Spoiled Kindle experience: too much sloppy content

[speakingtohe]

Originally Posted by Ti-Ron

From my experience, ebooks don't have the same traitment as a paper book. Normaly, editor ships the book to be converted to outsourced production. Many times, there's no proof reading on the e version...

Maybe it is a case of too many books too little time. One company that seems to produce quality ebooks is Hard Case Crime. They only publish older books and less than 6 most years. I think that most of the books are still in copyright but long out of print.

My understanding is that it takes two weeks to a month for a meticulous person to scan, proof and format a 200 page book. I could be way of base here and scanning and OCR are much improved since I tried it 20 years ago, but even at a minimum wage you are probably looking at $400 to $1,600 per book. Well worth it IMO, but if you have 20,000 books you want to publish you have to have a lot of meticulous people working on them and I doubt there are enough qualified people available to proofread all of the books being sold on Amazon.

Helen .

[ilovejedd]

Originally Posted by jswinden

Another reason many eBooks suck compared to their printed versions is that publishing companies don't always keep the digital files they used to create the printed version. When they go back a few years later to create the eBook version they have to quickly and inexpensively digitize the text and then format it for an eBook. This process often produces spelling errors and if they don't thoroughly edit the digitized text we will see these errors.

Yeah, I've noticed quite a number of obvious OCR errors on back catalog titles. There's less likely to be errors with new titles from BPH. Alas, the same could not be said for majority of indie ebooks I've tried. I've taken to steering clear of independent ebooks unless it's recommended by my OCD friend.

[Elfwreck]

Originally Posted by speakingtohe

My understanding is that it takes two weeks to a month for a meticulous person to scan, proof and format a 200 page book. I could be way of base here and scanning and OCR are much improved since I tried it 20 years ago,

Scanning takes literally minutes per book, if one is allowed to destroy the original. (I'm assuming "chop the spine off" counts as destruction.)

OCR software is much better now. And while good formatting takes substantial time, and so does a line-by-line proof, a read-through proof by a competent editor shouldn't take more than a few hours for a 200-page book--maybe 1 day if it's troublesome, has a lot of unique names or specialized terminology.

A really careful proof would indeed take many days; a proof pass that would catch errors on the title page, repeated uses of "tlie" for "the" (common OCR error for some fonts), and line-break errors, would not.

Originally Posted by ilovejedd

Yeah, I've noticed quite a number of obvious OCR errors on back catalog titles. There's less likely to be errors with new titles from BPH. Alas, the same could not be said for majority of indie ebooks I've tried. I've taken to steering clear of independent ebooks unless it's recommended by my OCD friend.

Maybe the industry needs OCD OCR software?

speakingtohe: the other thing to consider is the return on investment. $1000 is fairly significant if the volume is low and the sale price is low.

[speakingtohe]

Originally Posted by Elfwreck

Scanning takes literally minutes per book, if one is allowed to destroy the original. (I'm assuming "chop the spine off" counts as destruction.)

OCR software is much better now. And while good formatting takes substantial time, and so does a line-by-line proof, a read-through proof by a competent editor shouldn't take more than a few hours for a 200-page book--maybe 1 day if it's troublesome, has a lot of unique names or specialized terminology.

A really careful proof would indeed take many days; a proof pass that would catch errors on the title page, repeated uses of "tlie" for "the" (common OCR error for some fonts), and line-break errors, would not.

Well I did specify meticulous

I have read a couple of books that have die for the or he for the. Never thought about it being the font.

By meticulous I was talking about a person who sits there with a paper book and compares line by line. I believe that is what HarryT often does.

I also think this is not done by most, but my impression from a google search is that proofreading companies charge $6 per page for easy to proof books or documents.

My main point although perhaps poorly expressed is that even if Amazon was to take on the burden of having all of the books they sell proofread in the most basic form, their might not be enough proofreaders out there to accomplish this in a timely manner.

I would think that Amazon has the books that they actually publish proofread, but I don't know.

Maybe they could have a separate rating system, with choices of:

Definitely proofread
Sort of proofread
Possibly proofread
What a joke

Myself, I only buy books from authors I love, and generally they have been around for a while and the quality is better than acceptable. If it is a tad crappy, chances are I am fairly forgiving because I love the author.

New authors I get from the library, as well as many from familiar authors, which may seem limiting, but I can always find a book to read and enjoy most of them. Quality of library books has improved immensely in the last 3 years.

But on the article cited in the original post, why would anyone buy book 20 from the kindle store if a good portion of the previous 19 were extremely inferior? I'd be kind of embarrassed to whine about it publicly.

Helen

[Elfwreck]

Originally Posted by speakingtohe

I have read a couple of books that have die for the or he for the. Never thought about it being the font.

If the font has particularly thin tops to curves (or the scanning specs lightened the text), the OCR program can miss the fact that the "h" is one letter, and instead split it into two. If the "t" is clearly separate, that gets read as "tlie;" if the crossbar is long enough to almost touch the "h," that often gets read as "die"--and then, of course, a spellcheck program will miss it. (And search/replace can't be used; it has to be manually checked.)

Books with those problems are also prone to having "diat," "dian," "diere," "diese" and "diis." The "tl" problem isn't as common for longer "th" words; the internal dictionaries seem to be better at recognizing that "there" is more likely than "tliere."

These are problems that don't need the intense, meticulous check to fix; they just need someone flipping through pages enough to notice *one* example, and then doing quick, focused clean-up on the rest.