future image processing will show the losses up more clearly than presently. With memory space relatively cheap it seems to me that compressing might turn out to be an unfortunate economy.

Link to post
Share on other sites

... Other factors are more important, like the choice of ISO ...

According to Stefan Daniel's Reichmann interview, if I understood correctly, the choice of ISO also affects the M9's rate of writing images to card, or at least did at that time.

 

I wonder if anyone has done any experimentation with that aspect?

Link to post
Share on other sites

Well, yes, although the difference is more clear if one (over)processes the image strongly.

 

It is very instructive to try and make an example of 2 images, compressed and uncompressed of a surface with a fluent grade of highlights, expand 20 steps (bins) of this highlight area over the full 256 histogram step (while in 16 bit mode in PhotoShop) and eyeball the images close by (on top of each other in two layers). More than 10 fold expansion of a highlight area is certainly overprocessing. Even then image noise at 160 ISO is dominating the digitization noise and it is hard to say which is which.

 

(Still as I said, I shoot uncompressed... :) )

Link to post
Share on other sites

According to Stefan Daniel's Reichmann interview, if I understood correctly, the choice of ISO also affects the M9's rate of writing images to card, or at least did at that time.

 

I wonder if anyone has done any experimentation with that aspect?

 

No, there is no effect of ISO setting on write times to SD card.

 

Tested 7 exposures in continuous mode at 1/60 sec. exposure time each, with an Ultra II 16 Gbyte card including writing to card (times measured from pressing the shutter release button until writing led stopped signalling):

 

160 ISO uncompressed 34.7 seconds

 

1250 ISO uncompressed 35.2 seconds

 

160 ISO compressed 19.7 seconds

 

1250 ISO compressed 20.2 seconds

Link to post
Share on other sites

Wow, that is a big difference between compressed vs. uncompressed write times. That explains why the buffer seems to fill faster when shooting uncompressed. I am shooting everything compressed as hitting the buffer limit is too frustrating.

Link to post
Share on other sites

Given the easy availability of true lossless compression algorithm's ( like zip ), it is astonishing to me that we are still having this discussion. Leica should simply use one of them.

Surprisingly Leica had actually thought of this, even back in 2006. Unfortunately the processor wasn’t up to the task. The lossy compression scheme they came up with instead was extremely fast and has stood the test of time.

 

The M10 will probably sport a faster CPU and offer lossless compression. Until then it is either lossy compression or no compression at all.

Link to post
Share on other sites

I'd appreciate some help here.

 

Both Nikon and Canon offer cameras that shoot 3 sizes of RAW file, 4/10/16 and record any two sizes in their dual card bodies. Does anyone suggesting that shooting large RAW files on them is also wasted, or is the compression formula unique to Leica.

 

I expect Nikon and Canon to have good reason to produce large RAW files, presumably because it makes a difference to the end product. No ? Do any users of Nikon and Canon cameras here shoot medium RAW thinking along the lines of this thread ?

 

Bruce Frazier, the RAW file expert promoted the use of 16 bit RAW. Is his logic now superseded by Leica's look-up tables. Seems to me that Adobe have pushed their products pretty hard to bring us 16 bit manipulation and many of have spent small fortunes on computers that will handle the bigger files.

 

I'll admit to being sceptical of the Compressed file being as good as the Uncompressed file and it seems illogical, but it would be really useful to know whether Canon, Nikon, Apple and Microsoft are wasting their efforts on 'our' behalf.

 

M

Link to post
Share on other sites

Nikon offers a lossy compression scheme that is somewhat similar to Leica’s, only the lossy compression step doesn’t compress quite as much and there is a second, lossless compression step (this was introduced in 2004 with the D70, I believe). Alternatively you can choose a completely lossless compression, though not in all models.

 

Bruce Frazier' date=' the RAW file expert promoted the use of 16 bit RAW. Is his logic now superseded by Leica's look-up tables. Seems to me that Adobe have pushed their products pretty hard to bring us 16 bit manipulation and many of have spent small fortunes on computers that will handle the bigger files.[/quote']

The colour depth used for image editing is one thing, the bit depth of raw files another. You are well advised to do all your image editing within a 16 bit RGB or L*a*b space, whether the original raw data are 14 bit, 12 bit, or 14 bit compressed to 8 bit. No image processing whatsoever is done with Leica’s compressed 8 bit data; in the raw converter it is expanded to 14 bit first.

Link to post
Share on other sites

In regard to the lossy NEF compression, see Jeffrey Friedl's Blog , an interesting comparison which shows a demonstrable but not a visible loss of data.

 

In regard to Nikon's, Canon's and Leica's decisions to offer an uncompressed file as well, the quickest reason to hand is simply that "people want it." It's quicker than explaining the same thing over and over. :)

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...