rrmerkov

I was about to post my second post to the photo gallery when I received the error message that my file was too large. The photo before processing with Topaz DN was 15 meg. After processing the file was 25 meg. So, is Topaz Denoise doing something in addition to reducing noise?

fredpnm Loc: Corrales, NM

Depends, if you are comparing a submitted RAW file (or JPG) size to the resulting DN .tif file it can be a larger file.

You might get a better insight into your question if you told us something about the file format of the submitted photo and what you told DN the file type to save as the finished image.

AndyGarcia

It happens with all my photos I run through Topaz Denoise Ai. It does not bother me. Why don't you ask Topaz. I've always found their tech support people very helpful.

rrmerkov

Thanks for your reply. The format of the file that I submitted was JPG and the format of the file tpye that DN saved was also JPG.

rrmerkov

Thanks for your reply. I will try asking Topaz' support.

fredpnm Loc: Corrales, NM

Thanks for the information. I opened a .jpg file, ran standard noise reduction on it, and then saved it as a .jpg. The initial file was 1.41 MB and the saved file was 2.07 MB. So there is an increase in file size - not great but still, as you pointed out, a larger file. It will be interesting to hear the reason for this from Topaz.

kymarto Loc: Portland OR and Milan Italy

If you sharpen you have fewer similar pixels that can be grouped together in the compression algorithm. Very detailed jpgs are much larger than those with less detail.

kymarto Loc: Portland OR and Milan Italy

There are various levels of jpg compression. When a highly compressed jpg is opened and then saved, it may be saved with lower compression and therefore will be larger in size even if is of the same detail.

So even if you process a jpg and remove detail, and then save it as a higher quality (lower compression) jpg it will be larger than the original.

Some programs like Photoshop have different levels of jpg compression, from highly compressed/low quality to less compressed/higher quality. The lowest quality setting can yield an image a small fraction the size of the highest quality setting.

If you take a medium quality jpg and resave it at maximum compression/lowest quality, and then reopen and resave at the highest quality, it will be many times the size of the original but even lower quality than the low quality image--because it is lossy compression, always decreasing quality no matter how high the setting, and what is lost can never be regained, even if you save later at a higher quality.

In fact, every time a jpg is simply reopened and resaved, the quality gets lower no matter how high the quality setting. This is why I only ever save in jpg after all editing is done.

CHG_CANON Loc: the Windy City

What is the purpose of a 15MB attachment? Consider the logic and methods of resizing all your online posts to 2048-pixels on the long side of the image based on these techniques:

Recommended resizing parameters for digital images

rrmerkov

Thanks for the info and the link. I was not aware of the recommended resizing.

srsincary Loc: Cary, NC

My output .tif file produced by Topaz DeNoise AI is bigger than the input .tif file. I noticed that yesterday. I have version 3.2.0. This might have happened with older versions too - but yesterday I was doing bulk processing (usually I just call the PS plugin) and so the output files had different names from the inputs, allowing me to compare sizes.

gwilliams6

Folks whether you process your shots in Photoshop, Lightroom, Capture One, Luminar 4, or any Topaz software, once you work on the files in post you WILL add image size as you have changed and added data to the pixels with your adjustments. There is no free lunch.

That is perfectly normal and unavoidable unless you remove elements in the photo. You can still resize any image to be what size you prefer during output.

Cheers

OhD Loc: West Richland, WA

It may be that DN converts multi-pixel, monochromatic blobs of noise into the same number of pixels with different colors to create a gradient between the boundaries of the blob. This would not be as readily compressed as a blob and would increase the file size accordingly. You might inspect some of your images - maybe crop a noisy picture down to a few hundred pixels, then apply DN to a copy of the crop and compare them both for file size and visually to test my theory. Comparing histograms of the two samples might be informative.