CompuServe Thread

#getting to video tape

13 messages in this thread
#113106From: George OmuraJul 15, 1994 3:54 PM
Hi there, I've read the replies to your question, just to let you know where this message fits in. I recently aquired the DPS Par board and I love it. The quality is very good. Previously, I had rendered to a 3/4 Umatic (Industrial quality video) and the results were good. The PAR board makes Umatic look like a home movie. Keep in mind, however, that quality is really quite subjective. I've "crash recorded" stuff right off the PAR board onto Betacam SP decks and the results are quite good. The video editor I work with thinks the quality is great. I haven't put a wave form monitor or vectorscope to the output of the PAR, but since it's comming from DPS, I'd think the signal would be good. Understand that the PAR board will do simple cut and splice edits. It won't let you do fancy fades and wipes between scenes. Use video post for that stuff. Also, to get the best results from the PAR board, you have to render to a resolution of 720×480. In fact, this resolution is part of the reason stuff looks so good off the PAR board. Using my old method, rendering for a Targa + for step-frame animation, the resolution was 512×480. Also, the higher resolution tells you that the output bandwidth of the PAR board is quite high.
#113172From: Don LandisJul 15, 1994 11:26 PM
Your observations are identical to mine. I also once used a targa+ to Umatic and the 50% increased resolution, lower signal to noise, increased K factor quality of the PAR over the targa+ more than makes up for the compression artifacts that some on this forum claim exists but few have ever seen. I use the PAR output betacam SP CAV and the quality is running rings around some output I recently saw done on my competitor's SGI/softimage/ wavefront system. My demo quality won over the Abekas output. Now don't get me wrong on this. I think the stuff rendered with the SGI was rendered to frames, not fields and the colors were washed like he had a gamma set to 3 or so. Bottom line is my 3DS/PAR stuff was just plain supperior in appearance. One thing you do need to adjust though is your resolution on the PAR. It should be 752X480, not 720X480. Take care and keep in touch. Don Landis
#113600From: George OmuraJul 19, 1994 11:50 AM
Thanks for the corrx. on resolution. I'm interested in your comment comparing the PAR to the Abekas output. Do you mean it's actually better? Equal? I was thinking of putting together an Abekas subsystem (Exabyte tape, ASDG software etc.) for high end stuff, but if you're saying what I think you are saying…. G.O.
#113665From: John TissavaryJul 19, 1994 7:38 PM
The PAR (and the MAX, for that matter) does not compete in output quality with the Abekas machines. If you're talking industrial or cable, then its a different story, but for mainstream broadcast the consensus seems to be that jpeg and wavelet (ImMIX CUBE) compression is usually not acceptable. John Tissavary (LUNA cie)
#113714From: John TissavaryJul 20, 1994 2:57 AM
>>The PAR (and the MAX, for that matter) does not compete in output quality with the Abekas machines<< I should have said "does not compete directly" – sorry for any inconvenience this has caused <g> John Tissavary (LUNA cie)
#113820From: Don LandisJul 20, 1994 1:54 PM
>>but for mainstream broadcast the consensus seems to be that jpeg and wavelet (ImMIX CUBE) compression is usually not acceptable. Have you heard anything about CBS buying bunches of the Imix Cubes for producing these tabloid shows? Seems what these broadcast guys say and do are two different stories.
#113889From: John TissavaryJul 20, 1994 10:13 PM
I'd recieved some info from an editor at a TV station in LA who'd been woed by ImMIX and duly impressed. He really loved the unit, and said it definitely had a place in the broadcast world, but probably not for finals on network stuff. I think the ImMIX gets considerably less compression than the PAR or MAX, though, because it uses a SCSI array and some proprietary hardware to do the wavelet thing. It also lists for over $40K. John Tissavary (LUNA cie)
#113992From: Don LandisJul 21, 1994 10:21 AM
>> It also lists for over $40K. The base unit will set you back $65000. while you can sit the machine up for 40 it won't be capable of much. The figure I stated is what my quote came in at last winter for a working productive system.
#114065From: John TissavaryJul 21, 1994 4:25 PM
I guess 65K is with the extra disk arrays or something, and the larger monitor. I was pretty shocked at the crappy monitor you get with this system, and ImMIX claims it's to "keep costs down" – yeah, right! <g> John Tissavary (LUNA cie)
#113999From: George OmuraJul 21, 1994 10:49 AM
Thanks for the reply, John. I'm wondering if anyone has done an "ACID TEST" comparing the various compression schemes vs non-compressed video. It would be interesting to see the differences side by side.
#114066From: John TissavaryJul 21, 1994 4:25 PM
I don't know if there's really a way to quantify these sort of compression schemes, since they really do different stuff to every different image they compress. I mean, you can know what the compression ratio is, but the artifacting will be different on each image. So it's really a judgment call, IMHO. Like Don says, it's never come up for his clients. It has for one of mine, but not for another. I would do some demos on the units, show them to clients (don't tell them what equipment you used), and if they complain, then go with Abekas or single frame on that particular job. In any case you can't really go wrong with the PAR or MAX, because just their preview capabilities will increase your production BIG TIME (they did mine, at least). John Tissavary (LUNA cie)
#114075From: George OmuraJul 21, 1994 7:16 PM
Thanks for the advice… keep in touch g.o.
#113819From: Don LandisJul 20, 1994 1:54 PM
>>, but if you're saying what I think you are saying…. What I'm saying is that in my recent experience AP the PAR to betacam SP has satisfied all my clients that were looking for "beta SP quality" My competition that uses an Abekas A-66 has put out some (in my professional opinion) pretty lousy quality renderings from his SGI. This was indeed disappointing for me but knowing animation like I do I could see that the quality was probably due more to the improper use of his hardware than anything else. I will say that if your clients are real picky on detail and quality the next step up from the PAR would be the Abekas output. This is my decision here and if I ever do run into a customer that is not satisfied with the PAR I will go this route. One other thing to remember about the PAR: The PAR does NOT do the same quality on frame grabbing as it does on simple animation. QF of 12-14 are common for rotoscoping frame grabs but QF of 23, BLF 250 is the rule for animation. This means that animation incorporating rotoscoping will be of lesser quality in the rotoscoped portion than simple animation. The reality is I've done a few jobs of rotoscoping and a few slomo and time lapsed jobs all requiring frame grabs at reduced QF and even these jobs the artifact question never came up.