Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Does AJAs IO HD solve REDs poor down conversion from 4K to SD?

Carl Dieker

Well-known member
Joined
Nov 11, 2007
Messages
300
Reaction score
0
Points
0
Website
www.prospectdeck.com
I posted this question before at

http://www.reduser.net/forum/showthread.php?t=18227

but gave the thread a somewhat unclear title heading. Maybe this time I will get some replies....

Anyone? Since I soon will be doing some work for video artists whose initial delivery is SD I am really curious about AJA IO HD:s SD conversion capabilities when it comes to dealing with RED:s 4K resolution. At least I remember a thread last spring about the tedious testing by lots of users with different software and hardware solutions trying to down convert 4K to SD (a BMW clip) and as I recall it no one seemed to find a really satisfying solution for this problem. Now when we get 1080 4:2:2 and 4:4:4 (cropped) out of RED in playback mode and when the latest AJA driver does true progressive coding, 2995 $ for the IO HD seems a pretty good deal. And of course, in real time. Or?

:help:
 
Glad someone posted this question YET AGAIN. The fact is that, contrary to popular belief, going from 4k to SD DOES NOT look great. Why? Because of the TONS of aliasing you will be blessed with if you ever even think of burning that to a dvd. Mpeg streamclip as well as several other solutions have not helped one bit. Hopefully there will be a solution posted that doesn't involve endless, tedious experimenting. No one seems to have a clear cut answer and this down conversion issue absolutely killing my business.
 
All the SD Ive made looks fine, go straight from RAW- (Redcine) to 10bit 720x486 uncompressed. No problems here, proxies will always give horrible SD results... If this Is what you are talking about.

if you load your XML into Redcine from crimson I can make beautiful SD from the RAW, and center cut, letterbox, however you would like.

What exact delivery looks aliased and what workflow are you using?
 
I don't know how Redcine or RedAlert do on going down to SD, because I've always gone to 1080 HD for mastering, and then derive my SD from that. But the idea that Red can't go to SD cleanly is poppycock, and I'm tired of hearing it. Kona cards do pretty good, but I like more control, so I put a little gaussian blur at the HD (or higher) stage, then down-convert by scaling in Final Cut. It's as smooth as you want it to be; you can adjust the level of blur as needed to get rid of the aliasing while keeping enough sharpness. Sometimes I'll add a tiny bit of sharpening to the SD, the same as going from film to SD.

When I get a chance I'll check out doing the downconversion in RedCine, but I think it makes more sense to master in HD these days because it is so painless and doesn't require too much storage. Even if the deliverable is in SD, I like my master to be at least HD when shooting Red, so I can have something nicer for my reel, and it's more futureproof. And, since I'm doing the main color correction in HD, there are less relative artifacts than doing CC at the SD level.

Just my take on things . . . .
 
Its not RED giving poor result when downscaling.
Its the operator who doesnt unterstand the necessary technologies and working principles.

24P-29.97 is an art in itself, and judging the quality and characteristics of bicubic interpolators is not only debated, but often comes down to personal taste even among engineers.

Red downscaling looks perfect or bad - if done the exact way which is good, best or unsuited for the wished scenario. The same is valid for any solid high-res source (as 35mm scan or hdcam).

Regarding the question: The aja up/downconverters i have seen so far were good. Not as good as dedicated $$.$$$ AR/scale/FR-converters, but when opnly converting scale, they were good IMHO.
 
Red downscaling looks perfect or bad - if done the exact way which is good, best or unsuited for the wished scenario. The same is valid for any solid high-res source (as 35mm scan or hdcam).
.

Exactly.
 
Great. Yet another thread with posters claiming that there are no problems down-converting Red footage to SD.

Let me state flatly that the AJA down-conversion is unacceptable. The usual disclaimer; not a knock on AJA, just this specific issue. Anyone who says otherwise has zero credibility with me. Sorry, nothing personal, but the only advice I'd be likely to take from them is to duck in a gun fight.

Maybe those claiming acceptable software results are viewing on monitors that hide the problems. Hopefully it's not because they don't know a jagged edge when they see one.

I don't want to flame anyone, but anyone claiming that the problems exists only for the incompetent are just plain mistaken.

Down-conversion should not be some black art; it should be straight-forward and fast enough to be practical. No one, absolutely no one, has been able to demonstrate to me a satisfactory software down-conversion. Period. I had a lot of people claim to be able to do it, but none were able to produce results even approaching those obtained by dedicated hardware.

By the way, it's not just a need for speed; no software solution works. Really smart guys have suggested doing a bit of pre blur. Fine, if you want your footage to look worse than BetaCam original; it's not acceptable for me.

If you want to down-convert Red footage without jagged edges, the only solution I've found after EXTENSIVE search is hardware. End of story.

I apologize in advance if I've offended anyone; it's not my intent. I'd love to have a fast, effective software solution to a very vital requirement, but have found none.

Best regards to all,

Leo
 
Hi Leo, which particular hardware are you talking about ?
 
Great. Yet another thread with posters claiming that there are no problems down-converting Red footage to SD.

Let me state flatly that the AJA down-conversion is unacceptable. The usual disclaimer; not a knock on AJA, just this specific issue. Anyone who says otherwise has zero credibility with me. Sorry, nothing personal, but the only advice I'd be likely to take from them is to duck in a gun fight.

Maybe those claiming acceptable software results are viewing on monitors that hide the problems. Hopefully it's not because they don't know a jagged edge when they see one.

I don't want to flame anyone, but anyone claiming that the problems exists only for the incompetent are just plain mistaken.

Down-conversion should not be some black art; it should be straight-forward and fast enough to be practical. No one, absolutely no one, has been able to demonstrate to me a satisfactory software down-conversion. Period. I had a lot of people claim to be able to do it, but none were able to produce results even approaching those obtained by dedicated hardware.

By the way, it's not just a need for speed; no software solution works. Really smart guys have suggested doing a bit of pre blur. Fine, if you want your footage to look worse than BetaCam original; it's not acceptable for me.

If you want to down-convert Red footage without jagged edges, the only solution I've found after EXTENSIVE search is hardware. End of story.

I apologize in advance if I've offended anyone; it's not my intent. I'd love to have a fast, effective software solution to a very vital requirement, but have found none.

Best regards to all,

Leo


Thank you Leo for another honest post concerning to this utterly maddening dilemma. Look I use the Red as much as anyone else here. WHY IN THE WORLD should this be so difficult an issue? Nobody wants to degrade these wonderful images that come out of our Red cameras, and yes, I understand that the images in SD wont we nearly as good as the original 4k images. What I don't see is why there isn't a simple fix to this. An AJA Kona Card? OK, but For Standard? I could understand if I were going out to Bluray even HD-DVD, but WHY should I invest thousands more for STANDARD? Ok let me make is SIMPLE, for all the experts who don't understand what I mean or why I'm complaining: HOW DO I GET RID OF THE UGLY JAGGED LINES IN SD? it's not an art, it shouldn't be rocket science, it's a simple question. Can somebody finally answer it?
 
I will be making tests with the AJA IOHD in one of the coming days. The plan is to see how the quality holds up when outputting 1080p directly from the RED and onto the IO HD into both SD and HD. I have a setup that will reveal any downscaling issues - and I'll post the results for everyone to judge for themselves.
 
Great. Yet another thread with posters claiming that there are no problems down-converting Red footage to SD.

Let me state flatly that the AJA down-conversion is unacceptable. The usual disclaimer; not a knock on AJA, just this specific issue. Anyone who says otherwise has zero credibility with me. Sorry, nothing personal, but the only advice I'd be likely to take from them is to duck in a gun fight.

Maybe those claiming acceptable software results are viewing on monitors that hide the problems. Hopefully it's not because they don't know a jagged edge when they see one.

I don't want to flame anyone, but anyone claiming that the problems exists only for the incompetent are just plain mistaken.

Down-conversion should not be some black art; it should be straight-forward and fast enough to be practical. No one, absolutely no one, has been able to demonstrate to me a satisfactory software down-conversion. Period. I had a lot of people claim to be able to do it, but none were able to produce results even approaching those obtained by dedicated hardware.

By the way, it's not just a need for speed; no software solution works. Really smart guys have suggested doing a bit of pre blur. Fine, if you want your footage to look worse than BetaCam original; it's not acceptable for me.

If you want to down-convert Red footage without jagged edges, the only solution I've found after EXTENSIVE search is hardware. End of story.

I apologize in advance if I've offended anyone; it's not my intent. I'd love to have a fast, effective software solution to a very vital requirement, but have found none.

Best regards to all,

Leo

Leo,

I'm not offended at all; I appreciate where you're coming from. Here's where I'm coming from:

I've been doing high-end post since 1991, and have worked on most every platform available from that time until now. Some do scaling very well; some not well at all, and I'm talking about both hardware and software. My favorite scaling algorithm of all time is still the DF/X Composium, a dedicated hardware/software platform from the early nineties. I also like some of the interpolation algorithms in the Nitris, which I used to own. It's all about the algorithm, and that's true for software AND hardware. Digital hardware is also running software to accomplish its task.

You obviously want a solution that just does the scaling perfectly with no adjustments, no user input required. I agree that the AJA downconversion is not up to snuff, and that's why I do it my way (see my earlier post). One challenge with scaling, whether done in dedicated hardware or in software, is the balance between "sharpness" and "smoothness" (lack of aliasing). It's the same issue that Graham talks about with the OLPF on the Red. Red goes for something more smooth, while Sony goes for something more "sharp", including the accompanying increased aliasing. Take something shot on film with a Master Prime or one of the Panavision Primos that has a lot of fine almost-horizontal lines in the image, transfer it to SD in a telecine. At the telecine stage, if you leave it at the full focus on the film, those almost-horizontal fine lines are going to be very aliased. One way to get rid of the aliasing is to defocus the lens shooting the film very slightly - it won't look soft (because you still have plenty of detail for SD even with a little defocus), and it won't be aliasing. You choose how much based on the image content. I do the same in FCP with HD to SD downconversions, to help out FCP's interpolation a bit. And don't tell me it doesn't work with great results, because it does. I can put the balance between sharp and smooth wherever the footage looks best. I've been doing this at a very high level long enough to know what is good and what isn't.

To summarize, the question of downconverting Red footage to SD is exactly the same challenge as downconverting other HD and above formats, and film itself. Like you, I would like better interpolation algorithms in software solutions, but have found a workaround that works well.

Some random notes:

-The scaling quality of some programs like FCP and Motion varies widely depending on the version. When FCP 6 came out, it was a huge step backwards in scaling quality. I and others screamed bloody murder, and after several version upgrades, it's gotten better. It's still messed up depending on whether your footage has a interlacing tag on it or not; this can be changed to none in the browser window and it works better. Yeah, it sucks that it's not straightforward like it should be, but it absolutely is possible to get good results.

-You stated that the only acceptable scaling you've found is hardware. Which hardware are you referring to? Just curious; I'm always interested in hearing what is working for people. I've seen hardware downconversions that are aliasing like a mofo; it's not just software.

-You also mention "fast" in your desired requirements. Always a need, isn't it? My downconversions go faster than realtime on my Mac Pro (recent 8 core), so that's better than running footage realtime through a box.

-I understand what you mean about monitoring. For monitoring SD, a real CRT is an absolute must (BVM preferably). I also check things on LCD's, Plasma's, and so forth, but it must look good on a pro CRT to pass. Interlacing changes everything when viewing originally progressive footage, so any conversion needs to hold up to that.

-I am starting to see more functionality in software encoders that tackles the problems of aliasing in downconversion. Telestream's Episode Pro, for example, has a "preprocessing lowpass for large downscales" option in its resize menu. It does a pretty good job much of the time, but sometimes it's not enough or it's too much; I prefer adjustability.

I hope I haven't offended you either, and if you've tried what I suggest and still can't get what you want, we'll just have to agree to disagree. I never said that there were no problems downconverting Red footage, I just said that it's the same issue as downconverting other HD formats and film itself (in the digital realm), and that it IS possible to get good results with FCP. It's just an extra two steps, unfortunately. The AJA downconversion is okay for quick monitoring, but I'm not satisfied with it as you aren't and have taken matters into my own hands. AE unaided is not good enough IMO; Motion is not bad. Flame's interpolation is adjustable, so choose your poison, whether sharp, smooth, or some sweet spot in between.

It would be nice sometime to look at the same monitor as you at the same time and discuss and compare.

Best Regards,

Matthew Verkler

P.S. My gunfight advice is to return fire while moving to cover. :-) And be caught up on your prayers beforehand because there won't be time for that . . .
 
I will be making tests with the AJA IOHD in one of the coming days. The plan is to see how the quality holds up when outputting 1080p directly from the RED and onto the IO HD into both SD and HD. I have a setup that will reveal any downscaling issues - and I'll post the results for everyone to judge for themselves.

I'm sure we all look forward to your tests!
 
Thank you, Matthew, for your post.

Here's my view of why Red presents a down-convert problem that other formats do not suffer.

When we shoot 35MM film, the footage is "down-converted" either in telecine or via the HD deck's built-in hardware. Both produce very good results. In other words, no problem exists; we get alias free footage in real time and can then edit with no problems.

Shooting traditional 2/3" HD is equally seamless; the decks handle the down-conversion beautifully.

Red on the other hand, presents a serious problem to those who don't possess HD decks or dedicated, and relatively expensive, hardware.

This comes as a most unpleasant surprise to many Red producers; they shoot, edit beautiful HD images, then attempt to down-convert with an AJA card or software such as FCS or Compressor with unacceptable results.

To date, no software solution has been found to be acceptable. AJA Kona can't cut it, Scratch can't do it; Terranex works, HD decks work. Other hardware may work as well, but I've been very disappointed with a couple of them. Out of the box, test, pack it right back up and return. No, I won't name them.

Perhaps Red's new format which scales more directly to 1080 will achieve acceptable down-conversion via software. Just speculation; I'm not an engineer.

I'd love to see Red address this issue, either with a software solution that works, and works fast enough to be practical, or a hardware device.

I'm using the HD deck solution which works well.

Best regards,

Leo
 
This comes as a most unpleasant surprise to many Red producers; they shoot, edit beautiful HD images, then attempt to down-convert with an AJA card or software such as FCS or Compressor with unacceptable results.

Have you tried SD conversion using RedRushes? If you do all of your conversion into SD using RedRushes, it looks fantastic. Very clean. It has some limits currently (like no audio throughput), but they seem to be working those out.
 
Leo, maybe it should be stated again, what is the problem with SD downconversion: guys, it is aliasing! That doesn't mean the downconverted footage looks bad, unsharp or anything - but it has alias flickering.

I did some tests with the sample footage Leo provided in the other thread and came to the same conclusion as he does. I used some different algorithm for the downscaling like Lanczos and Precise Bicubic but none of them cut it.
Hence I will try some other when there is time like Spline-64.

I think the best solution would be some kind of anti-aliasing filter afterwards, or a rolled off low-cut filter - which would behave like a natural fall off.
 
hi guys,
i just finished the 1080p online of my new short shot with my RedOne at 4K. Before mastering to HDCAM SR, i exported a 1080p movie file (apple pro res HQ) and when i took the movie file and put in a new SD sequence in FCP and sent it to compressor to make a SD DVD, the film looked horrible, just as you guys are saying, alot of aliasing etc... but then, instead of making a SD sequence in FCP, i just took the exported 1080p movie file (proresHQ) straight into compressor (without going through FCP) and made a SD file... and by God it looked great!! no aliasing on the dvd at all!! i don't know if this helps, but i may be an easy workaround...
cheers
 
Leo,

Thanks for your follow-up. To summarize what you're saying, other formats have successful downconversion capability as inherent parts of their workflow, whether that be film and telecine, or HD and a deck's internal downconversion or Terranex, and that Red does not have this built-in to the workflow. I'll buy that for a dollar.

Yes, there is a need for better tools in software and/or hardware to handle this requirement. I'm with you on that.

I was responding to the idea that Red footage itself somehow was not able to be downconverted to SD properly, which clearly isn't true. Red does have a bigger challenge than HD, as it starts out as having much higher resolution, and so aliasing becomes more likely as we downconvert to SD.

With your current HD deck solution, how are you getting to the HD stage? RedCine? RedAlert? I haven't had aliasing problems going to HD size in RedCine, using the Sinc filter. From there, it's the same problem as going from any HD format to SD, which you successfully use a deck to accomplish. For the software tools, it's certainly not as straightforward. Compressor, for example, won't look good on many types of footage unless you turn on "Frame Controls" and use the "Best" resize filter. As default, frame controls are turned off, and if you do use the Best resize filter, render times increase dramatically. And even then it might not be good enough.

I guess I've gotten so used to using workarounds that I don't think about it consciously too much. In the old days (of pre-Flame early nineties), workarounds were how we accomplished pretty much all cool stuff. You are right to ask for better tools for downconversion; the world shouldn't have to suffer these medieval conditions! I'm not being facetious when I say that; I complained to manufacturers in the past, but then just gave up and used workarounds.

By the way, I hope you're not in Gustav's path, and if you are, stay safe and good luck!

Best,

Matthew Verkler
 
HI
I get a lit bit confused.Please help me. Is AJA KONA 3 Hardware down/up conversion or not?

Is from the AJA website:
"KONA 3 also includes a variety of 10-bit broadcast-quality features, such as hardware-based up-, down-, and cross-conversion to and from HD, and adds a live hardware HD/SD keyer for compositing bugs, live clips, and other elements over video."

 
Leo, maybe it should be stated again, what is the problem with SD downconversion: guys, it is aliasing! That doesn't mean the downconverted footage looks bad, unsharp or anything - but it has alias flickering.

I did some tests with the sample footage Leo provided in the other thread and came to the same conclusion as he does. I used some different algorithm for the downscaling like Lanczos and Precise Bicubic but none of them cut it.
Hence I will try some other when there is time like Spline-64.

I think the best solution would be some kind of anti-aliasing filter afterwards, or a rolled off low-cut filter - which would behave like a natural fall off.

Sometimes I use FCP's flicker filter on minimum to get the last bit of aliasing, but it does soften the picture more than I like. A "rolled off low-cut filter", otherwise known as slight blur, should be applied on the HD image BEFORE scaling down; doing it after just makes everything blurry. I sometimes add a little sharpening on the SD image, as would be done in any telecine or in camera with SD cameras, because SD needs a little punch, especially with talent's eyes in wider shots.

Good interpolation (the algorithm for scaling) eliminates the need for these techniques, but good interpolation in software is hard to come by, unfortunately. As Leo has lamented.
 
Back
Top