Monday, 19 May 2014

Sean Michael Robinson: A Proposal

Previously on 'A Moment Of Cerebus':
Dave Sim, working with George Peter Gatsis, has remastered the first two collected volumes of Cerebus to restore details and quality in the artwork lost over the thirty years since they were originally published (as detailed here and here). After Cerebus' original printer Preney Print closed its doors, Dave Sim moved his printing to Lebonfon in 2007 as at that time they were still capable of working with photographic negatives and making printing plates as Preney had done. And then Lebonfon switched to digital scanning and printing - a technology which struggles to faithfully reproduce Cerebus' tone without creating moire patterns (as detailed in Crisis On Infinite Pixels). 

The following proposal was written by Sean Michael Robinson in response to Dave Sim's Weekly Update #31 (posted on 16 May 2014):
...I think we could be looking at compensation "going forward" for what Sean and George are doing and will be doing (as opposed to what they have done so far). It depends on how fast we want to go, in a lot of ways. That's definitely a question for George and Sean:  which would YOU do if you were me? Compensate you guys so as to be "front of line" in your working day or stick with the volunteer basis because all of this is going to take a while no matter which way we go?


Hey Dave,

I've read back over what you've written a few times now, and I'm going to take you at face value here when you're saying "how would you do this?" This. This huge document is how I would do this.

When I started chiming in with advice and bossy opinions a few months ago I had no idea what this would turn into. But not that I've been enmeshed in the process, I want nothing more than to see this project completed by someone, in the most effective way possible.

First off, I want to be clear about one thing -- the prior work on the restoration is not mine. I've done nothing so far other than to suggest ways that the already done work could be modified to print without half-toning, and eliminate the moire that was plaguing the pages. If I had worked on this from the beginning, I would have done things very differently, and I don't think it gives either George or I much credit to say that the current work represents me in any way, for better or worse. It's George's work, with my band-aid.

So, there are two distinct but overlapping projects here -- creating digital "negatives" for the Cerebus/High Society books, and creating digital "negatives" for the entire Cerebus series.

I realize that from a financial perspective, it might be tempting to address just the first project, since you still have some stock of the other books. But from an economy of scale perspective, I think it makes the most sense to, once we've definitively worked out our procedure, proceed on the whole enchilada. Several of the things I'm going to propose here depend on having trained people working in tandem -- it'll be easier to keep people if there aren't large gaps of time between production times.

But it's workable either way.

First off -- for those of you who aren't Dave, let me say a little about myself. I'm a freelance illustrator, designer, and writer. I had a book out last September by powerHouse/Random House, which I illustrated, co-wrote, designed, and did pre-press for. I've done pre-press for numerous clients over the years, mostly albums, but books and smaller publications as well. Perhaps more important for this project, I'm a cartoonist and print-obsessive.

Most of my work over the past year has been local work in San Diego. This month? About 30 hours of illustration for a financial app. Interior and color illustration and logo for a local band's debut album. A process portrait for an independent movie (drawn in stages so the actor can be filmed "drawing" on each stage). Posters. Sometimes (this week, for instance), I have more work than I can possibly handle. Other times (next week? Looking likely) I have little on my schedule other than, finally, making some business cards and a real website and drumming up some future business.

As an illustrator, I work at a rate of $50 an hour. I'm stating this upfront because, frankly, this is what is competing for my time. When I'm writing this right now, I have a roughly penciled album cover on my drawing board that I should be tightening up. I'll be teaching a watercolor class this afternoon for the same rate. Any time helping out on Cerebus, even just writing responses to tech posts here, is in direct competition with that work.

I told you over the phone, Dave, that I've been chiming in here, arguing, testing stuff, throwing out opinions, because I think you're the greatest living cartoonist in North America, and that it's a crime that your books aren't in print, in beautiful editions that are making you money and spreading your work around. With the right procedure, with the right capitol in the right places, I think we can make this entire thing happen in a matter of months.

So -- procedure.


You'll have to get a computer person to come in and advise you here before you buy anything. Perhaps John Funk can advise you? You'll need--

a. multiple hi-resolution (2400 dpi) large format document scanners that
b. can either work together on the same computer or
c. that each have their own computers

Older SCSI scanners could be attached in series to the same computer, but with newer USB scanners it seems to be more of a technical problem to have them all connected simultaneously. This is something that someone with more know-how than myself could hopefully solve in a few minutes.

The most affordable scanners that fit these requirements are the Mustek A3 2400, which are priced at $270 apiece. Even if you had to purchase a used PC for each of these individually, you could still be in the $650 range per scanning station.

The purpose of the multiple stations? To save on labor and time. 2400 dpi= an awful lot of time.

And where does your labor come from?

I have two associates whom I hire every time I have a high-volume technical task. They've worked for me doing scanning and photo adjustment, they've worked for my friend Tim doing audio book quality control. Depending on the client and the volume and the task, they generally get paid between $20 - $25 an hour.

I'd suggest you ask John Funk to find a few people similar to my associates in the Kitchener area. A technically-minded young person(s), ideally, someone who you can pay $15 to $20 an hour to sit and feed the scanners for a few hours a day for a few weeks. (I see that the national minimum wage in Canada is currently $11. $15 seems like a good enough amount over that to make up for the mind-numbing aspect of the task, and to ensure that you can keep the person once you've got them trained).

I'll have a Mustek 2400 arriving at my house in the next two or three days. As soon as it arrives, I'll do a few speed trials, and figure out exactly how long each scan will take.


A. negatives

Every negative in your possession, at 2400 dpi greyscale.

I'm now convinced that you're right-- 2400 dpi is the way to go, for half a dozen reasons. (Anyone reading this who thinks otherwise—I'd advise you to take an hour to read Dave's posts and my responses over the past two months. )Yes, there will be only a marginal improvement on the page from 1200 dpi to 2400 dpi, and only in the area of the teeny tiny lines (and possibly teeny tiny tone). But it matters to Dave, and it matters to me, and although the scanning will be much longer, the adjustment should be significantly less labor intensive-- no rescues of little lines necessary. It will also be significant in the future-proofing department.

B. original artwork

All original pages in your possession, at 2400 dpi color (possibly a slightly lower resolution, if there's a lower native resolution on whatever scanner we go with, that makes mathematical sense to the destination size)

And lastly--

C. printed pages, when missing negatives and original art

This could be done in San Diego, by my team, if need be.

I'd like to approach this book by book, so that the pre-press work could be happening simultaneously. So, scan every available negative for book X, then every available original page for book X, then any missing pages of book X from the best available printing.


I'd take a look at the first batch of negative scans and, using a representative page, put together a baseline exposure/adjustment. i.e., scanned on this type of scanner, at this exposure, the negatives will look the most like the original printings when adjusted at Y brightness control. I'd then set up a Photoshop script to do this as a batch adjustment, i.e. generate a second version of the scans using that baseline adjustment and letting the computer do the work without any assistance.

Meanwhile I'd take a highlighter to a Preney copy of the phone book in question and hilight any areas that are obvious flaws in the original printing (not too different than the notes I sent you via mail on those High Society pages. Was there dirt or pencil schmutz on the original negative? Do any corners need squaring? Any line breakup on the original negative? Missing tone?)

Next I'd have my associates Mara and Matt work through those pages represented by negatives only, and make those changes. Anything beyond basic adjustments like these I'd do myself, on a case-by-case basis.

As I was implying before, this works better when there's a large volume of work. If I had dribs and drabs of stuff for them, say, an hour at a time, it doesn't make sense for me to train them to do a task. I'd just do it myself. But at a high volume of work, it doesn't really make sense to pay me to do something when someone else who can do it equally as well will do it for less pay.

All of this is setup to minimize the amount of labor. You'd prioritize it by type of action.

Just setting up a batch conversion or adjustment (i.e. "Computer, that thing I just did? Do the same thing to these other 1,000 files") is essentially free, once you make the script. So there's no loss trying this, other than computing time and a little electricity in setting it up and doing test prints of your results and seeing how close that gets you.

Having trained and skilled but not necessarily expert people do guided work on pages is much cheaper than hiring someone who could be making $50 an hour doing something else to do the same work. So, you prioritize those people second, and give them all the work that they can do as quickly as an expert. (This is basically the situation at a small publisher, where most of the scanning and adjustment will be done by no-pay or low-pay interns, under the guidance of whomever is designing the book). I toyed with the idea of asking volunteer fans do this work as well, but then you're in a similar position as before. It's significantly easier to train someone in person, it's easier when it's people I trust and have worked with before, and it's easier to say "I need you to do this now, this week" to someone who's being compensated for their labor. I feel like if we farmed this out to fans, you'd have to have a significant vetting process, otherwise you'd be back to having to re-do work as it comes in, spending more on the back end because you tried to save money on the front end.

And then lastly, you have the expensive work. Guiding the whole shebang, looking over the pages, doing most of the adjustment of the original artwork, which is more labor-intensive than the negatives, as the pages have to have a certain amount of panel-by-panel and even section by section adjustment to insure no breakup of lines. (Mentioned this last week, but, unlike the negatives, which have already been contrast-adjusted all those years ago, all ink lines are not created equal. You can have lines break up because of how grey they are, not because of their thin-ness, which just means you have to take a lot more care in how you adjust the page, especially if you want to maintain the original overall balance in the first place.)

(Alternatively, if you wanted to avoid this in the first place and all you wanted to do is create duplicates of the Preney printings, I think 2400 dpi negative scans plus book scans of the missing pages could get us there with almost no adjustment, relying on automated adjustment and conversion for most of the work. This is in all likelihood how Lebonfon would have handled the books, had you paid them to do the negative conversion. The main difference being, they were scanning the negatives at an unacceptable resolution, namely, 600 dpi. This is by far the most expedient and economical way to do this job).

So, what's the price tag for all of this? For the whole 6,000 pages?

We'd need to test a few things, and that would give us some data to extrapolate out over the entire enterprise. Either take Lebonfon up on their offer, or find another printer that will offer to do a free or reduced rate sample for you. And then ask John Funk to scan the first four “Night Before” pages at 2400 dpi. Get me those files, and I'll go through them from scratch and time myself. Meanwhile, when the Mustek scanner arrives, I'll do some speed tests here. When we run a sample, we can compare head to head 2400 dpi pages adjusted by me from the original pages, and 1200 dpi reductions of the same. Possibly throw in a few other versions if desired.

Just some rough numbers to give us an idea of the scale we're talking about--

Let's say I personally am working for a reduced rate of $40/hr. Given the volume involved here, that certainly makes sense.

If I personally adjusted every page of original artwork at, let's say, 15 minutes average per page--

2,000 pages X 15 min= = 500 hours X $40/hr = $20,000

Let's say that, given that the majority of the pages would be coming from negatives or original printings, that most of the work on these could be done through automated processes, but that, on average, the tweaks and adjustments could be kept to 10 min per page by my San Diego associates--

6,000 pages X 10 min/pg = 1000 hrs X $20/hr = $20,000

For the scanning, I'm going to have to do a few speed trials to make this have any kind of accuracy. But let's say you have 5 scanning stations and one operator. 10 pages an hour seems reasonable at that resolution. (might be faster. Will need some actual numbers here)

8,000 pages / 10 pg/hr = 800 hrs X $15 = $12,000

Add in somewhere between $3,000 and $5,000 for equipment, another $5,000 management time and inevitable overruns, and you arrive at somewhere around $60,000.

For some perspective here, Colleen has paid $40,000 for the first book (?) worth of A Distant Soil restoration.

I don't see a cheaper way to do this unless you have the resources (and unpaid interns) of a publisher at your disposal.

If this seems like a large number to anyone, keep in mind that it represents several months of working time for at least four people.

There is a budget alternative for the first two books that, to most readers, will be indistinguishable from this--

Take George's files. Batch convert the whole thing to bitmap, print out test pages, and do finer-tuned adjustments when necessary. Print the book and move on.

I'd be happy to do this for you, next week or whenever the work lets up, if you'd like me to. I think to most Cerebus fans the result would be pleasing.

But to get to the level of minutia that you and I have been discussing, to make these printings equal to the level of detail in the Preney books, it's higher-res scans and files.

Scan all available negs at 1200 dpi (¼ the speed of 2400 dpi). I'll do some test adjustment, and batch convert and adjust the whole thing, with no additional changes.

But for something more involved than this, I just can't do that work on a volunteer basis. I hope you understand. It's already a struggle to get my own work off the ground again while spending so much time per week drawing for clients. I can't take on what essentially amounts to a full-time job for six months without compensation close to what I'd be getting for other work.

That said, if you have someone else that can do the work for cheaper, I'd be happy to consult on the work. These are all commonly-known procedures, and there are many people that I'm sure would do a great job of it. I'll also always be happy to take a look at files for you or give advice on anything you need. (Speaking of which—I'll be getting those adjusted portfolio files to John Funk this week-- just a little bit of sharpening required and then some test printing)

Anyway, unless we decide to go forward on any of these things, consider this my finished statement on all of this. This is what I think it would take to get it done. I think there's a marginal difference between 2400 dpi and 1200 dpi, yes, and it does make the scanning stage more expensive, but I think the amount of hassle saved futzing with the scans, and the amount of detail present in the finest areas, would more than justify the scanning hassle.

Lastly, I just want to say that, in my experience, up-front is the best place to spend your money, whether it's recording music, drawing, or pre-press. You draw it right first in pencil so you won't have to white out later on. Ya know?

Now, it's time to draw. Let's see if I can follow my own advice this time!

Sean Michael Robinson

Dave, I'll be sending you my book as soon as my order from Random House warehouse comes through! (don't have any copies right now) Thanks for your interest.

The other (probably long-shot) economic alternative--

Give a publisher (say, IDW) a contract to print an edition of the books, say, a single print run, in exchange for them creating a digital negative for you, from your specifications. I have no idea if a publisher would go along with something like that, but it might be worth making a few calls...

(As I said before, the economics involve make it cheaper for a publisher to do this kind of work, especially a publisher that works with interns)

Another possible money-saving device-- as I said above, don't use original artwork at all, only original negatives. Huge savings of labor. The only downside is you're resigning yourself to any line breakup present in the original printings.

Lastly, even if we were going back to scratch, I'd still be strongly in favor of using George's files and restorations for the early pages of Cerebus. The negatives are in horrible shape, and with those materials, there's no advantage to moving to a higher resolution. So I'd be in favor of just bitmap-converting his work and going from there...)


Okay, two more things--

I see in my rough math above that I have double-dipping--2,000 pages being adjusted (from the negatives) by one team of people, and then the same pages (from the original artwork) being adjusted again by another person. This obviously wouldn't need to happen.

Secondly, I hope it's clear here that I'm trying to spell out a procedure that could be used regardless of who's doing the work. I think having it happen by essentially three teams, tiered by their level of expertise, should be the real take-away here.
A. 2400 dpi. Scanning, stage one.

B. Batch conversion/adjustment, i.e. automated procedure for the entire thing. test printing, comparison to Preney editions.

C. Only adjust problem pages/problem areas/ previously flagged areas, and have those adjusted by people trained in those procedures.

D. Only really problem pages (or pages scanned from original art) adjusted by someone with line art expertise.
I hope that, even if nothing else comes of this, that this procedure can be of some use to you.


Be sure to visit Sean's website Living The Line.


Anonymous said...

Colleen didn't pay $40,000 for the first book. She wrote she raised $40,000 and that some of that money was used to pay ebay and paypal fees, ship items sold, purchase of packaging, and money to pay an assistant and her personal overhead. This covered two books and 4 issues of the new comics. This was covered on Facebook.

uthor said...

I would look into renting the scanners or selling them off once you are done. No need to spend $3000-5000 on a bunch of high end scanners that won't be used for more than six months. Maybe keep one once everything is done and be out only $1000 in the end.

Not a lot in the grand scheme of things, but it's a bit of money saved.

David Birdsong said...

I have always been an adherent to the idea that more resolution is always better. I would go with this method if it were me.