Jump to content

Is FilmFabriek aware that their sound scanner has audio problems?


Recommended Posts

Thanks for all the feedback, everyone!

I have not read anything as yet. I hope to catch up on the thread this week. Have not had much time for anything. I've been working on getting the cine' film Archive reels into ABC order as well as half a dozen other jobs. About 80% done with ABC 16mm reels. Then I will reshuffle them into better ABC organization to allow for working space. Once I have final organization, I will label the cans top and sides. Then I have to computerize it all. 

 

DSC09415.jpg

ABC's are fantastic! I can even find my split reels now...filed under 'S.'

 

It has been a lot of work, as many of the reels were unknown and had to be reviewed some on the rewinds to figure out the content. Some will have to be broken down as they contained various films. Computer is no big deal. I just make files for each film. No need for database software to buy or learn. I can even put screenshots from the films in the files. It is searchable. It is fantastic!

I use the same system I use in the Daniel D. Teoli Jr. Archival Collection. I'm not one to complicate things. I'm not an overthinker...I'm a simpleton.

 

Screenshot%2003-21-2022%2009.07.57.jpg

 

You need room to work on the shelves...pressure packed is not good. So, I will keep shuffling until I got it half-ass right.

 

DSC09417.jpg

 

Even though I don't have room to keep the Archive all organized in one spot, it is better than it was.

 

DSC09422.jpg

DSC09424.jpg

 

Got many hundred small reels. Hate em. 400 footers are my fav. The small reels will go into ABC boxes I use for my optical disc library. My dream house would be a small ranch with a roof that doesn't leak with 50 chrome wire shelf units from floor to ceiling. 

Shelving is king...it separates the hoarder from the non-hoarder...a place for everything and everything in its place!

 

film%20collector%20Dennis%20R.%20Atkinso

Internet Photo: Fair Use

 

I've closed down the cine' Archive more or less to new acquisitions and have got rid of about 380 8mm reels.  I've kept about 535 8mm reels for now. I hate 8mm, just too low Q, especially with archival material. There is no use acquiring more and more 16mm film if it can't be scanned. But for now, I'm keeping the bulk of the 16mm with the hope something changes in the future. And if it never gets scanned, someone else can deal with it when I kick off. The pickers may disperse it on eBay, or it may all end up in the nearest dumpster. Just no telling. 

I had met an old gal in the Post Office a while back that took an interest in my film work. She saw me unpacking some film shipments there and we got talking. Over time she had expressed an interest in buying me a cheap scanner for $50k. She was retired from biz and well off, she had few interests except a trip once in a while.  I guess she was nostalgic from when she was a girl in school and watched the films. She didn't like the stag films, but she was not too prejudicial, so she overlooked it.

Anyway, last time I talked with her I ran down the scanner options with her.  Lasergraphics will never answer their emails.  I would not be comfortable with sending them a f'ing penny. And the FilmFabriek won't scan the sound right. So, I told her I was not comfortable taking her $$ as I had no place to go with it.  What a mess with these GD scanning companies. 

Well...that is how it is on 3.21.22 

 

Edited by Daniel D. Teoli Jr.
Link to comment
Share on other sites

On 3/9/2022 at 10:17 PM, Tyler Purcell said:

I have talked to FF about it, they didn't understand what I was talking about really. 

I think the problem is the tension motors, not the capstan itself. I think the capstan isn't strong enough to remove the tension motors which are not consistent. It needs a larger diameter capstan which can have more control over the film. Also, a cassette tape machine has a huge flywheel and pinch roller which holds the tape to the drive motor. There is no real way to do that on these machines. 

Magnatech gets away with this by having two big capstans. 

 

Just getting back on this thread today. After I posted earlier in the week, with hopes of reading the replies, I found another few hundred 16mm reels in storage to organize. So did not have time. I've found I have 2 or 3 copies of some films...all because of not knowing what I have. That is a big problem when you got stuff all over the place. 

That is terrible Tyler. I mean, it is good you were at least able to discuss it with them. It is not like Lasergraphics where they won't answer emails. But FF is not much better if they can't or won't understand the issue. OK, they will answer your inquires--but do nothing about it. Reminds me of the Russian post software I use. They eventually answer emails, but you are in the same boat as FF...same problem and nothing has changed.

I'm thinking FF should lower their machine way down in price if it does not do sound right. $45K for a 'so called' sound scanner, that does not do sound, is way too much. 

 

Can the sound be fixed in Audacity? Or does it get out of synch? How have you worked with this sound issue Tyler?

Well, I got half an hour, now about 26 minutes, so will see what I can read here.

 

Edited by Daniel D. Teoli Jr.
Link to comment
Share on other sites

  • Premium Member
4 minutes ago, Daniel D. Teoli Jr. said:

Can the sound be fixed in Audacity? Or does it get out of synch? How have you worked with this sound issue Tyler?

Nothing can be done from my experience. The only way to solve it is to scan with AVI and use their own software which can link the avi file to the aiff file. This is low-quality image 8 bit solution. For some reason when you scan that way and work with that software it all works fine. If you do DPX, if you drop it into Resolve, it does not sync. I haven't researched why, it's easier to just fix the sync. I simply look at the waveform and align it to the 2 pop and adjust the speed of the audio till it matches. It's not difficult, just takes a bit of time per clip. 

We rarely do audio work, maybe once every 2 - 3 months. 

Link to comment
Share on other sites

On 3/10/2022 at 8:11 AM, Dan Baxter said:

To be fair optical audio off a Lasergraphics is unmitigated shit. Professional companies just transfer optical audio separately.

 

Are you serious?? I thought Lasergraphics was King or at least Queen.

Now, I'm not interested in hi-fi or multi track crap. Some films have narration, and some films have synched dialogue. Maybe an old Soundie here or there from the 40's. That is all I'm after. 

There is a guy on LinkedIn that transfers sound like you speak about, off films. I wondered why anyone would do that if the scanner can read the sound. But I guess scanner sound is not good enough for the big boys. 

Here is a 16mm sound film Perry scanned for me a few years ago when I first started... 

 

 

It is a raw scan. OK, Perry has a $$ Lasergraphics model, but won't their El cheapo Archivist do as well with the sound?

Perry also did one by Deanna Durbin. I thought sound was great. Beautiful little film. As soon as I put it on YT, they took it down within a few hours and banned me for 2 weeks. That was my intro to being a film archivist. I paid $35 - $40 for the film, maybe it was $160 to scan, + shipping both ways and boom I get banned for my troubles. Plus, I paid Perry to burn me some discs. I didn't know how to burn Blu-ray back then, make MP4 or even make titles. I'm an old fossil still photog from the 70's. I knew nothing of cine' film and video.

Here is the 'banned film' that Perry did.

Internet Archive Search: Deanna Durbin teoli

(Raw scan)

 

Edited by Daniel D. Teoli Jr.
Link to comment
Share on other sites

On 3/10/2022 at 10:42 AM, Perry Paolantonio said:

That's completely ridiculous, and also untrue. It is in fact, among the best optical soundtrack reproducers currently available.  

We've had filmmakers who are intimately familiar with their own soundtracks tell us they've never heard their optical tracks sound as good as they do after we've scanned them.

Honestly, what is up with you and spewing information about a machine you have apparently never even used? Why do you insist on spreading misinformation like this?

 

Don't know who is telling the truth. I hope to get back to the thread later to see who else chimes in on this subject. I hope Robert has something to say on it.

And it is not that I don't believe you Perry, it is that I know nothing about the scanner competition. So, I do not know where Lasergraphics fits in with sound reproduction. 

Like I said previously, I was very happy with the optical sound you did for me Perry.

Edited by Daniel D. Teoli Jr.
Link to comment
Share on other sites

1 hour ago, Daniel D. Teoli Jr. said:

Don't know who is telling the truth

If it were me, I'd believe the person who uses the machine every day. But I'm biased and clearly don't have a clue what I'm talking about. That being said...

Here is what it boils down to: 

  1. The lasergraphics hardware reader captures an audio frequency range of up to 20kHz. This is more than an optical soundtrack can reproduce. Much more than a 16mm optical track can reproduce. The gap between the top end of the frequency range of the signal and the top end of the sound reader, at least for the frequencies within the range of human hearing, will be mostly white noise. 
  2. "Hiss" is white noise. White noise is random. Film grain is also random. The white noise you hear is the sound of the film grain in the frequency ranges above top end of the signal. In the case of 16mm that means almost everything between 6kHz and 20kHz is white noise. A 16mm soundtrack would have been mixed to cut off the signal above 6kHz because a 16mm optical track projected at 24fps can't reproduce frequencies much higher than 6kHz. 35mm can, in part because the film is moving at a faster speed per second through the machine. If 16mm film ran through the projector at 48fps, it would have a frequency response similar to 35mm.  
  3. In a projector or a telecine that uses a light and a photocell, there is a hardware low pass filter that cuts off just above the top end of the signal. This removes the hiss above 6kHz resulting in something that is more immediately pleasing to the ear. HOWEVER, it will also remove valid signal above 6kHz, weak as that signal may be. 
  4. The optical track reader in the Lasergraphics scanner captures the sound at a sampling rate of about 80Khz. It does this by compiling a long, skinny image of the soundtrack, with one line of the image representing a single audio sample, 80,000 times every second. (And unlike traditional optical reproduction, slower speed through the optical reader means better resolution of the image of the soundtrack, because the number of samples is constant - run the film faster and you get fewer samples per given length of film. Run it slower and you get more). In any case, the long strip image is a digital image representing the optical track on the film. This means it sees the grain as well as the signal. And the grain is random. And random grain = white noise. 
  5. The signal you get from the optical track reader is the same, regardless of whether noise reduction is on. Without noise reduction, you get some white noise from the film grain. With it on, you get the exact same signal without the white noise. because the white noise is the film grain and the film grain is gone (see below)
  6. The optical track reader on the scanstation reads the soundtrack while the film is going around a roughly 2" diameter drum. This is a relatively tight radius and has the effect of flattening the film, resulting in a sharper image of the soundtrack, and getting better sound reproduction than a reader in the middle of a straight run of the film, or a picture of an entire frame of film, a la AEO-Light.

Lasergraphics Optical Track Noise Reduction: I've spoken at length with Lasergraphics this week about what exactly is happening under the hood here, so this information is from their engineers. I was also incorrect about how they do their noise reduction. When you capture with NR on, they are not applying noise reduction to the captured sound. What they're really doing is grain reduction on the image of the soundtrack *before* the image is converted to an audio file. This removes the white noise caused by the film grain, without affecting the image of the waveform. Thus, the same signal with and without noise reduction, but no white noise (film grain; randomness) on the NR version.  

We are actively investigating whether we want to start capturing optical audio with the noise reduction on, now that we have this information about how it's doing that noise reduction. We'll likely do a lot of testing on this before making a decision. In the mean time, we will continue to capture as we have been because even though it's capturing some noise (film grain) it's also capturing the full signal. And it's easy enough to remove the noise later if need be.

The track reader in the Archivist is different - conceptually, it functions like AEO-light, by looking at the overscanned frame's capture of the soundtrack (however, it is Lasergraphics' implementation, and isn't the same code as AEO-Light). Each frame's length of soundtrack is stitched to the next frame using some overlap, to form a long image that is then decoded into sound. Both AEO-Light and the Archivist software-based sound reader suffer from the same issue: If the film is warped when the image of the film is taken, stitching is harder and you may also get warbling in the sound, due to the out of focus image of the track. Additionally the resolution of the soundtrack in these cases is significantly lower than in the hardware Lasergraphics optical track reader, and is dependent upon the resolution you scanned the picture at.

Archivist Software/AEO-Light reader: Let's assume you're scanning a 5k (full overscan) image. Just to make the math easy, let's look at 15 frames of footage, scanned at a speed of 15fps (not relevant here, but is below): That's about 3800 vertical pixels times 15 frames, or 57,000 pixels of soundtrack information per half second of footage. Some of that gets thrown out because it has to be used to overlap with the next frame when stitching so the number would be lower. Figure 200 pixels per frame overlap, and that leaves you with about 54,000 samples per 15 frames. 

Lasergraphics hardware reader Assuming you scanned at 15fps, that 15 frame length of soundtrack is represented by 80,000 samples (because it took 1 second to go through the hardware reader, at 80,000 samples per second). That is all dedicated to the soundtrack. There is no overlap necessary because it's separate from the image camera. Scan it at 7.5fps (the standard speed for 4k HDR on the ScanStation), and that same 15 frames of soundtrack is represented by 160,000 samples. 

 

Based on specs, sound quality, and implementation, the hardware optical reader in the Lasergraphics scanner is gets you better reproduction of an optical soundtrack than a traditional projector/telecine exciter bulb and photocell setup, and in most cases a much better track than is possible with AEO-Light or implementation like it, that are based on scans of the picture+track.

(and for Dan Baxter: we tested this on our Scanstation with and without NR, a Philips Shadow HD Telecine across town, and on AEO light. Happy to try other track readers, but we don't have ready access to any) 

Edited by Perry Paolantonio
  • Like 1
  • Thanks 3
Link to comment
Share on other sites

6 hours ago, Daniel D. Teoli Jr. said:

I'm thinking FF should lower their machine way down in price if it does not do sound right. $45K for a 'so called' sound scanner, that does not do sound, is way too much.

I think it's pretty much already as affordable as they can afford to make it. The thing is huge.

I don't see what the point of selling it as cheap as possible would be - they would have less budget for R&D, to provide support, and to manufacture it. It's still in active development with improvements being made to it, it's not an abandoned product like many other scanning systems.

Sound isn't what they're focused on - remember it's a scanner - it doesn't have to do sound at all, that's just an optional feature.

5 hours ago, Perry Paolantonio said:

(and for Dan Baxter: we tested this on our Scanstation with and without NR, a Philips Shadow HD Telecine across town, and on AEO light. Happy to try other track readers, but we don't have ready access to any) 

Without getting into an argument over audio quality, an interesting thing of note is that Blackmagic's audio reader is Optical + Keykode + 16mm Mag all for $3.5K. That makes it incredible value really.

Link to comment
Share on other sites

10 minutes ago, Dan Baxter said:

Sound isn't what they're focused on - remember it's a scanner - it doesn't have to do sound at all, that's just an optional feature.

100%

And after buying their optional 16mm sound head, I'm even more convinced that it's primarily a visual scanner.  Thank God for AEO Light and iZotope RX (if necessary, often not).

  • Like 1
Link to comment
Share on other sites

18 minutes ago, Todd Ruel said:

And after buying their optional 16mm sound head, I'm even more convinced that it's primarily a visual scanner.  Thank God for AEO Light and iZotope RX (if necessary, often not).

Yeah I was actually going to say that iZotope RX is only $400, and that's what the professionals use to fix up the audio if they need to. I'm sure they could find a way to perfect the optical audio reader but that would take away from development on some other more important part that needs improvement (for example the new "wetgate pump), and then it would probably need to cost a LOT more than what they sell it for.

Link to comment
Share on other sites

On 3/26/2022 at 1:28 AM, Dan Baxter said:

Without getting into an argument over audio quality, an interesting thing of note is that Blackmagic's audio reader is Optical + Keykode + 16mm Mag all for $3.5K. That makes it incredible value really.

Wait a minute. You made this about sound quality. In case you forgot, these are your words: "To be fair optical audio off a Lasergraphics is unmitigated shit."

But once again, you are comparing apples to oranges. The Cintel sound reader (as I understand it) is essentially what you get in a projector: an exciter bulb and a photocell. Please correct me if I'm wrong. Because if I'm right, you could just as easily say Cintel is overcharging by thousands for $10 in parts. 

 

Link to comment
Share on other sites

13 hours ago, Perry Paolantonio said:

But once again, you are comparing apples to oranges. The Cintel sound reader (as I understand it) is essentially what you get in a projector: an exciter bulb and a photocell. Please correct me if I'm wrong. Because if I'm right, you could just as easily say Cintel is overcharging by thousands for $10 in parts.

No that can't be right - it reads KeyKode, you wouldn't be able to read KeyKode the way you're describing. It has an LED light according to the specs, and I would think it's a line-sensor imager. It's essentially a mini-scanner but one that only needs to be able to see black-and-white, that's why there's a difference in quality between different machines and designs - if the film is not perfectly flat you'll get distortion just like you'd get with the image on a Spirit Telecine or a Scanity if the machine has developed a wobble.

That's why I do not think that the LG audio reader is perfect - the image-stack has to be precise and perfect for the audio reader to be perfect (for KeyKode it can be less precise as it just needs to be able to read numbers and letters), and we're talking about a much smaller area of film where the margin for area increases exponentially compared to scanning the image that way on a Spirit. The best way to test whether the hiss is 100% from the film or not is to scan the same film multiple times and check the hiss - because if the audio is perfect then just like the video there would be no introduced noise and it would be exactly the same on each capture save for the minute differences accrued by the film's position relative to capture (but that won't affect noise or hiss).

Link to comment
Share on other sites

12 hours ago, Dan Baxter said:

No that can't be right - it reads KeyKode, you wouldn't be able to read KeyKode the way you're describing. It has an LED light according to the specs, and I would think it's a line-sensor imager.

This is incorrect and a quick look at the Cintel manual shows how it works:

781765200_ScreenShot2022-03-29at10_04_21AM.thumb.jpg.86df430a7eff53be4134bc1392b81b92.jpg

 

First, the soundtrack reader and the keykode reader are separate - you thread the film for one or the other through different readers. 

The optical soundtrack reader is in the upper left. the mag reader is upper right. The keykode reader is on the bottom of the assembly, below the two orange screws. This is all in the manual, from which I took this screenshot. 

There may be a camera for the optical reader but my guess is that it's using a traditional photocell that's illuminated by the LED shining through the track. This would make sense since that's how other Cintel scanners did it. This is far simpler and cheaper and given the price of this part, it seems more likely..  

Another clue is in the specs. The Cintel gives specifications for wow and flutter, which would mainly be a factor in a system where the sound is captured in real-time (or close to it, as the Cintel does) and would be subject to fluctuations in motor speed. A system that uses a line scanner to capture the frames can easily get around this in a couple ways, and wow and flutter isn't much of an issue. If it works like the scanstation, it uses the known position of the film to track and eliminate wow and flutter (see below). Wow and flutter simply aren't an issue on the ScanStation. In our recent tests, the Philips Spirit had major issues with this, due to film shrinkage and the motor constantly struggling to keep up. 

 

12 hours ago, Dan Baxter said:

That's why I do not think that the LG audio reader is perfect - the image-stack has to be precise and perfect for the audio reader to be perfect

Again, nobody is saying it's perfect, except you in your straw man arguments...

But the ScanStation's method is better. The Lasergraphics reader has a camera that is zoomed in on the soundtrack. The film is wound around a free-spinning chromed drum that effectively flattens the film out while the image of the track is being taken. The image captured by the soundtrack reader uses the edge of the film to align the soundtrack into a perfectly straight line, basically in the same way the scanner does horizontal alignment for Super 8 - using machine vision algorithms to "see" the edge of the film, and make sure the track is precisely placed before any sound is extracted from it. This is far better than any system that relies on rollers alone to guide the film through the reader. And in the case of the cintel, you have to manually adjust using thumbscrews on the outside, to position the reader over the track. 

 

12 hours ago, Dan Baxter said:

(for KeyKode it can be less precise as it just needs to be able to read numbers and letters),

Said with such confidence, and yet ...wrong.

Keykode numbers are there for us humans. Keykode readers read the burned-in barcodes that correspond to those numbers. We have a $20 USB barcode scanner we use daily for checking in and out customer materials. It's nothing fancy, the technology for that thing has been around for decades, and it's not expensive or exotic hardware. The Cintel uses a completely different reader for keykode, and it's likely a typical laser scanner like you'd have in a supermarket, only smaller, to read the barcode. Reading the numbers wouldn't make any sense. 

 

12 hours ago, Dan Baxter said:

The best way to test whether the hiss is 100% from the film or not is to scan the same film multiple times and check the hiss - because if the audio is perfect then just like the video there would be no introduced noise and it would be exactly the same on each capture save for the minute differences accrued by the film's position relative to capture (but that won't affect noise or hiss).

This is not a practical test and won't work the way you assume. In the ScanStation, the position of the film in the scanner is based on the frame (from the end user's perspective), though it is using a closed-loop servo motor which means it knows down to a fraction of a rotation where it is at all times. The current frame is the one in the gate, but that position varies. Because the ScanStation uses a constant motion transport with an oversized gate, and a camera that takes an image of the frame at high speed while it's in motion, the actual position of the film within the camera gate will almost never be in exactly the same place on two passes. This is by design, to allow for shrunken film. Basically, the frame is never in the same place relative to the edges of the gate, on each pass. And that's fine. There is a target location it tries to hit and if there's shrinkage, it may drift from that target one way or another over time. The scanner compensates for that by speeding up and slowing down the motors when necessary, in tiny increments. This affects the speed of the film through the soundtrack reader as well. 

The optical audio reader, however, is simply reading samples as image of the soundtrack goes by. The drum the film passes over has an encoder, so it knows where it is, and this is correlated with the captured samples so that the fluctuations in motor speed are irrelevant - preventing wow and flutter.

But to do the test you're suggesting would only work if you could match the two captured tracks down to the sample, which isn't a simple task. In any case, it's pointless, because as I said above, the soundtrack reader is a better design that ensures that

  1. the film is held flat when the soundtrack is read
  2. the film edge is tracked to perfectly align the track
  3. the encoder in the sound reader is used to correct for any fluctuations in motor speed, eliminating wow and flutter

The Cintel is doing what it's doing because by design it's a scanner based on an old telecine model - realtime capture - and everything is optimized for that. Based on the specs they give, I'm pretty certain that it's using a light and a photocell to read the soundtrack. That's why it can only operate within certain speed parameters, and why it needs a specially designed capstan to compensate for wow and flutter. 

I mean, on the Cintel you have to manually turn dials to adjust the position of the track reader for optimal results, and the film is not being held flat while the track is being read, it's happening in a roughly 6" gap between rollers. That's far from precise. 

 

 

Edited by Perry Paolantonio
Link to comment
Share on other sites

 Interesting I stand corrected on a couple of points there thanks. I should have realised KeyKode is just a barcode reader!

On 3/30/2022 at 1:50 AM, Perry Paolantonio said:

There may be a camera for the optical reader but my guess is that it's using a traditional photocell that's illuminated by the LED shining through the track. This would make sense since that's how other Cintel scanners did it. This is far simpler and cheaper and given the price of this part, it seems more likely.

Well THAT's a stretch. This isn't a Rank-Cintel part that was re-appropriated for the Blackmagic scanner like the Sondor sound reader -> DFT Scanity, that part was not even ready for sale when Blackmagic launched their Cintel scanner it came a few months later. Obviously they planned to have it from the start as it has to integrate into the film path and plug-in so the audio comes out, but it quite clearly wasn't ready when they launched the scanner as it wasn't for sale (the 16mm kit though was). You can check that yourself easily using the Internet Archive's Wayback machine. It's like how many of the Lasergraphics features which you've mentioned were added later on.

On the price - before you said it should be even cheaper if it's doing the audio the way you allege: "if I'm right, you could just as easily say Cintel is overcharging by thousands for $10 in parts." Really they're overcharging? The list price on the Lasergraphics 16mm mag sound readers is $6K each (there's one for the mag-strip and other one for full-coat). Arri's KeyKode reader does KeyKode and nothing else. FF charge €1.3K for each sound reader, so that's €2.6K for 16mm optical + mag (no keykode no 35mm ability) which is almost as much as Blackmagic's product. Also FF's is more basic and looks like it does audio the way a projector does it. So if Blackmagic was charging $6K for 16mm mag only they'd be on-par with the biggest competitor's pricing and they charge half that for a part that does mag + optical + keykode which is why I was pointing out that IMO the product is very good value.

The LG software does free optical audio extraction BTW so that's very good value too! I presume it didn't originally do that in 2013 right?

While there may be some cleanup applied - what you're alleging they're doing just isn't possible. You cannot remove hiss on-the-fly with a $3,000 part and a laptop operating it if the noise is 100% a property of the film and therefore different film-to-film. Try removing the hiss sometime in free software instead of iZotope Rx and see how far you get.

On 3/30/2022 at 1:50 AM, Perry Paolantonio said:

This is not a practical test and won't work the way you assume. In the ScanStation, the position of the film in the scanner is based on the frame (from the end user's perspective), though it is using a closed-loop servo motor which means it knows down to a fraction of a rotation where it is at all times.

That difference though doesn't matter - you can still do a comparison as you're just looking for deformity. To re-state what you've said in another way, the light might have tiny bit of uneveness from top to bottom, and maybe the lens is introducing a tiny bit of geometric deformity that is almost imperceptible - bit it doesn't matter if the film is 40 pixels in a different direction when photographed as those difference are so tiny that if you overlap the first scan and the second one the pixels will be *almost* identical. Then if there's scanner noise etc you can easily detect exactly how much of it there is. By the way the same thing would happen with a pin-registered scanner as well - the film won't be in *exactly* the same position, it'll just be closer to being in the same spot (instead of 40 pixels maybe it's 4 pixels or something of gate-weave) and require less effort to stabilise.

People have run this kind of test before - including with HDR on and off and checking different settings and how it affects the scan quality. For example on the website LaserGraphics claims of the RGB tool: "Built-in color grading tools for easy dye fade correction, applied during scan, eliminates secondary post-processing step." Everyone I know says that isn't true - the RGB tool is not suitable for dye-fade correction at all, it's very basic and only useful for a little bit of tweaking such as for dailies but using in the way described on the LG website degrades the scan quality and that you do need to do dye-fade correction after the scan using the proper software. It may be fine for a quick preview or a proxy, but it's not what you want for restoration. I know I'm not telling you anything you don't know, but I'm just pointing this out for the benefit of others because there are companies that will use the RGB tool exactly as described on the LG website as a shortcut.

Edited by Dan Baxter
Link to comment
Share on other sites

10 hours ago, Dan Baxter said:

Well THAT's a stretch.

How is that a stretch? it's how sound was read for many decades, including on some late model telecines. Blackmagic bought Cintel for a reason, so it's logical they'd have used some of that tech in their new scanner. As for the timeframe, you clearly have no concept of the work involved in designing a scanner like this. The BMD Cintel is based on late model Cintel scanners, but it's still a different machine. There is no way they could have taken a Cintel sound reader and plunked it onto the new scanner, because it's a completely different machine.

A ton of work goes into the design (mechanical and electronic) of these things, even if you're just modifying existing models. If they were behind on something and needed to get the scanner launched for, say NAB, dropping the audio module for a while until it's ready seems like a reasonable decision. Don't forget the cintel design team at BMD is/was tiny. Obviously they planned to include that  from the beginning, because the ports were there in the first release I saw at NAB in 2014. 

In any case we're both guessing so I've asked Blackmagic. We'll see what they have to say. 

10 hours ago, Dan Baxter said:

On the price - before you said it should be even cheaper if it's doing the audio the way you allege: "if I'm right, you could just as easily say Cintel is overcharging by thousands for $10 in parts." Really they're overcharging?

No. you're completely missing my point. I'm not saying they're overcharging.

You seem to be implying that because BMD is charging $3500 for their reader, Lasergraphics is overcharging for theirs. This conveniently ignores the fact that the Lasergraphics reader is a totally different design with a totally different way of doing things (I am almost certain) in a totally different class of scanners. As I spelled out above - there are significant differences in how the Lasergraphics reader is working that make it a better design, even if you take the camera vs photocell argument out of the picture. Better design costs more for a reason. 

I was just turning around your argument to point out that if BMD is not using a camera (as I suspect is the case), the parts involved to reproduce sound are cheap. Of course it costs a bunch of money to design, manufacture and sell these things, and I have no problem with the price they're charging. It's proportional to the cost of the machine.  I don't think they're overcharging.

 

11 hours ago, Dan Baxter said:

The LG software does free optical audio extraction BTW so that's very good value too!

In the current lineup, the Archivist is the only way to get the software reader because they don't offer a hardware reader for it. This was confirmed with Lasergraphics last week. The ScanStation comes with the hardware reader. You cannot run both the free software reader and the hardware reader on the same machine. 

That said, while it's a good value, it is subject to the same issues as AEO-Light is, in that warped film will result in worse sound than the hardware reader. This is in part because the hardware reader is holding the film flatter, but also because it's sampling at a significantly higher rate. 

 

10 hours ago, Dan Baxter said:

While there may be some cleanup applied - what you're alleging they're doing just isn't possible. You cannot remove hiss on-the-fly with a $3,000 part and a laptop operating it if the noise is 100% a property of the film and therefore different film-to-film. Try removing the hiss sometime in free software instead of iZotope Rx and see how far you get.

What exactly am I "alleging" here? Are you talking about blackmagic? A low pass filter isn't necessarily done in software, though it can be, and in software you can fine tune it to the specific audio you're working with. But if the BMD audio reader works as I expect it does, it's all analog and the filtering is done in hardware: light through the film to a photocell, which converts the light to a voltage. Hardware low pass filters for audio are not especially complex - they can be passive circuits inline with the analog audio signal. Here's a basic explanation of the concept.

A low pass filter over 7kHz for 16mm or 15kHz for 35mm (BMD's stated specs for optical audio reading) would eliminate the hiss you hear. Because 16mm sound is mixed to cut off above about 6kHz, and the BMD scanner's frequency response for 16mm caps out at 7kHz, they are likely applying a low pass filter to cut off everything above 7kHz, which would mostly be noise. But as i've now explained at least twice in this thread, there can still be signal above 6kHz, which this kind of filter would ALSO eliminate.

In other words: Would it remove the hiss? Of course it would. that's the point. Would it remove signal within that hiss that's above 7kHz? Also yes. 

Hardware cost for this filter? again, under $10, probably. 

 

 

Link to comment
Share on other sites

5 hours ago, Perry Paolantonio said:

In any case we're both guessing so I've asked Blackmagic. We'll see what they have to say. 

I am still waiting to hear back from BMD, but have spoken with an owner of one of these who has taken the unit apart. He's also someone who would know the difference between a phototransisitor and a camera. There is NO camera inside for the optical track reader, it works like a projector. Which checks out, as I outlined above, with things like the wow and flutter specs. And they are almost certainly using a hardware low-pass on the audio to cut off everything above 7kHz and 16kHz for 16mm and 35mm respectively. 

Link to comment
Share on other sites

Okay Perry I'm going to try and keep this brief.

I wish you wouldn't start with "you clearly have no concept ..." that's rude and I try to extend you better courtesy than that. Cintel was dissolved a decade ago and the rights sold to Blackmagic so explain exactly what you're talking about please.

18 hours ago, Perry Paolantonio said:

In the current lineup, the Archivist is the only way to get the software reader because they don't offer a hardware reader for it. This was confirmed with Lasergraphics last week. The ScanStation comes with the hardware reader. You cannot run both the free software reader and the hardware reader on the same machine.

I wish you would stop parroting claims made by LG to you. I do not doubt they told you this - what I am sceptical of is the truthfulness or accuracy of what they are saying there. In my humble opinion you shouldn't take what LG says (Steve? Stefan?) so literally and then be stubborn about it. What they may have meant is that you can only use one at a time which would make sense. Do you think a company buying their second, third, or fourth ScanStation would accept them crippling an important software feature on the newest one? Of course not. The software extraction works fine, it's free, and the hardware optical/keykode reader is not included in the base price of the scanner it's an extra:

z5LCTfI.png

From last year's base config. It would not make any sense for them to remove an important software feature - it gives their scanners a key competitive advantage over their competitors including Arri, FF, Blackmagic, DCS, and Kinetta.

As far as the Archivist goes - I suggest you get LG to send you the current configuration sheet so you know what you're talking about there. Here is a photo of one:

v8sN13D.jpg

That one clearly has the plastic rollers, but notice that the plate that attaches to the scanner so you can then attach one of the magnetic sound readers is not included in the base configuration, it's an additional product. So to get for example 16mm SepMag you need to buy the mounting plate product AR113 and the sound reader. As you can see from the photo, where the optical/keykode reader goes is just like on a ScanStation (they've even had to cover-up the hole for it) so it would be able to mount the optical/keykode reader's mounting plate if they wanted to offer that - but they don't. The film path is slightly different so that may be a factor I'm not sure, but it's more likely that most of the customers for those won't have much need for 16mm-only optical sound ability.

12 hours ago, Perry Paolantonio said:

I am still waiting to hear back from BMD, but have spoken with an owner of one of these who has taken the unit apart. He's also someone who would know the difference between a phototransisitor and a camera. There is NO camera inside for the optical track reader, it works like a projector.

Great! I'm corrected on my assumptions, and you should send him some sound film and then do a comparison against your ScanStation.

Edited by Dan Baxter
Link to comment
Share on other sites

First off, the image you're showing above is an early model. You've showed pictures of this one before I think. We've discussed this already: The Archivist has an optional MAG reader. There is NO optical track reader available for the archivist, it is software only. 

What you are showing in the picture above is an archivist that was built on a ScanStation deck plate (possibly due to supply chain issues? They are the same footprint in all other respects). Look at the picture on their web site. There is no square opening above the mag reader. This square opening is where the back of the optical reader enters the inside of the unit. The capped screw holes in the photo above are where the optical head attaches to the deck plate. 

THE CURRENTLY AVAILABLE ARCHIVIST HAS NO SUCH OPENINGS OR SCREW HOLES:

products-archivist-medium.jpeg.bce4383006dd2f673d03657f42996303.jpeg

 

The reader you see there is magnetic. It is not optical. HARDWARE OPTICAL AUDIO READERS ARE NOT AN OPTION ON THIS SCANNER. 

Direct from Lasergraphics, on January 5, 2022 when you first insisted this is available: "the Archivist does not have a mechanical optical sound reader (it is camera/software-based); it does offer magnetic sound readers" -- Look at it. Where does the optical reader go? It is a three-dimensional object that sits on top of the deck plate and extends into the chassis where the connectors are. There is nowhere to put this. 

You keep quoting an old price sheet. But man, I hate to break it to you: Now is not two years ago. 

 

 

Edited by Perry Paolantonio
Link to comment
Share on other sites

BTW, I'm happy to test the same track on a Cintel. The person I talked to sold his or he'd have done it. His assessment was that their ScanStation's audio was miles ahead of the cintel, and that the quality on the Cintel reader is "garbage". Which honestly doesn't surprise me given the manual adjustments (actual knobs) you have to tweak to align the reader and the track. Anything remotely shrunken is going to be a problem in that thing. 

But I'm game if someone has a sprocketless Cintel they'd like to run that film through. I wouldn't want to put that film through a sprocketed machine due to the level of shrinkage. 

Link to comment
Share on other sites

10 hours ago, Perry Paolantonio said:

First off, the image you're showing above is an early model. You've showed pictures of this one before I think. We've discussed this already: The Archivist has an optional MAG reader. There is NO optical track reader available for the archivist, it is software only. 

No I haven't, I just took that one off this website. The photo you're referring to there has been photoshopped. But obviously they may now be manufacturing the face-plate differently.

11 hours ago, Perry Paolantonio said:

The reader you see there is magnetic.

That's right, optical is software-only which I already said - you didn't seem to read what I wrote. What I said was that it would be technically possible to attach the mounting plate for the optical reader, nothing else. The film path itself is different through (only slightly different, but there's one less roller on the left side and I think the magnetic readers are mounted in a different position as well).

Again nothing I said is inaccurate - with the Archivist they do not attach the magnetic sound mounting plate if you don't buy it, whereas if you buy the ScanStation it comes with both mounting plates already there even if you don't buy the audio readers.

11 hours ago, Perry Paolantonio said:

You keep quoting an old price sheet. But man, I hate to break it to you: Now is not two years ago. 

Less than a year actually, and we're not at war Perry I just have a different opinion on some things than you do. You keep quoting the LG website when I've pointed out to you that much of what they claim on the website is false or misleading. Or you quote what they tell you as absolute truth. I'm fully aware that they tell different people different things - this happened last year. They said conflicting things to different people I know at exactly the same time (both of them were going through the process of purchasing scanners from them). When I say "conflicting" I do mean a direct contradiction on something quite important. So again on software audio extraction for the SS, no I do not believe they would disable that feature on new machines.

10 hours ago, Perry Paolantonio said:

But I'm game if someone has a sprocketless Cintel they'd like to run that film through. I wouldn't want to put that film through a sprocketed machine due to the level of shrinkage. 

The person/company operating it is more important than which one they have IMO. Just send something less valuable like a couple of trailers or something. Why don't you ask sometime on the Facebook group how people clean their P/T rollers - be polite about it and non-judgemental and I think you may be in for a bit of a shock with how different people/companies operate. If film slips a little bit against abrasive dirt it causes those cinch damage marks that are commonly heaviest at the end of projection reels (see here). So old slippery P/T rollers with dirt on them have the potential to damage film. With some scanners you don't have a choice not to thread through the P/T rollers - the Cintel included - hence one reason why how the machine is operated is important.

Link to comment
Share on other sites

As an example, this claim:

RDwQvz8.png

Which I've referenced to before, isn't just likely to mislead customers - it's anti-competitive because they're effectively stating "you don't need to buy a professional colouring suite, you can do everything you need with our in-built software". So any company that makes the professional post-production software products could make a complaint if they wanted against that claim IMO.

In a previous job I used to have a boss who would literally label things in the shop with false and misleading labels and whenever she was finished doing that and would leave I would have to go through and re-label everything correctly so that we weren't misleading our customers. One day she even told me not to do it "the customer will work that out for themselves" is quite literally what I was told. Yes I am serious, and no of course I didn't oblige.

So when it comes to referencing claims directly from LG in my humble opinion it would help a LOT if the claims made on their own website were honest and truthful. You CAN do a complete dye-fade correction as shown on the website, but it WON'T be to the same quality as if you did it in post-production with the correct tools. Fine for some uses perhaps, not what you'd want for professional work.

I noticed that you never said I was wrong about this point, but neither did you acknowledge it as a valid observation Perry.

Edited by Dan Baxter
Link to comment
Share on other sites

6 hours ago, Dan Baxter said:

No I haven't, I just took that one off this website. The photo you're referring to there has been photoshopped. But obviously they may now be manufacturing the face-plate differently.

Ok, whatever you say boss. Just for the record, every single product catalog photo you've ever looked at for anything (shoes, clothes, film scanners, food), has been photoshopped too. 

 

6 hours ago, Dan Baxter said:

That's right, optical is software-only which I already said - you didn't seem to read what I wrote.

You have said or implied many times in this thread and others that you can get an optical track reader for this machine. Two examples:

On 4/1/2022 at 3:58 AM, Dan Baxter said:

As you can see from the photo, where the optical/keykode reader goes is just like on a ScanStation (they've even had to cover-up the hole for it) so it would be able to mount the optical/keykode reader's mounting plate if they wanted to offer that - but they don't.

On 4/1/2022 at 3:58 AM, Dan Baxter said:

The software extraction works fine, it's free, and the hardware optical/keykode reader is not included in the base price of the scanner it's an extra:

I am pointing out that the image of the machine that is on their web site, WHICH IS THE MACHINE THEY SELL, is incapable of mounting the optical reader without major modifications that cannot be easily done. The image you are pointing to is clearly an early model built on a Scanstation deck plate. That's why they covered the hole you see. 

And for the record, no, it's not available as an extra for the Archivist. you're quoting an early price sheet you got from an Australian reseller. Contact Lasergraphics, ask them if you can get this, tell us what they say. They're the final authority, not your reseller. 

6 hours ago, Dan Baxter said:

Less than a year actually, and we're not at war Perry I just have a different opinion on some things than you do.

Well, there's the difference. You've got opinions, I've got facts. 

 

6 hours ago, Dan Baxter said:

You keep quoting the LG website when I've pointed out to you that much of what they claim on the website is false or misleading.

Show me an example of something that's false or misleading. Don't keep bringing up the Dye Fade Correction, because that, in fact, is a feature that exists on the scanner. It's called "Auto Color" and it's basically the same as auto color in DaVinici Resolve. We don't use it because it typically only works for a single shot and is a bit heavy handed, and it defeats the purpose of doing a flat scan. But I know several home movie transfer services do use it. it works by using the built-in scene detection in the scanner and when the scene changes it apples a new auto color correction for that shot. As you might imagine, this doesn't work perfectly. 

When I'm in the office I can post an example from a dye faded print with this on and off. It's a real thing that has existed on Lasergraphics scanners for as long as I can remember. 

 

6 hours ago, Dan Baxter said:

So again on software audio extraction for the SS, no I do not believe they would disable that feature on new machines.

I think this is the crux of the problem. You have "beliefs" about facts that are easily provable. I mean seriously? We have a ScanStation and you're telling me things are different than what I see in front of my own eyes *AND* have confirmed with Lasergraphics?

We have the latest software. The software soundtrack reader is not there. Does not exist. No software reader. None. not there. No options in the software for software reading. Lasergraphics tells me that if you have the optical reader, you do not have the software reader. 

But hey, why should anyone believe them? they only made the machine. 

Link to comment
Share on other sites

32 minutes ago, Perry Paolantonio said:

I am pointing out that the image of the machine that is on their web site, WHICH IS THE MACHINE THEY SELL, is incapable of mounting the optical reader without major modifications that cannot be easily done. The image you are pointing to is clearly an early model built on a Scanstation deck plate. That's why they covered the hole you see. 

And for the record, no, it's not available as an extra for the Archivist. you're quoting an early price sheet you got from an Australian reseller. Contact Lasergraphics, ask them if you can get this, tell us what they say. They're the final authority, not your reseller. 

Again Perry you didn't read what I said. The hardware Optical/Keykode reader is an option for the ScanStation not the Archivist, I never claimed otherwise. Also no what I'm referring to comes from Galileo Digital not Gencom.

35 minutes ago, Perry Paolantonio said:

Show me an example of something that's false or misleading. Don't keep bringing up the Dye Fade Correction, because that, in fact, is a feature that exists on the scanner. It's called "Auto Color" and it's basically the same as auto color in DaVinici Resolve. We don't use it because it typically only works for a single shot and is a bit heavy handed, and it defeats the purpose of doing a flat scan. But I know several home movie transfer services do use it. it works by using the built-in scene detection in the scanner and when the scene changes it apples a new auto color correction for that shot. As you might imagine, this doesn't work perfectly.

Okay we are WORLDS apart here. Let me put it this way, if I was the sales agent I would be very unhappy with the claims on the website.

38 minutes ago, Perry Paolantonio said:

We have the latest software. The software soundtrack reader is not there. Does not exist. No software reader. None. not there. No options in the software for software reading. Lasergraphics tells me that if you have the optical reader, you do not have the software reader.

And yet you literally have no evidence for this nonsensical claim! None whatsoever.

x3E47Hz.png

You know why you don't have it? My guess - you've got a 2013 model that was never licensed for the software sound extraction. It's included for new customers (at least up until 2021) and if you want to complain that you don't have it don't take it up with me take it up with Steve or Stefan or whoever you're talking to at LG. It would hardly be unusual for a new software feature not to be given to old machines for free though - as you've seen in the Arri price sheet you need to pay for the software in-scan stabilisation if you have an old one and it's free/included with a brand-new Arri XT.

FWIW unless there's been some alteration in the past couple of months, on the current ones all the software features are included for free except 2-flash HDR which you need to license. That is good value.

53 minutes ago, Perry Paolantonio said:

We have a ScanStation and you're telling me things are different than what I see in front of my own eyes *AND* have confirmed with Lasergraphics?

Message me in private if you prefer and I'll tell you exactly what their contradictory claims were - made mere days-apart to my mates. I also confirmed this "discrepancy" with a 3rd-party buying one at around the same time. So while I'm not claiming malice or anything like that, I can say that it was of significant importance. So far as I'm concerned that company tells different people different things and I've seen sufficient evidence to form that opinion.

Link to comment
Share on other sites

59 minutes ago, Dan Baxter said:

And yet you literally have no evidence for this nonsensical claim! None whatsoever.

I have two working eyes and a brain. And I also have email and text message confirmation from Stefan, the CEO of Lasergraphics and Steve, the owner of Galileo Digital, the reseller that it is either/or. You cannot have both, it is one or the other. 

1 hour ago, Dan Baxter said:

Again Perry you didn't read what I said. The hardware Optical/Keykode reader is an option for the ScanStation not the Archivist, I never claimed otherwise.

Ok, whatever you say. The record on this is pretty clear to anyone who wants to search this forum that this isn't true. I don't have time to dig up every claim but you have in fact said that the Archivist has a hardware optical reader, more than once. 

 

1 hour ago, Dan Baxter said:

You know why you don't have it? My guess - you've got a 2013 model that was never licensed for the software sound extraction.

No, your guessing wrong. We don't have it because we have the hardware reader. The software reader is software. It's part of the application and does not require a license, as it's part of the base module application. If you have a hardware reader, the software reader is disabled You get one or the other. Again, confirmed with Stefan and Steve separately. Why is this so hard to understand? It has nothing to do with model year of the scanner as it's purely a software option, and it is Lasergraphics' decision to give you one or the other. 

 

1 hour ago, Dan Baxter said:

It would hardly be unusual for a new software feature not to be given to old machines for free though

No, this is not how Lasergraphics does it's software updates. If it was there but not licensed, it would say as much in the license window where that stuff is listed. It does not. If you pay for a support contract you get the same software that runs on machines shipping from the factory today. The only features that are not necessarily enabled, are hardware based. For example, the original machines didn't have variable tension adjustment for the dancer arms. Without a hardware update to add that feature, you cannot use those settings in the software. But they're there. 

You are the one making unsubstantiated claims in a public forum. I'm not going to message you privately about this. You are spreading misinformation about something you do not have direct experience with, and in many cases you're making assumptions based on incorrect guesses as to how things work. Your claims are not credible, IMO, and I wish you would simply stop talking about things you don't know about. 

Link to comment
Share on other sites

Dan's posts are from a hobbyists point of view, they have very little relation to real world professional film scanning / restoration work - yet he thinks he's in a position to "school" @Perry Paolantonio about film scanners and film handling. It's mesmerizing that trolling exist in such a niche community.

Too much misinformation and guesses being shared here and on other forums, it makes professional people run away. Dan is not alone unfortunately.

Be careful who you take advice from - especially when it involves delicate film elements and expensive equipment meant to be operated by skilled professionals.

 

 

 

  • Like 2
  • Thanks 1
  • Upvote 1
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...