Photo Editing

[CSULB71]

Since video and photo enhancement techniques are similar, I hope this question will get some responses here. What I'm doing is editing/enhancing photos for personal use, mainly to view the photos on a computer monitor.

In trying to avoid the quality loss that happens when a JPEG file is repeatedly saved due to compression of the data each time, I've been starting the editing process with a Bitmap (.BMP) version of the photo because the file size doesn't change each time you save the file, like a JPEG file does. I know starting with a RAW or TIFF file is best for this, but I don't want to work with those huge files.

So, does using a BMP file avoid/minimize loss of quality because BMP's don't recompress each time they're saved?

[hech54]

In trying to avoid the quality loss that happens when a JPEG file is repeatedly saved due to compression of the data each time

I've heard this nonsense SO often recently it's pathetic. Where are you getting this crap from? People need to learn the difference between "save" and "save as".

[poisondeathray]

So, does using a BMP file avoid/minimize loss of quality because BMP's don't recompress each time they're saved?

another option is lossless compression , like png

[lordsmurf]

In trying to avoid the quality loss that happens when a JPEG file is repeatedly saved due to compression of the data each time

I've heard this nonsense SO often recently it's pathetic. Where are you getting this crap from? People need to learn the difference between "save" and "save as".

There is some truth to it. But it's not the JPEG format -- it's crap software.
Photoshop users rarely complain about this. It's always people who user freeware or cheapware. That's simply because the JPEG format is really horrible from that program. We see the same thing with MPEG-2 and MPEG-4. There are enough variables within these photo and video formats that you can screw it up easily.

Save as max level-10 compression, and it will be fine.
If you're still really concerned, then use a TIF.

[CSULB71]

People need to learn the difference between "save" and "save as".

I know the difference -- I use "save as" to a different file name all the time during the editing iterations process. I was just asking if there was any truth to what I'd read/heard years ago (can't remember where) that using BMP's was better than JPEG's because of no/little compression loss. I'll try using PNG on a few photos to see if there is a noticeable difference vs. BMP. Thanks for the suggestions.

[Steve Stepoway]

In trying to avoid the quality loss that happens when a JPEG file is repeatedly saved due to compression of the data each time

I've heard this nonsense SO often recently it's pathetic. Where are you getting this crap from? People need to learn the difference between "save" and "save as".

There is some truth to it. But it's not the JPEG format -- it's crap software.
Photoshop users rarely complain about this. It's always people who user freeware or cheapware. That's simply because the JPEG format is really horrible from that program. We see the same thing with MPEG-2 and MPEG-4. There are enough variables within these photo and video formats that you can screw it up easily.

Save as max level-10 compression, and it will be fine.
If you're still really concerned, then use a TIF.

Some people who think they know the difference between "Save" and "Save As" obvioulsly don't. JPEG is inherently a lossy compression. It doesn't make any difference whether you are using Photoshop or GIMP or anything else. JPEG can never be lossless. If you open a JPEG image to edit, the image is decompressed. When you save it, whether under the same name with the same compression parameters ("Save") or under a different name with different parameters ("Save As") the edited image has to be recompressed. That is a lossy process, no matter how high you crank the "quality" parameter. A JPEG quality of 100 is not lessless. Read the specs and you will see that there are still round-off errors. Repeated cycles will degrade the image. Maybe not enough to notice at first, but eventually you will. Most people are not going to go through enough cycles for this to matter, but it does happen. True, if you crank the quality factor to over 90, you are unlikely to notice the difference, even with several cycles. But it remains a lossy compression. If you are really concerned, use a TIF (as LS suggested) or a PNG or an uncompressed BMP.

[CSULB71]

Steve Stepoway:
So you're saying using a BMP (uncompressed?) is better than doing/saving the iterations as JPEG's? That's what I thought, and why I asked the initial question. But, of course, not as good as using PNG or TIF.

[Steve Stepoway]

Steve Stepoway:
So you're saying using a BMP (uncompressed?) is better than doing/saving the iterations as JPEG's? That's what I thought, and why I asked the initial question. But, of course, not as good as using PNG or TIF.

I don't have much experience using BMPs; I mostly use PNGs, or native XCF format if I'm using GIMP. I did run into some compressed BMPs one time, and that experience wasn't good. Exactly what the problem was, I didn't bother to figure out. I just knew that they didn't look good, with lots of compression and/or quantization artifacts. When working with JPEGs, I keep the quality factor to 90+, and over 2 or 3 cycles I've not noticed any problems. But I know they're going to show up at some point, so why worry? Temporary disk space is cheap, so I figure "keep it uncompressed as long as you can.".

[lordsmurf]

It doesn't make any difference whether you are using Photoshop or GIMP or anything else.

Yes, it does. Quality between software is highly variable.

there are still round-off errors

I never suggested otherwise. But the loss at max quality is minimal, if any.

Temporary disk space is cheap

Not really. Most photographers I know operate, at minimum, on a multi-TB RAID-1 setup. More often, it involves several RAID mirrors, with redundant backups. We don't take any chances. A 2x2TB RAID-1 is $300+, not exactly "cheap".

...............

I'm continuously amused at the people who think photos must be saved in some kind of super-uncompressed format. That's not how the industry runs (newspapers, magazines, PR, marketing, etc). Zillions of JPEGs are emailed around daily. If you're looking to be a studio print artist with large canvas prints -- sure, save uncompressed. But you'll be working with DNG, NEF, CR2, etc. Never BMP, and rarely even TIF.

There are valid reasons to use each of them. If you plan to open/edit x infinity, sure, save as TIF.
But the assumption that "JPEG is bad" is, as hech54 first said, a really dumb myth.

[CSULB71]

I've been gone awhile working with my scanned slides. I'm completely happy with the quality level that's been maintained by using the BMP format during the many "enhancement" iterations. Now I'm at the point of sharpening many of the pics, and then converting them all to final JPEG's.

Question: for best final image quality, should I sharpen the pics as BMP then convert to JPEG, or convert to JPEG first and then sharpen? I've tried it both ways, but I can't tell the difference (old eyes aren't the best test ). I'm inclined to think that if I sharpen them as BMP's, the compression to JPEG would loose some of the sharpness?

[poisondeathray]

I've been gone awhile working with my scanned slides. I'm completely happy with the quality level that's been maintained by using the BMP format during the many "enhancement" iterations. Now I'm at the point of sharpening many of the pics, and then converting them all to final JPEG's.

Question: for best final image quality, should I sharpen the pics as BMP then convert to JPEG, or convert to JPEG first and then sharpen? I've tried it both ways, but I can't tell the difference (old eyes aren't the best test ). I'm inclined to think that if I sharpen them as BMP's, the compression to JPEG would loose some of the sharpness?

Sharpen as BMP. If you do it as jpeg, you incur 1 round of generation loss (jpeg to jpeg, assuming your final format goal is jpeg)

If you did all your manipulations in 1 go, you wouldn't need all these stages or BMP

BMP is uncompressed. You can use lossless compression as well to save file space (e.g. png) .

If you're familiar with audio, the analogy would be FLAC for png. JPEG is lossy so the analogy would be MP3. BMP is uncompressed so the equivalent would be WAV audio

[Cornucopia]

The way I see it is you use what is appropriate for the task.

The 1st task is the taking of the picture with the highest quality available in the situation (your choice here). These days, this usually means Camera Raw Files (best quality) or TIF (nearly best, but not usually high-density or logarithmic, so a little limited) or JPEG (great for compatibility & size but so-so for quality of master and limited color/DR palette).

Then, the next (2nd) task is editing. You may be doing it alone or working in concert with other people/agencies. Alone, pick what you want as long as it maintains quality. If others are involved, you want something that is interoperable. Interoperabiity include both Universality and Consistency. I would break this up into items which have LAYERS and those that don't. For Layers, PSD/PSB, EXR, and TIF are common (TIF is most common, but not always consistent). For non-layers, PNG, TIF, BMP, JPEG. (BMP is probably most Universal because of its history with MS PC's, but isn't always consistent and has no real ICC support; JPEG - even with possible compression & palette limitations, DOES/CAN have real ICC support; PNG also, though it isn't nearly as universal).

Then there's the 3rd task of distribution to end users. This is really where JPEG shines.

I say, to maintain the best quality always make as few changes as possible - hopefully all together - and do your compression at/near the end, when you're ready to distribute.

Compared to HD/2k & Stereoscopic Video files , a couple of TB overall for Photo Hard drive space IS cheap, and that's a working Pro setup, not a hobbyist's. So I'm in the uncompressed/losslessly compressed camp when it comes to internal storage of intermediate files.

Scott

[CSULB71]

Actually, I have already finished many pics by sharpening as BMP, then converting to final JPEG -- I just thought I 'd ask in case doing it the other way was better.

poisondeathray quote: "If you did all your manipulations in 1 go, you wouldn't need all these stages or BMP".
Oh, if only I was that good! It takes many iterations to get the color and exposure to my liking (most of the slides are 30 - 40 years old, and have undergone color change/deterioration, plus the scanning process introduces a bit of "off color" of its own), and I prefer to save every so often in case I want to revert to an earlier version. That's why I asked my original posted question about minimizing quality loss when saving multiple times.

cornucopia: This is all for my personal/family use, so I scanned the slides and saved as BMP, did my editing/enhancements with Photoshop Elements (as BMP), then convert to final JPEG.

[Cornucopia]

What I and I think poisondeathray were suggesting was that:
If you have Photoshop, you can do alot of your color adjustments through "Adjustment Layers" instead of using explicit, permanent actions.
This has the benefit of being post- re-editable, and multiple ADJ layers can be applied and/or temporarily turned off for comparison, etc.
What it does is not BAKE it into the original, and only until the final export to your distribution copy is it baked into the copy.
AND
It's applied all at once, so there's only 1 "generation loss". VERY handy.
Of course, this does require that you save in PSD/PSB format for the intermediate.
Think of it as a "Modification Script" that you keep previewing and editing until you're ready, and only then do you RUN the script, once.

BMP is ok. Like I said, YES, it's universal, but it isn't consistent with how it gets displayed, because of lack of ICC support. If I were going to do it lossless (and not with layers), I'd stick with TIF or PNG. But if it works for you - GREAT!

Scott

[budwzr]

Photographers / Videographers, and art people in general, are notorious for passing around a lot of baloney because they want to exclude the unwashed masses from their craft, and the best way to do this is via obfuscation.

There are some people around that will tell you vinyl records sound better than a CD. These people don't understand digital, and hold on to a nostalgic view in spite of the facts.

So I'm going to tell you the truth, that .PNG is the best format to be in for still images.

[aedipuss]

There are some people around that will tell you vinyl records sound better than a CD. These people don't understand digital, and hold on to a nostalgic view in spite of the facts.

heh? you are too young or have never experienced true hifi equipment. you don't seem to understand how sound waves work. they are analog. digitally sampling music at intervals leaves some pieces out. cd quality is not nearly as complete a representation of live music as vinyl, but you need high quality equipment to hear the difference. an ipod and earbuds really suck sonically.

[Cornucopia]

Photographers / Videographers, and art people in general, are notorious for passing around a lot of baloney because they want to exclude the unwashed masses from their craft, and the best way to do this is via obfuscation.

There are some people around that will tell you vinyl records sound better than a CD. These people don't understand digital, and hold on to a nostalgic view in spite of the facts.

So I'm going to tell you the truth, that .PNG is the best format to be in for still images.

You're joking, right?

PNG is GOOD for 24bit images or 32bit images, when using most standard, modern software (which supports it). It ain't good for low-bit or high-bit images, not for certain browsers, not for layers, and not for older or non-standard software. IOW, it might OFTEN NOT be the best format.

heh? you are too young or have never experienced true hifi equipment. you don't seem to understand how sound waves work. they are analog. digitally sampling music at intervals leaves some pieces out. cd quality is not nearly as complete a representation of live music as vinyl, but you need high quality equipment to hear the difference. an ipod and earbuds really suck sonically.

You're joking, right?

Yes, Digital leaves out stuff, but CD quality only leaves out stuff that WE CAN'T HEAR, same goes for HIGHEST BITRATE MP3/AAC, etc.
Vinyl is analog, and has all the other FUN analog stuff, like HISS+other noise, usage WEAR, Distortion, Highly modified/colored EQ. Well produced, high bitrate digital doesn't have that (at least not within the limits of human hearing).
iPods are quite NEUTRAL as far as their playback capabilities go. Yes, they don't have Class A amplifiers and yes, there might be some EMI+RFI leaking in, but otherwise, their response is pretty FLAT (aka CLEAN and UNOBTRUSIVE).
Headphones are headphones - they're ALL analog. Pick whichever sounds the best to you.

I bet it would be quite easy to set up an A/B/X test where you sit in a room listening to headphones, and you don't know whether they're connected to live microphones in another room, or to those same microphones recorded to digital, then converted to high-bitrate AAC, then played back through an iPOD to those same headphones. Very likely even you, aedipuss wouldn't be able to tell the difference.

Now, where were we with the Original topic?

Scott

[budwzr]

So cutting a master and stamping out vinyl discs is lossless?

[lordsmurf]

You never hear this kind of nonsense from within the music community.
Or in the photo community.
I know Cornucopia is in touch with these fields at that level -- as am I -- and I agree with him.

@CSULB71
Be careful sharpening. Most people tend to overdo it.
Keep an "unsharp" master, and only sharpen prior to final print/export version.

[bendixG15]

I just "envy" people with golden ears and platinum eyes whe hear and see what no others can.

They have my sympathy .............................

[Cornucopia]

Nope. In the analog world, WEAR=LOSS.

Scott

[aedipuss]

I bet it would be quite easy to set up an A/B/X test where you sit in a room listening to headphones, and you don't know whether they're connected to live microphones in another room, or to those same microphones recorded to digital, then converted to high-bitrate AAC, then played back through an iPOD to those same headphones. Very likely even you, aedipuss wouldn't be able to tell the difference.

not likely. i can hear the transistors in cheap consumer amps switching on and off. it's not a joke i've worked in the audio field in the booth.

unfortunately cd specs were a bad compromise and ended up at 41.1khtz, a mediocre sampling rate at best. and mp3 conversion even at 320bps is a laughable imitation of live music

go ahead and but laugh they're my ears and i like them.

[Cornucopia]

Yeah, I've been "in the booth" for >20years now. And according to my most recent ear check, still have ~17k-19k hearing (at 50! because I'm careful with SPL).

MPEG chose 320 precisely because FOR MOST PEOPLE, it was indistiguishable from LPCM. They did trials and everything. I guess you must have those "golden ears" then.
44.1k was for some time a compromise, which compromise has long since been made inconsequential with cost-efficient phase-linear digital FIR filters and vastly decreased Jitter from modern components. If it's PL below the Nyquist frequency, 22.05k vs. 24k isn't really a BIG difference.
What CAN be a difference is doubling that frequency. You may not be able to hear it, but you can sometimes "sense" it, particularly in Ambience/Reverb and with Binaural/Phase-Surround encoding.
Of course, Analog LPs roll off well below that, unless you were to find a rare CD-4 or "Half-speed" Mastered pressing and used a Shibata cartridge. And even then, it will only last for a few plays before it's worn down and lost.

I like my ears too. So, too, do my customers (lately, classical/choral/organ stuff).

Scott

[budwzr]

Photographers / Videographers, and art people in general, are notorious for passing around a lot of baloney because they want to exclude the unwashed masses from their craft, and the best way to do this is via obfuscation.

There are some people around that will tell you vinyl records sound better than a CD. These people don't understand digital, and hold on to a nostalgic view in spite of the facts.

So I'm going to tell you the truth, that .PNG is the best format to be in for still images.

You're joking, right?

PNG is GOOD for 24bit images or 32bit images, when using most standard, modern software (which supports it). It ain't good for low-bit or high-bit images, not for certain browsers, not for layers, and not for older or non-standard software. IOW, it might OFTEN NOT be the best format.

Of course I'm talking high quality, as in for editing, printing, incorporating into video, stuff like that.

If old or crappy software doesn't support it, tough tittie.

[CSULB71]

Be careful sharpening. Most people tend to overdo it.
Keep an "unsharp" master, and only sharpen prior to final print/export version.

True -- it's easy to go too far. Using PhotoShop Elements 7, I've settled at 30% sharpness for the average scanned slide image, and often follow that up with 10% "unsharp mask" as that looks a little better to me. But, is there really a benefit to even doing the secondary "unsharp mask"? Or, would doing it in reverse order make a difference?

Edit: Never mind my "sharpening" question above -- I did some more research on the subject and found lots of info that says it's much better to use the "unsharp mask" method instead of the "sharpness" adjustment because there's more control over the parameters, so that's what I'll be using from now on.