Competitive comparsion of an SD film transfer

The wide range of film telecine methods, equipment, internal workflows and post production practices have a direct impact on the visual outcome of a telecined home movie film. Lets take a look at the differences.

“If I were to transfer my once in a lifetime films to DVD, choosing a film transfer vendor could be a real gamble. Unless I assume they all produce the same results and all I have to do is shop on price… ” The consumer.

Introduction

Rumble House Media Group Inc.
May 2008 – June 2010

There seems to be hundreds of film transfer facilities all over North America who make various ‘best there is’ claims regarding their film to DVD transfer processes, while there are still others who shy away completely from saying anything at all on their web site, yet still claim they have the superior method without any substantiation. So what does this all mean? Do we have a level playing field?

From our viewpoint, much less the consumer, there is no means to compare telecine qualities even from a well documented competitor’s web-site with lots of selected film samples, because it’s still using their samples, methods and processes. Our web site is no exception. It’s not like going to Walmart and buying a pair of shoes where a direct comparison between competitive shoe vendors can be made immediately before making a commitment to buy.

In many instances, the examples shown on a good number of media transfer web-sites are not really representative of the actual condition of much of the film out there. One always seems to see the best frames or video clips, leading to a false conclusion that all film will look wonderful after using Vendor A’s process regardless of the current state of the film. The bulk of the film we get reveals a different story. Most of the film is quite average to poor in condition to say the least, with admittedly a scattering of gems in between. Lots of aging, poorly handled film and the effects of inconsistent lab processing – film grain, scratches, permanent “crap in the gate”, color casts, deep exposure flares, poor registration, focus problems, stripped sprockets, vinegar syndrome, mold and over\underexposed film segments. Welcome to the world of amateur film. One might conclude that given the sad state of the existing film body in the first place, why bother with any significant effort to recover those images and be satisfied with just the ‘spirit’ of the film. Cheap will do, right. That may very well be the attitude of some film transfer houses. But wouldn’t it be incumbent on a transfer house to preserve the quality of the film at hand using the best means possible, regardless of its current condition? After all these are peoples memories, paying good money and will probably never convert them again. Not an endearing legacy. But, lets leave that debate over my attached Facebook blog page -see below. You can beat me up there.

Original film condition aside, we wanted to see how we compare with the competition’s telecine process on the whole, so we instituted a comparative study whose goal is to determine our position in this industry – very good, good, mediocre, bad or sideways. We initially engaged in this effort more from a marketing research standpoint than anything else, but it has yielded much more than that in terms of other valuable technical information that will not be covered here. But we do want to share the results of what we found when comparing the different telecine methods.

The telecine method its own of course represents only part of the equation. The quality and performance of the image sensor plays a large part in any imaging system. Consideration in image sensor size, pixel density, pixel quantization and technology, the grade and properties of the glass – optics, sensor post processing and other aspects of the whole telecine chain affects the final outcome.

We have adopted the frame by frame telecine method of film transfer coupled to a 3CCD imager in a modified Moviestuff Sniper configuration. What does that say? How does this film to video method measure up to other processes, systems and technologies? To find out, we have assembled an 8mm film test reel that contains varying degrees of film quality in terms of grain, scratches, color casts and levels of light exposure. A vendor list was then composed and the same film reel was sent to each one in sequence with instructions on what to do and produce.

Each of these film qualities will stress the telecine system at hand end to end, yielding a final result unique to that vendor. We then apply some simple rules for evaluation purposes, the ones you can judge readily like visual presence and texture and the ones you aren’t readily obvious like effects of video horizontal line resolution, color rendition and degree of image cropping. All vendor returns have been referenced against selected film frames extracted from the same film reel that has been manually scanned using a professional Nikon 12MP camera (in uncompressed TIFF) and corrected in Photoshop to luminance, gamma, color saturation and hue levels that offer the image in its best light.

Note: Still frames scanned and processed this way, will produce the best results film conversion can offer given the existence of the ultimate film transfer machine. As such, comparing these optimized still frames against the faster exposed DV based video frames will seem unfair, so take the comparisons for what they are and make your own conclusions. At the same time be objective with what you see based on what is compared across all video frame samples for the evaluation test at hand.

Any test clips and still frames presented have been compressed for web page loading considerations, but they will still be good enough to see the differences between vendors.

disclaimer — Very Important to read before continuing

All assessments, comparisons and comments made in this report were done with objectivity in mind. It is not meant to downplay, ‘pump up’, denigrate or allude in any way as to what process is better, “good to have”, “could have been better” or worse than any other vendors process. However, any technical aspects to a transfer that is not so visually apparent, may or may not be commented on subjectively based on our technical experience and accepted qualities that define ‘good’ digital video.

Telecined film properties to examine

We will look at how faithfully the original film has been reproduced once digitized and processed to video by the respective vendor. We have used Windows DV AVI as the video format for these comparisons. DV is about 500 lines of horizontal resolution with a frame size of 720×480 pixels and frame rate of 29.97fps, color space 4:1:1. This ensures a more consistent foundation for comparison. Image frames will be extracted as they are from the video stream as close as possible to the frame to review (sometimes not all of the film has been captured and processed ??). All captured frames may have already had some color correction done and degrain filtering as part of the vendors post process workflow. Minimum compression has been applied for web display purposes.

Anomalies that can be introduced or controlled at film capture time – telecine cycle

• has image detail been preserved (assuming original images were sharp and in focus)
• flicker – a telecine method artifact
• hotspots – very bright unnatural whites, details lost
• focus – could be in original film sequences
• white\black balance – color temperature settings. original black IRE setting
• the milky look – poor pedestal or brightness setting
• cropping, tilt, zoom – film gate opening and alignment
• film registration – partial pre -post frames appearing top or bottom
• crap in the gate – intentional introduction of optical obstructions during telecine
• video banding – due to poor gamma settings at capture or during post process
• dynamic range – how well has detail has been preserved in dark or bright scenes
• light gradients – uneven light distribution over film frame
• optical system – quality of lens, camera resolution and exposure settings
• sensor technology – 3CCD, CMOS – dynamic range : 8, 10 or 12 bit pixel quantization

Anomalies that can be eliminated, reduced or ignored in post processes

• bad colors – correction for casts, aging effects, environmental effects, storage, poor film processing, questionable original film stock – best effort
• image contrast – luminance correction in an attempt to meet IRE standards white clip, gamma spread, gray scale linearity, 7.5 IRE black pedestal – for maximum richness and image depth
• film grain and scratches – natural property of aging film, environmental and/or abuse factor. Can be reduced using wetgate process or in post using sophisticated restorative software based image processing tools.

Anomalies that can’t be readily removed in pre or post activities

• crap in the gate – recorded right on the film itself at origination time
• random hairs and other gate obstructions – recorded on the film at origination time
• exposure flares – exposure to unintended light on undeveloped film
• film judder – film camera problems
• image jerkiness – steadiness of user at recording time. Can use “Steady” filter in post
• poor film quality – bad original film stock, poor lab processing
• embedded mold – ‘snowflakes’ in the emulsion

List of Transfer Vendors in Survey Abbrev – FBF = frame by frame; RT = real-time

Note: some vendors could not produce a DV AVI file for direct comparison, so the supplied DVD MPEG files were extracted and converted to DV AVI. The degree of MPEG compression may already have introduced some image degradation and not due to the transcode.

There will be five reference images used for comparison and 10 selected transfer houses compared as noted above:

• an image that has reasonable overall tonal qualities
• degree of cropping of the image and image detail – grain qualities,
• image with very dark components where details appear to be hidden
• image with very bright components where details appear to be hidden
• image with good degree of tonal quality along with small details

Rather than identifying each transfer house directly by name, they will be shown by a generic name like Vendor X. HomeDVD will put its name on its work. I don’t want to ruffle any feathers (there are two other vendors that offer the same quality or better than HomeDVD). This is an article to show the wide variety of transfer results that are out there. Pricing differences aside, the overall conclusion here is that not all transfer houses produced the same value in work for money that is paid.

The smaller gauge 8mm film format was used in this survey because it’s the most challenging to transfer. Aside from pushing the telecine system (lenses, imager, post processes and workflow), the physical film body itself is not the greatest in quality thus the many visual anomalies it carries would be tough to hide without applying some very special post processes – usually translates to $$$.

Some vendors have applied various degrees of filtering to the video to minimize the film grain and video noise, but in so doing have taken the edge off the video and thus making the video look softer overall. Sometimes this works well, sometimes not without using the proper degrain filters that leave contrasting edges intact while reducing the real spatial and temporal noise. All HomeDVD sample images are not filtered at all, but have been color and density corrected (film grain and DV mosquito noise mixing products will be evident).

If you are interested, double click on the image of interest and it will expand for a closer look. Click outside the image boundary and it will retract back.

Average Tonal Quality

Comparison images not available – problems with links Needs attention.

Image issues?????
*****************************************************************************************************