Ritrovata and Eclair Cinematic Universes

[rrenault]

I meant in the sense of "American philosophies" as opposed to "European philosophies" with FiM's dedication to high fidelity reference quality encodes intended for home viewing as an example of something the French film establishment might perceived as an "Anglo-American way of doing things", since it's an "infringement" on the "traditionalism" of the Franco-Italian film establishment.

Look at Studio Canal. They'll call up David M for Don't Look Now or Tinker Taylor Soldier Spy, but "never in a million years would we want to subject our precious legacy museum pieces like Breathless and Discreet Charm of the Bourgeoisie to that crass American techno-fetishism".

[tenia]

I think that's over-thinking it. It's likely to be down to which team is handling what project, and it's not a matter of philosophy than skills here, as David M is (fortunately) not the only one able to output proper encodes like these. Gaumont and Pathé for instance have no problems, and they're not the only ones.

In terms of philosophies, it's not down to this, which is more binary, but to the restoration itself, meaning to the labs. In this case : yes, why not. But even then, not all European labs are working in the same way. Eclair and Hiventy aren't even working in the same way since they clearly each output their own recognizable results.

[Drucker]

They trust a lot works like those of Ritrovata and Eclair, perceived as "reproducing the native characteristics of film stocks".

Sounds to me like they are just copy + pasting the words on a press release of some sort but that's just my two cents.

[tenia]

This one might look like this, but it's just because it's the usual way of summing these things up. But this aside, they're not, they do have developped an opinion on this that is trusting these labs' work ane results.

To me, there is way enough elements to disprove this, but seemingly not to them, and that's despite the exercice I've done mostly, actually, to respond to their specific arguments (which felt like a starting point as good as any). For instance, regarding the Argento movies in this set, they've told me the disparity in restored looks (ie the Ritrovata ones being Ritrovata'd and the Luce Cinecitta not being Ritrovata'd) that "it might be because some movies are pre-1975 and the others post-1980". But there already are enough movies affected by this that are on both sides of this calendar threshold ! So nope : it can't be it.

It does make me wonder what would be required for such people to really and deeply start questioning what is happening. I mean, for starters : Eclair and Ritrovata can't be both right. Yet, that's how they treat their works : as a homogeneous corpus of properly restored movies, despite Ritrovata clearly not working like Eclair do (and vice-versa), because otherwise, well, we wouldn't be as easily able to pick who did what, wouldn't we.

To me, it seems as if when thinking we might be reliving the homogenenous magenta-push plague from 20 years ago but in yellow and green-blue, they get stuck on the yellow and green-blue thing rather than the homogeneous part of it. Which actually is THE issue.

[mhofmann]

Look at Studio Canal. They'll call up David M for Don't Look Now or Tinker Taylor Soldier Spy, but "never in a million years would we want to subject our precious legacy museum pieces like Breathless and Discreet Charm of the Bourgeoisie to that crass American techno-fetishism".

I honestly don't think that's the case and it sounds a bit conspiratorial to me. Poor David M. has nothing to do with how labs grade their movies - and I'm sure he's encoded plenty of Ritrovata and Eclair restorations!

[Black Hat]

An article from the latest issue of Filmmaker Magazine on color timing in restorations.

Okay, which one of you is Bingham Bryant?

These worlds colliding, or rather, unmasked, would be quite the discomforting crossover event.

[rrenault]

Look at Studio Canal. They'll call up David M for Don't Look Now or Tinker Taylor Soldier Spy, but "never in a million years would we want to subject our precious legacy museum pieces like Breathless and Discreet Charm of the Bourgeoisie to that crass American techno-fetishism".

I honestly don't think that's the case and it sounds a bit conspiratorial to me. Poor David M. has nothing to do with how labs grade their movies - and I'm sure he's encoded plenty of Ritrovata and Eclair restorations!

Yeah, I guess you're right.

[mhofmann]

An article from the latest issue of Filmmaker Magazine on color timing in restorations.

Okay, which one of you is Bingham Bryant?

No idea, but his feature 'For the Plasma' was just announced as an OCN partner label release.

[tenia]

New one with the 4K resto of Vittorio de Seta's Lost World, restored in 2019 by Ritrovata and one of the worst offender I've seen lately.
Interestingly, it's a WHOLE DIFFERENT restoration than what was done in 2008, already by Ritrovata, under De Seta's supervision but without their trademark color-grading : the AR, the length (it was 105 min on DVD - once converted back to 24fps - it's now 116 min), and the text panels (including, at times, the text itself of these panels !).

[tenia]

Potemkine's upcoming release of Laloux' La planète sauvage will feature a new "color corrected grading respecting the original colors and tints". From what I understood, it's a newly commissionned color-correction and not a licencing of Camera Obscura's color-corrected master.

In other news, we can add to the lists :

* Ritrovata :
Big Guns

* Eclair :
Vie privée
7 hommes en or
most probably Le procès des doges

Interestingly, Hotel de France has been restored and graded by Ritrovata but their signature is quite subdued, except in a few places like are usually where their signature is the most visible and is, well, the most visible here too.

On a sadder note, it seems like we'll need to monitor what Hiventy are doing, as I've been seeing some similarities in now several of their restorations, with a color signature that, in some ways, could be seen as close to what Ritrovata is doing in particular on yellows and blues, but is not exactly the same and tends to be, in particular, more saturated.

This can be seen on Red Sonja, Le gendarme de St Tropez, Cat's Eyes, and in a more subtle way on Astérix et Obélix contre César (and in a as-subtle but less frequent fashion on Le pacte des loups and Astérix et Obélix Mission Cléopatre). In some respect, one could argue Basic Instinct fits within this trend.

[tenia]

Like nicolas just posted here, Le magnifique can be added to the Ritrovata'd list.

[tenia]

Arthur Joffé's Alberto Express can be added to the Eclair'd list (in a particularly obvious way - it reminds me of what Eclair was doing like 10 years ago on stuff like Etat de siège or Muriel).

[tenia]

For those still keeping track of this (and who might still think there's no real correlation between those labs and the end-results overall color timing) : I'm on the verge of litterally proving statistically how much of these corpus of restored gradings are down to the labs and pretty much only the labs.

[perkizitore]

:-"

[tenia]

EURGH.
Fotokem are credited for the color grading, though.

[nicolas]

EURGH.
Fotokem are credited for the color grading, though.

RItrovata only did the scanning for the film as the negative is probably with SC.

Here are some caps of the 4K. The colors in most shots are too natural for Ritrovata yet Lynch obviously pushed for a warmer look compared to the old master.
I think it looks flat-out gorgeous in 4K Dolby Vision.

Spoiler

Spoiler

Spoiler

Spoiler

Spoiler

Spoiler

Spoiler

Spoiler

Spoiler

Spoiler

Spoiler

[tenia]

This doesn't look Ritrovata'd.

[mhofmann]

It's not.

[tenia]

As written a few days ago, I was working on trying to statistically schematize and analyze what I was seeing when it comes to Ritrovata'd and Eclair'd movies, showing how it's not a matter of production countries or periods, and that's it's unique to them, being aside from pretty much everyone else, and that even Ritrovata and Eclair are not working towards similar fashions so they two can't even be discussed as an item.

I've finalised that this morning, and my main article (in French) has been updated to include the findings : https://testsbluray.com/2021/05/08/rest ... echniques/

The database and scoring system do have their limits, the obvious ones being that it's only an incomplete sampling, and that since I've scored myself the presence/absence/intensity of their signatures, it's a bit biased. This being written, I did try a forced model in which I downgraded every signature I thought wasn't that intense, and the sampling still being what it is (277 entries for Ritrovata or Eclair works, 441 entries total, and so many undeniable signed works), it makes no significant difference anyway.

It's also a simpler model that it would need to be, as at best, it'd be capturing other labs possible signatures (like the Fox Deluxe stuff, or some of the movies that can easily be tracked down to being graded by Hiventy), but it'd mean either going to discrete outputs, and the theoretical models for these suck, or trying to build some kind of radar-type map, and honestly, it took me long enough already to do this first passes to try and prove once again this horse is already long dead anyway.

[nicolas]

As written a few days ago, I was working on trying to statistically schematize and analyze what I was seeing when it comes to Ritrovata'd and Eclair'd movies, showing how it's not a matter of production countries or periods, and that's it's unique to them, being aside from pretty much everyone else, and that even Ritrovata and Eclair are not working towards similar fashions so they two can't even be discussed as an item.

I've finalised that this morning, and my main article (in French) has been updated to include the findings : https://testsbluray.com/2021/05/08/rest ... echniques/

The database and scoring system do have their limits, the obvious ones being that it's only an incomplete sampling, and that since I've scored myself the presence/absence/intensity of their signatures, it's a bit biased. This being written, I did try a forced model in which I downgraded every signature I thought wasn't that intense, and the sampling still being what it is (277 entries for Ritrovata or Eclair works, 441 entries total, and so many undeniable signed works), it makes no significant difference anyway.

It's also a simpler model that it would need to be, as at best, it'd be capturing other labs possible signatures (like the Fox Deluxe stuff, or some of the movies that can easily be tracked down to being graded by Hiventy), but it'd mean either going to discrete outputs, and the theoretical models for these suck, or trying to build some kind of radar-type map, and honestly, it took me long enough already to do this first passes to try and prove once again this horse is already long dead anyway.

Absolutely incredible article from start to finish. Thank you for your time and effort in doing such an extensive, necessary breakdown into this sad development. I don't think many of us would mind if the textures of some classic films lean warmer compared to older, magenta-tinged masters if the *restoration* (not reimagination) of the original color scheme happened on a level that includes proper reference material. Very careful guess-work also wouldn't cause massive uproar if you base it on the look of another film by the director / DP of the same era. If both of these options fail for whatever reason, the goal should be to make the film look as natural as possible, respectful of its filmic nature. As you've proven numerous times throughout your article, Rémy, this is impossible to be the case for all of the films Ritrovata in particular re-imagined with their gradings as so many of them look similar across decades, countries and irrespective of any DP / director supervision.

I think a great example is The Straight Story as we've just seen. The 4K master is warmer than the old one but it didn't nearly showcase the many sun-drenched moments as well as the new one as well as the rich colors in the gorgeous sunset I've made a screenshot of. We're now much more invested into the world of the film because it looks more realistic and faithful due to its warmer look. But due to the absence of blanket tints, scenes that are purposefully shot to look cooler (such as the latter part of the film), appear on screen exactly like that and serve as careful contrasts to the warmer look of Alvin's journey earlier in the film. I can't imagine anyone not being happy with how the film looks now.

[tenia]

There are plenty of new masters (or even old, actually) that can lean warmer or colder, that's fine by me, the issue with Ritrovata is that they're always the same kind of warmer, and the issue with Eclair always the same kind of colder. While I can imagine a couple of movies originally actually looking similar for whatever reason, the larger the corpus grow, the less likely it is to tie together 150+ movies. That's where hypothesis' analysis start, and I really typed this article with this frame of mind : can it be a given period ? can it be a given country (I haven't tried "a given region" but it'll give the same result) ? etc etc. At some point, it's just trying to find what actually ties together all these gradings, and, well, the fact is, it can't be the timeframe because the concerned movies go over 50 years, it can't be a country because, it can't be a rightholder because, it can't be having someone supervising it or not because... because everytime there are countless examples that just don't fit this.

And the article stems from keeping being given those arguments again and again : they. just. don't. fit. And the more it goes, the more they can't. They just can't. You cannot have such a wide variety of impacted movies that actually have this one thing in common that nobody has been able to give me within the past 10 years.

The addition also stems from again these arguments being thrown at me, this time because of L'homme de Rio (done by Hiventy, and that unsurprisingly don't retain the 2013 heavy yellow tint). I'm just... tired of it. 277 freaking movies are concerned. How many more before somebody just accepts there is at least something to look into ? Because at this point, there isn't more I can to show how the labs are THAT intrusively leaving their marks on these works, but what are naysayers actually providing to rebute this ? Nothing. I recall discussions years ago from people being very bothered by my questioning, rebutting it as if there was an obvious answer to this state of things, to all these movies weirdly now looking kinda the same and kinda guessable who did the grading before who was the DP... but never telling me what the obvious answer was. And honestly : I think they never did because there isn't any.

But if there is : please people, give it to me, I'll fill in my Minitab table and we'll check. But when a model explains 92% of the results with such a statistical significance, my brand-new training tells me there is 0 reason to go looking for other factors, because the one already explaining 92% of the results is absolutely plenty enough. Though I do wish that if naysayers think they have a better answer, they would just tabulate the results themselves and then tell me what they found. I did my work, they should do theirs.

[diamonds]

As written a few days ago, I was working on trying to statistically schematize and analyze what I was seeing when it comes to Ritrovata'd and Eclair'd movies, showing how it's not a matter of production countries or periods, and that's it's unique to them, being aside from pretty much everyone else, and that even Ritrovata and Eclair are not working towards similar fashions so they two can't even be discussed as an item.

I've finalised that this morning, and my main article (in French) has been updated to include the findings : https://testsbluray.com/2021/05/08/rest ... echniques/

Excellent, tireless work tenia. One sincerely hopes this receives its due attention from home viewers as well as some movers in the industry.

A minor correction regarding one of the screenshot captions: Rien ne va plus and La fleur du mal were shot by Eduardo Serra, not Renato Berta.

[tenia]

A minor correction regarding one of the screenshot captions: Rien ne va plus and La fleur du mal were shot by Eduardo Serra, not Renato Berta.

Thanks for catching this one, it's now corected !

[tenia]

Oh and don't hesitate to make the article go round a bit. I've already discussed with both labs, they know what my methodology anyway so I don't think I have anything to lose for it being out there as much as possible.

And if somebody is able to get this to slow down or even better stop by making this article visible to someone with such reach, then it won't have been for nothing.

[nicolas]

Oh and don't hesitate to make the article go round a bit. I've already discussed with both labs, they know what my methodology anyway so I don't think I have anything to lose for it being out there as much as possible.

And if somebody is able to get this to slow down or even better stop by making this article visible to someone with such reach, then it won't have been for nothing.

I couldn’t be happier if the article achieved just that but I’m feeling a little pessimistic in that regard after having attempted something in that vein myself (albeit on a smaller scale) by pointing out all the flaws in the worst UHD: Kino Lorber’s encodes of Fear and Desire and the Kubrick shorts in 4k by making caps and BDInfo graphs for all of them with comparisons to their old 2012 BD. Not to go too much off-topic, hence only a brief quote from what I wrote to Geoff D on the other forum, but neglect is probably the most likely outcome for now - but still, if we don’t speak up in these matters, absolutely nothing happens in the slightest as others certainly won’t.

“I decided to copy-paste my post [about the Fear and Desire UHD] into the KL main thread and, naively, I thought that I *might* catch our friendly Kino Lorber Insider’s attention as what they released here is beyond anything remotely acceptable even for the lowest of standards they set for themselves. Verbally, I couldn’t possibly match his disrespectful, passive-aggressive tone, hence all the evidence with caps and charts beyond my usual talk. I expected and got a modest blast of animosity from a few die-hard Kino Cult members but then, genuinely surprising to me… an endless void of pure ignorance and silence. As soon as the regular posters in this thread (their key buyer group anyway) started breaking down their studio deals again as if nothing happened, I knew that’d be the end of that story and nothing will change to the better, not even an inch.

Examples like Fear and Desire are even more sad as they reveal how much better physical media could be if the people in charge only cared a little more beyond their immediate 9-to-5 job compensation, rank, industry contacts,… and build a legacy of contributing something meaningful to this niche.”

These labs go the easiest route as well, following their boss’s guidelines even though the colorists probably know that they’re not doing the right thing for the films and that they could do better if they only got the chance. Many authoring houses also could do better, I’m sure of that, but if their cheap, half-a**ed encodes are approved anyways, why bother wasting more time, rendering power, personnel etc. It’s funny that both of these parties often go hand-in-hand. Badly graded Ritrovata masters are often a pain to encode because of a lighter grain structure, crushed highlights due to the grade (which invites encoders to throttle the bitrate there) and yellow is trickier as a color anyway. That’s why Criterion BDs of Ritrovata and Éclair masters are generally unwatchable among a sea of macroblocking.