Compression Artifacts! What Compression Artifacts?

limo991 said:
I like the disclaimer. I wouldn't be able to that even if I I tried:) Lets keep comparisons realistic.
Click to expand...

Was this or was this not about HDV artifacts? The FX1 shots are HDV artifacts. How's it not realistic?
 
The frames I posted are normal everyday frames and that's what I see on HDV. Your frame has an explosion and actually compares to HVX in entirely different exposure. That's what I meant. It is a HUGE exageration! The cameras are very close in typical situations.

If you had only posted the Fx1 frame I bet nobody would say anything.
 
They are very close in normal situations.

Did I or did I not say that this was HDV at its VERY WORST? Did I or did I not say that this was the most artifacted footage anyone would ever see? Did I or did I not say that HDV doesn't normally look this way, but this is what can happen in a worst-case type of scenario?

I'll dig up the original .m2t files.
 
limo991 said:
So, the quality dropped for a few frames. The area on the right is entirely black on the HVX so we can't really compare anything. Can we have an FX1 frame without an explosion or an HVX frame with an explosion?
Click to expand...
You can have the entire FX1 stream. I'll dig it up and upload the whole thing. So you can see every frame with and without any explosions and everything.
 
limo991 said:
If you had only posted the Fx1 frame I bet nobody would say anything.
Click to expand...
Simple enough to fix. I removed the HVX picture and instead put in a shot of the FX1 without the macroblocking. I found a spot in the stream that looks comparable to that HVX shot and substituted it out.
 
HDV is compressed, its gonna have some artifacts during high motion no doubt about that, questions is how is it viewed by non-techs like Joe Bloggs thats more important

And if you are desperate to avoid the compression you have two options:

(1) Connect the HD component into a Black Magic Card and record to a PC, then transcode and work in almost any format you want, this workflow does work, check out the Music Video done by a User over at DVinfo.net

(2) Get the HVX200, can't beat its format, end of story

Anhar
 
Barry, is it possible that the HVX example is 1:1 pixel but the FX1 sample is zoomed in by 200% or something? The Fx1 frame is already much larger, so why zoom? The FX1 actually compares with the HVX at similar lighting.

Here is a part of the frame from the FX1 clip at 1:1 pixel:

 
limo991 said:
Barry, is it possible that the HVX example is 1:1 pixel but the FX1 sample is zoomed in by 200% or something? The Fx1 frame is already much larger, so why zoom? The FX1 actually compares with the HVX at similar lighting.

Here is a part of the frame from the FX1 clip at 1:1 pixel:

Click to expand...

That looks way more close to what I was thinking. I was expecting the HVX to win over the Fx1, but it made the fx1 look lik an old Hi8
 
It's a tricky clip. There is total light change every couple of frame from a strobe light combined with pure chroma parts (blue) and camera instability and HDV cannot code those. Not something you would shoot everyday:) I would love to see even a compressed version of the HVX. Mostly for seeing how it overexposes. I know the detail on strobe light will be better than HDV.

The first example from Barry, "before macroblocking" is actually blocking too much already. It is a matter of choice of frame. Under constant light and light motion, even with 4:2:0, the above is what the FX1 would look like. With the strobing I probably got one of the best frames. I frame or something:)

It is interesting to see this clip on slow motion. You can see how HDV works since this is a worst case scenario for long GOP.
 
Last edited:
limo991 said:
Barry, is it possible that the HVX example is 1:1 pixel but the FX1 sample is zoomed in by 200% or something? The Fx1 frame is already much larger, so why zoom? The FX1 actually compares with the HVX at similar lighting.
Click to expand...
I extracted the frame at pixel-for-pixel basis. But, they were upsampled to 1920x1080. Maybe that's the difference you're seeing.

You've got the same frames now that I had, so you can extract the same ones. No zooming was done. But they were aspect-ratio-corrected.
 
myfriendimage said:
That looks way more close to what I was thinking. I was expecting the HVX to win over the Fx1, but it made the fx1 look lik an old Hi8
Click to expand...
That's also what I was trying to point out with the HVX image -- this is what it's supposed to look like. I guess I just did the wrong thing by showing an HVX frame of it.

Under ideal circumstances they look very close. When things go wrong, they do not look close.
 
limo991 said:
I would love to see even a compressed version of the HVX. Mostly for seeing how it overexposes. I know the detail on strobe light will be better than HDV.
Click to expand...
I can find that clip. I'll root around on the Mac and see if I can find where it is.

The problem you'll run into is I think the white balance switch got toggled on the HVX so it's the wrong colors. Everything's a lot oranger on the HVX version of the clip.

I had both cameras strapped down with velcro straps on a Matthews 9" monitor platform, set on auto-iris and held above my head so I couldn't even aim properly but I had to try to get the shot over the crowd. Not exactly ideal cinematography technique!
 
limo991 said:
The frames I posted are normal everyday frames and that's what I see on HDV. Your frame has an explosion and actually compares to HVX in entirely different exposure. That's what I meant. It is a HUGE exageration! The cameras are very close in typical situations.

If you had only posted the Fx1 frame I bet nobody would say anything.
Click to expand...

Your new comparison is better, but to help understand this thread, I think it is ok to post your original in again.
http://static.flickr.com/41/100485297_c3c37b5c19_o.jpg
 
Last edited:
Barry. Thanks for all your info and help in regard to my artifacting question. I can definatly see where one could fall into potential problems.

Kind Regards
 
xray said:
Your new comparison is better, but to help understand this thread, I think it is ok to post your original in again.
http://static.flickr.com/41/100485297_c3c37b5c19_o.jpg
Click to expand...
Fine -- but would you also care to add the text I had when I posted that picture? Where I said specifically that it was to show what the detail was supposed to look like, and that I didn't try to find a matching frame but instead I looked for one where there was enough light to show the detail?

Or would you rather just show it out of context?
 
So you also did changed your text? I am not responsable if you change pictures or text after 20 postings -dealing the original image and text you posted-, but if you want (need it) I have a copy. Just to keep it clear I reposted the image otherwise it is hard to follow this discussion. For me, there is no need to discuss this anymore.
 
You must log in or register to reply here.
Back
Top