Wed 01-Feb-12 12:47 AM | edited Wed 01-Feb-12 12:48 AM by hujiie
I have had D700 so long and now I own two bodies. I use them for commercial use as well as my design works. When I got back one of the unit form service from Nikon, I was synchronizing the settings of two bodies and realized that I had Raw setting as lossless compressions instead of uncompressed. For me IQ is the key because often I enlarge the images more than it usually capable of.
Can anyone tell me difference in IQ b/w lossless compression and uncompressed?
By the way, I use 14 bits and after that, switched "uncompressed" RAW for the best.
Try as I did, when comparing NEF lossless compressed and NEF uncompressed I couldn't see any difference in the photos, even shadow and highlight areas where I would expect differences, if any, to show up the most.
On the other hand, though it was nitpicking to be sure, I did see differences in shadow areas when using 12bit vs. 14bit, with the 14bit photos having slightly better color rendition.
There is no difference between uncompressed and lossless compressed images. The latter is uncompressed by the RAW converter (i.e. NX2 or ACR) into the exact same file as the uncompressed file would have been.
Thank you for the reply. It seems to me no differences... As I think about it, when I print images with large format printers, I do not see difference b/w Tiff LZW compression and uncompressed TIFF files.
I guess as Ned implies, I have to test them in my workflow (enlargement upto 54" of the shortest ends of the photos) to see if they are the same.
Lossless IS lossless. The only difference - uncompressed takes more space on the card, but less time for recording while shooting, and lossless on the contrary: more time, but less space. Picture quality is NOT affected.
The reason the compression is called lossless is that there are data compression schemes that lose data when compressed. Early Nikon cameras dropped some information that was not discernible by the human eye and then there were some methods used in the Jpeg format that lost more and more data upon each compression.
> The only difference - uncompressed >takes more space on the card, but less time for recording >while shooting, and lossless on the contrary: more time, but >less space.
Note that this generally may not always hold true. The added time for the processor to compress the file can often be more than offset by the time savings of writing a smaller file. It's a balance between compression speed and write speed. As such, it can be camera/card dependent. I seem to recall there was a thread where someone did some testing but I can't seem to find it. In any event, if it is important, it is worth testing with your camera/card combination.
Just want to reiterate that in terms of IQ, which you are most concerned about, the files are identical. In fact, I wish Nikon would provide an option under ViewNX to uncompress the file upon import before saving to disk. (In which case it would be identical to a file taken as uncompressed.) I would prefer to store my backups as uncompressed.
To compress or not to compress for frame rate is a different issue. It is also complicated by the fact that there is a fast memory buffer which is between the compression and the flash card. By compressing, you get a few more shots before the buffer fills up and the following write to the card will be faster. Keep in mind that sports photographers often shoot jpg to maximize the shots before the buffer fills. That suggests to me that the camera processor is not the limiting factor. I would guess (and just a wild guess) that the conversion from Sensor(RAW) to JPG would be about the same as a RAW compression but it just depends on how they optimized the processor chip. Again, if it's important, give it a test. I haven't yet been motivated to fire off all the shutter clicks to find out.
I think my concern is that there is compression setting, even though it says lossless. In order to compress, the photo HAS TO sacrifice something to make the date small. IF not, why do we need the "compressed file".
Also the judgement depends on regular photography output workflow vs. unconventional ones (in my case, extreme enlargement).
After all, as I mentioned again, I need to test to find out. If I find the difference in MY WORKFLOW based on my needs, I will let you know.
Thank you for all. I thought this was much simple question.
>In order to compress, the photo HAS >TO sacrifice something to make the date small. IF not, why do >we need the "compressed file".
There need not be any loss. For example, consider the data . I could write which is smaller, but if you know how it is encoded you can go back and forth without any loss. Obviously an extremely simple example. Photographic data is usually very redundant (think blue sky). As such it can be encoded in a way that is much smaller than a brute force approach of recording a value for each sensor pixel. Of course, the encoding step takes processing power to encode the data rather than just writing what comes off the sensor. When your desktop software (NX2, Lightroom etc.) loads the file, it just undoes the encoding process and gets back brute force sensor data.
Now I guess there is the question of whether or not Adobe really, really knows Nikon's encoding scheme, but the same could be said for the uncompressed file.
I don't have the data to show it, but would expect that lossless compressed might consume more shoot to write time than uncompressed, as the in-camera processor has to execute the compression algorithm. Offsetting this would be the time required to write a greater volume of data in the uncompressed format. Nikon surely made this a choice for a reason.
Fri 03-Feb-12 07:35 AM | edited Fri 03-Feb-12 03:19 PM by briantilley
>I think my concern is that there is compression setting, even >though it says lossless. In order to compress, the photo HAS >TO sacrifice something to make the date small.
Lossless compression is just that - one (or more) complex algorithms are applied to the RAW data and are designed in such a way that the bits and bytes removed when compressing can be exactly re-constituted when uncompressing.
>After all, as I mentioned again, I need to test to find out.
No, you don't. This isn't workflow-dependent. After uncompression, the data is truly identical to the original - you really don't need to worry about this!
>Lossless compression is just that - one (or more) complex >algorithms are applied to the RAW data and are designed in >such a way that the bits and bytes removed when compressing >can be exactly re-constituted when uncompressing. >
As others have pointed out, there are two methods employed by Nikon for NEF compression. Originally there was the visually lossless, and now the option of lossless. With the visually lossless compression there is a two step compression, the first being lossey and second lossless. With lossless compression the first step is skipped leaving the second which is lossless.
Nikonian’s Bill Claff put together a rather nicedocument that describes this.
I guess the short answer would be because customers want it.
Expanding ... RAW uncompressed is sort of the accepted gold standard. If it didn't exist for people to test against there would always be doubts about whether or not the compressed version really was lossless. As an aside, this has a history. With earlier cameras (D100 time frame) Nikon seemed to say the compression was lossless but later said it was "visually lossless" which now corresponds to the NEF Compressed. Also, for some processor/memory configurations Uncompressed RAW may be faster. I've never taken a stopwatch to my D700 but haven't noticed a difference in burst rate (writing to the buffer) between compressed and uncompressed but other cameras with different processors could have different results. And don't forget marketing. When some people buy a 36MP camera, they are gonna want to see that 75Mbyte file!
What everyone is telling you here is definitely correct; I'm not clear why you are worrying about it It applies to any Nikon which offers the "Lossless Compressed NEF" option, not just the D700.
As was already explained, using Lossless Compressed can slow down frame rate slightly, due to the extra processing time required. Uncompressed is generally a little faster and is offered for that reason.