Video clip comparisons: JVCs, Pan AG-1980, Mitsubishi

[moxiecat]

I have a tape that needs to be transferred to DVD for a local organization. To do the job, I first went to my trusty JVC HR-S9911U. When the tape played with occasional jitter, I tried it on the new Mitsubishi HS-HD2000U I recently purchased (at Orsetto's recommendations). The difference in quality was surprisingly noticeable for two machines that were supposed to be similar. So then I decided to try the tape on my Panasonic AG-1980 and a used JVC HR-S9600U as well.

I'm hoping that the pros here on this forum might be able to look at these clips and assess them. I am mainly interested in how the different VCRs' picture modes affect the playback comparatively.

Some background information: The tape is a professionally-made VHS in good condition. The video clips are each about 12 seconds long, and show the tape's opening scene as well as a few seconds of the title screen. The TBC/DNR is on for all of the VCRs except the Panasonic, for which the TBC made the picture noticeably worse. The encoding was done with a Toshiba RD-XS35 recorder on the 7.8 bitrate setting. No filters are turned on in the recorder.

Here are the clips:

Mitsubishi HS-HD2000U in NORMAL (no processing): http://www.oneworldonehealth.org/video/mitsu_hshd2000u_normal.mpg
Note: This VCR has Normal, Soft, and Sharp modes, as well as an "auto" which apparently chooses from the three options.

JVC HR-S9911U in EDIT (no processing): http://www.oneworldonehealth.org/video/jvc_hrs9911u_edit.mpg

JVC HR-S9911U in NORM: http://www.oneworldonehealth.org/video/jvc_hrs9911u_norm.mpg

JVC HR-S9600U in EDIT (no processing): http://www.oneworldonehealth.org/video/jvc_hrs9600u_edit.mpg

JVC HR-S9600U in NORM: http://www.oneworldonehealth.org/video/jvc_hrs9600u_norm.mpg
Note: At the end of the clip is a sample of my "blue screen" from my 9600. It has a lot of weird patterns which my other units do not show. This is NOT apparent on the playback picture. But is this something that shows an error in the machine?

Panasonic AG-1980 in EDIT (no processing): http://www.oneworldonehealth.org/video/pan_ag1980_edit.mpg

Panasonic AG-1980 in NOR: http://www.oneworldonehealth.org/video/pan_ag1980_nor.mpg

My quick assessments:
- The Mitsubishi has an amazingly noise-free picture, but I do see some weird brightening/darkening going on. Its audio is crackly on the "ss's," which none of the other units had.
- The 9911 has more noise on both settings than I would have thought.
- The 9600 is significantly different from the 9911. It's "norm" mode is way too soft, to my eyes.
- The Panasonic's "Edit" setting is very noisy, but its "Nor" is actually not too bad.

I posted video clips because I believe that is preferred, but I can do still images too if needed (or anything else different). This is my first go at posting a comparison like this, so just let me know.

So what do you guys think? Which is the best clip? Thanks!

[lordsmurf]

What's the total running time of the tape?

[moxiecat]

25 minutes.

[orsetto]

Just for reference, the specific operating modes of some of these VCRs can be deceptive. There is no such thing as "all processing turned off" in the Panasonic AG1980, its DNR is always active. You can reduce its aggressiveness by setting the picture slider on the front panel to its middle detent, and switching the detail-norm-edit switch to edit, but some DNR luma/chroma processing is always on. The TBC is a separate function and can be turned on or off. On the JVC and Mitsubishi, the TBC and DNR processing are locked together so both are either always on or always off. The EDIT picture mode on the JVC is not available on the Mitsubishi, you can approximate EDIT by setting the Mitsu to SHARP but this isn't quite as "pure" as the JVC "edit" setting. OTOH, the Mitsu has newer TBC/DNR design than the JVC and can sometimes outperform the JVC if set to DNR/TBC "on" and picture mode to "Sharp": it depends on the tape. The "sharp" setting of the Mitsu is not nearly as artificial as the "sharp" setting on the Panasonic and slightly more natural than the JVC "sharp". These settings can be very subjective.

The crackling sound in the hifi audio playback of the Mitsu could have as easily occurred with the JVC or Panasonic: possibly the most consistently annoying problem with VHS transfer is the inconsistency of HiFi audio tracking tape to tape and machine to machine. If this specific project is very important, you have two choices. Assuming the video is far more important to viewers than the audio, use the Mitsu to play the tape but with the audio track set to Normal/Linear. This will result in more audible tape hiss, lesser tone quality monophonic sound but should eliminate the crackling noise. If compromising on the sound is not an option, and you need both picture and sound to be flawless, LordSmurf has sometimes advised doing separate transfers of the audio and video into a PC, using two different VCRs. With the right software, you can match the audio transfer from one VCR with the video transfer from another VCR. It isn't a task to take on lightly, you might want to have a pro do this for you if you think the project merits the expense. The only alternative would be to choose a compromise VCR that tracks the hifi audio without crackling and outputs a good (if not the best possible) image.

[moxiecat]

I'm pretty well versed in Final Cut Pro, so I could do separate jobs and match the video/audio--I've done that before. But then I would have to run the video into the Mac and encode it using a software encoder. (I've done that too.) But I have a hard time believing that the other six clips I've posted all have unacceptable video quality, to warrant this much effort.

For the purposes of this project, if one of the other six clips looks nearly as good as the Mitsubishi's output (or better?), that would be fine. I'm just interested in knowing which samples you guys think are the best.

[orsetto]

I have tried several times to view your clip links, using both Mac and Windows computers on different internet connections, and can't get any of the six clips to load: they all time out with a greyed Quicktime "Q" symbol in the center of the screen with a question mark overlay. If there is an issue preventing your clips from loading, people may be waiting to give their opinions to you until they can evaluate the clips.

[minidv2dvd]

orsetto - i used the right click save as option and copied them to the hard drive.

moxiecat - i'd pick the 9911u on norm

[moxiecat]

They're just MPEG-2 files, ripped using directly from discs burned on my Toshiba recorder using DVD Decrypter, and then edited and resaved in Womble. I would assume that any program you currently use to view files with the MPG extension would work (Power DVD, for example) That maybe does entail separating them from the browser.

However, I am also on a Mac using Quicktime, and they do work fine for me (Safari 3). Maybe your Quicktime needs a plugin. I haven't tried them on a PC but I will do so. I'm sorry, I don't know how else to present these clips without changing the quality of the original MPEG-2.

Thank you, Mini!

[orsetto]

minidvtodvd: thanks for the reminder about right-click-save, I must be in one helluva "instant gratification" mood today . Once downloaded as you suggested, I viewed on the PC and then burned a DVD to check on a reference CRT studio monitor.

moxiecat: I agree with you that the Mitsubishi NORM holds a very slight video edge over the JVC 9911 NORM with this particular tape, but the difference is small enough to not matter if you prefer the sound quality from the JVC. Both captures are very very nice. The others make it much more obvious the source is VHS and do not look nearly as good, the closest runner up is the AG1980 NORM.

Re the "brightening/darkening" symptom on the Mitsubishi: this clip is not long enough to really see it, but I know what you're talking about. The symptom is another one of those random interactions between certain tapes and certain VCRs. Today you see it on the Mitsubishi, tomorrow you might see it on the JVC with another tape, and a third tape might make the Panasonic do it. The TBC/DNR circuits in all these vcrs are a little odd and a mixed bag, each model filters the tape differently. Of course if one specific VCR, such as your Mitsu, is displaying consistent issues with brightness across a broad number of tapes, and none of the others shows the slightest issue, then the MGA could be out of alignment or defective. It would be unusual, but certainly possible with any VCR.

[moxiecat]

Thank you SO much for that effort, wow. Very appreciated, Orsetto.

The Mitsubishi is second-hand but never-used. I got it from someone who bought it as a backup to another VCR.

I was surprised that the 9600 was not closer to the 9911 in quality, but the "norm" setting in particular is significantly different. Does anyone think the 9600 should look better than this or did JVC make serious improvements in the circuitry from the 9600 to the 9911? (Or does my unit have problems?)

Unfortunately, the assessment so far that the 9911 on "norm" is the best takes me back the original problem: the 9911 has some jitter when playing the tape. It's not a lot; the picture just "jumps" once in awhile, but it's a symptom that did not happen on the Mitsubishi. The Mit had a much solider performance, aside from the audio issue (LOL, of course there must be something!). But I can try the 9911 again.

Anyone else have an opinion? I'm sorry if the clips are difficult to load; I didn't mean to hinder folks' willingness to help me.

[moxiecat]

One more thing to add: I do have longer versions of all seven of these clips. Not much longer--maybe about 10 seconds more. Both the opening scene and the title are longer, and there is another short scene showing the woman on a bus. I cut the clips back because I didn't want to overload either you guys or the server I'm using. But if anyone would like longer versions of any of these for comparisons, just let me know and I'll put them up too.

Thanks again for the help.

[themaster1]

noise analysis

Mitsubishi HS-HD2000U in NORMAL (no processing) / frame 252
http://img14.imageshack.us/img14/7267/mitsuframe252.jpg

JVC HR-S9911U in EDIT (no processing) / frame 233
http://img520.imageshack.us/img520/9924/jvcframe233.jpg

----------------
same frame analysed : no (not possible clipping on the jvc)
profiles quality identicals or close : yes (66% vs 62%)

result: mitsu win

[minidv2dvd]

the master - you've got it backwards don't you. check your own scores.

the jvc has the higher quality score on your tests. the mitsu has some nasty vertical banding in that frame section.

[lordsmurf]

1. is the 9911 on NORM
2. is the AG-1980P

No time to expand on this.

Understand this is only reflective of YOUR EXACT VCRs BEING USED. All of these results are all fairly close, but the minute differences in video quality highly depend on variables specific to those exact units. Anybody coming to this thread thinking this is some sort of comparison of models needs to realize that these results would likely different on a another set of the same models in another test.

That Mitsubishi is the worst of the lot, it's not even in the running. I've seen better consumer VHS VCRs. But even that could be a reaction to this tape -- it may be perfectly fine on other tapes.

Again, many variables -- tapes and power being but two of them.

[ministry88]

I'm no expert on VCR technology but to my eye the 9911 NORM and the AG1980 NORM look best. Not too soft, but not grainy which gives so much VHS its ugliness.

[2Bdecided]

If you capture with locked audio (no idea if your DVD-R machine will), it should be easy enough to swap in matching audio from another capture. I take the audio out to Cool Edit Pro (Adobe Audition), drop the original audio into the multitrack, drop the replacement audio into the multitrack, cut+drag to match, and timestretch if needed. Then just re-save the replacement audio.

It's dead easy if you have the right tools and you know what you're doing. About five minutes work.

Without the right tools, with drifting captured audio, and without a clue, I suppose it would take weeks.



Apart from the periodic brightness change, and if you're not going to do any subsequent processing, I preferred the mitsu_hshd2000u_normal.mpg - it removed most of the objectionable VHS noise without softening the picture.

The jvc_hrs9600u_norm.mpg was a joke.

If I was going to do post processing, I'd pick the one that kept most detail while delivering a stable picture - probably the jvc_hrs9911u_edit.mpg, though it doesn't seem to keep any more real usable information than the norm.

I think the Panasonic and the Mitsubishi over sharpen, but some people like this with VHS.

Cheers,
David.

[ZQX]

If you capture with locked audio (no idea if your DVD-R machine will), it should be easy enough to swap in matching audio from another capture. I take the audio out to Cool Edit Pro (Adobe Audition), drop the original audio into the multitrack, drop the replacement audio into the multitrack, cut+drag to match, and timestretch if needed. Then just re-save the replacement audio.

It's dead easy if you have the right tools and you know what you're doing. About five minutes work.

Without the right tools, with drifting captured audio, and without a clue, I suppose it would take weeks.

Yeah you make it sound easy. Fine if it is talking with inbetween moments of silence, but what about background music/noise etc? If you're chopping up your audio track and dragging/stretching segments around and cross-fading and patching in moments of complete silence etc... Man that sounds like a lot of work.

It's an important part of video editing thoough, so if you got any tips and techniques you should write 'em up here! I know Reaper can stretch audio while preserving pitch, for instance (just alt-right-click the end of a wav and drag it shorter or longer). It has an extremely good built-in "preserve-pitch-while-stretching" algorithm. I think Audacity has this ability too.

But hell, I have a few projects sitting here that remain a thorn in my side (eg syncing a VHS audio dub up with a DVD video rip, for instance - I keep putting this off). Another one is matching a pristine CD song rip to a band video clip. Frustrating stuff.

[moxiecat]

Thanks for all the posts.

If I were working with the video in a program like Final Cut Pro, I would definitely try matching video from one VCR with audio from another. It would be easy if you could set some markers and make sure everything matched up. But I'm nervous about doing it with Womble Video Wizard, which is a clunkier with its audio tools.

However, it seems as if the consensus seems to be that the 9911 on "norm" is the best video bet here. So if I can fix that jitter issue, that one had good audio as well. I can always try to re-run the sections where the picture visibly jumps and try to edit them in.

I'm surprised by LordSmurf's reaction to the Mitsubishi; I really thought the video on it was pretty good. Certainly close to the 9911 and Panny on their "norm" settings.

I will confess that I don't know how to read all the parameters that that Neat Video analysis, but thank you for doing that. That said, if the Mitsubishi rates a 62% and the JVC on "edit" rates a 66%, then doesn't that mean the results are actually fairly close? I do see the vertical band on the Mitsu in the blue box, but it is not in the closeup picture.

"The jvc_hrs9600u_norm.mpg was a joke."
I KNOW! Yikes. This is possibly the most surprising "find" of this entire experiment, how soft that picture is. I don't know if this means there is something wrong with my specific machine or the 9600 is much softer on the "norm" setting in general. It's bizarre since that machine is one of the best-respected on this forum.

I realize it always depends on the tape used and the specific VCRs. I was mainly interested in a general assessment of what my machines could do with a generally good-quality tape.

Thanks again.

[themaster1]

the master - you've got it backwards don't you. check your own scores.

the jvc has the higher quality score on your tests. the mitsu has some nasty vertical banding in that frame section.

Right the profile quality is better 66% i got it mixed up

same frame analysis:

mitsu
http://img221.imageshack.us/img221/6500/mitsu.jpg
jvc
http://img150.imageshack.us/img150/9731/jvc.jpg

Result: profile quality is better for mitsu (= more noise detected) BUT the luminance is not uniform for jvc
(=results biased somewhat)

I'd need a frame where there is no such uniformity but sadly you have removed the longer versions of your clips it seems and i can't find any now on the jvc .

[moxiecat]

I never put the longer versions online, but here the two you are are interested in:

Mitsubishi (longer): http://www.oneworldonehealth.org/video/mitsu_hshd2000u_normal_long.mpg

JVC on NORM/AUTO (longer): http://www.oneworldonehealth.org/video/jvc_hrs9911u_norm_long.mpg

The older clips linked in the first post are still online, but I will have to take all of them down eventually. They're all residing on the server of a nonprofit that may not be so generous with its space for me if the bandwidth goes crazy!

[robjv1]

Fixing the audio is a breeze with MPEG Video Wizard.

I had this same issue previously with hi-fi audio on my JVC SR-V101US and I use that program exclusively for editing. Just dub two copies of your tape (one with as perfect tracking as you can get in the linear audio mode, one with the hi-fi audio and who cares how the picture looks) on your DVD recorder and rip each of them to your computers HD as two big VOBs using DVD Decrypter or what have you. Drag the one with the perfect picture onto your timeline in the video track and right click on it to mute this videos audio track. Then drag the one with the perfect audio into the audio track below.

To get them in synch, you can use MPEG Video Wizards frame accurate editing to cut them each on the same start and end frame, or if you want to wing it a little, you can un-mute the video track and play the both tracks simultaneously while dragging the audio track until they are matched in real time. Once you get them pretty close in the default view, right click your timeline and select "1 frame" and then drag one frame at a time -- this way you'll be able to hear it real time and get it precisely where you want it. You will be able to tell that you have it right when the audio will have no "echo" at all in either direction.

Go ahead and hit export (the record button) and double check in the details that there is no audio re-encoding going on. If there is, then that is your clue that you didn't match the audio up just right -- and you can tell by how much if you zoom in on the timeline down to one frame and check out of the end of your timeline -- the audio and video tracks should be of exactly the same length.

Either way works -- no need to use any other fancy audio tools, time-expand, or any of that.

[lordsmurf]

To get them in synch, you can use MPEG Video Wizards frame accurate editing to cut them each on the same start and end frame,

Great method, I use this myself sometimes. But I do still use Premiere for more complex splicing work. Uncompressed captures for that one, encode out to Adobe MainConcept.

[ministry88]

Orsetto: THANK YOU for explaining that maddening brightness/darkening issue on the Mitsu VCR! MY GOD I thought I was going crazy when I tried explaining what I was seeing on my VCRs! Each VCR of mine behaves differently with this issue, depending on the tape and the setup. For example, my Toshiba doesn't do it at all UNLESS it's hooked up to my Pioneer 340H DVD Recorder. Then it sometimes does it for certain VHS.

On the other hand, my JVC HR-S7600U does it once in a while whether it's hooked up to the recorder or not. Strangely, my cheapo Panasonic VCR does it the least, but just last night when I was watching a tape on it it did it all over the place, but not at all on the JVC! Crazy. I guess there's nothing you can do. Some VHS just don't play well with some VCRs. I guess this is why all the pros have a wide array of VCRs for each funky situation that can arise with stubborn tapes...

[robjv1]

Yeah, MPEG Video Wizard is (for me) my most favorite application when working with VHS tapes captured to a DVD recorder. I've used it to match video/audio from multiple tapes and it works pretty much flawlessly for this kind of simple stuff.

Just FYI -- the only "bug" I can think of is that if while playing both tracks (video+audio in track one, and an additional audio track) to see if the audio is in synch using your ears is that if you skip ahead too fast in the timeline, the audio will sometimes sound out of synch though it really isn't. For some reason the program (at least on my computer) doesn't like it if you make huge jumps along multiple audio streams for this purpose, but if you do it in small increments (maybe 5 minutes at a time) you will get an accurate read on the situation. This is kind of the quick and dirty method though or used as a spot check after you've made your cuts.

I have encountered one or two instances where the audio does seem to have drifted a frame and upon close examination, one frame would be duplicated twice for some reason. I don't know if this is a product of my VCR or TBC dealing with a video error or an issue within the program, but I haven't seen in some time, so it might have just been a fluke thing.

[moxiecat]

Thanks again for everyone's help.

I love Womble's program too, but I have run into a different issue with it. Let's say you edit four sections together from the same VOB file (in order to take out commercials, for example). If the bit rate for the VOB is high, 8 mbps+ or so, Video Wizard will sometimes create a bitrate spike for no reason at an edit point when it saves your edited timeline as an MPEG. The MPEG can then cause the authoring and build in Adobe DVD Encore to fail (maybe other programs too--that's just what I use). When the MPEG is put through an bit rate analysis program, the spike is obvious, and since it's not in the original VOB, it's a spike that was created by Womble's program. I have worked with Womble to try to figure this out, but they have no answer.

So now, I keep my bit rates 7.8 Mbps or under to keep Encore happy. Has this happened to anyone else?

[robjv1]

I'm not sure if this exactly addresses your issue since it sounds like you are working from one VOB -- but I do know that working with MPEG Video Wizard, if you're using DVD Decrypter then you need to rip the DVD with the file splitting set to "None" instead of "By File" or "Automatic". This is because at whatever point where there is a change in the VTS (VTS_01 --> VTS_02) then you will get a glitch (perhaps what you are seeing as a bitrate spike) in the audio and video that is quite noticable. I believe LordSmurf suggested this a few years ago on here and it completely allieviated my issue with it.

I've never had a problem otherwise though with the program and I pretty much do all my DVD recorder captures in high-bitrate modes (XP mode or FR-80 mode on my JVC DR-M100). I do however do all of my authoring in DVD Lab Pro 2.

[lordsmurf]

Yes, VOBs should never be merged like separate files. The breaks between them are rarely correctly re-merged. It's always best to rip as the single stream it is, based on the IFO data on the disc. Once you lose the IFO data, you're sort of on your own, and it may give all kinds of errors, depending on software.

Follow this guide: http://www.digitalfaq.com/guides/video/edit-dvd-recorder.htm

[moxiecat]

Thanks for the advice, but yes I'm aware of the file splitting and have always had it set to "none" in Decrypter. And it always only happens at the edit point(s) I create when editing sections of one VOB together. Sometimes it doesn't happen, so it's not a constant error, but by lowering the bitrate I've been able to avoid it for the most part.

While Womble is creating the spike (verified by the analysis program), it's possible that Encore is more sensitive to it than other authoring programs.

[lordsmurf]

Encore tends to only like videos encoded out by Premiere.

[moxiecat]

I've been OK with Encore, as long as the bitrate is below 7.8 and the encoding was VBR. If the encoding is higher than that or CBR, I just have to avoid too many edit points in the file and then it's usually OK too. But, I use MPEGs created with both my Toshiba recorder/Womble and Final Cut Pro/Compressor in Encore regularly, with very few problems.