Noise in Sony Hi8 EVO-9850 Playback

[Jim Norman]

I've started capturing about 100 hours of Hi8 video and am encountering problems with the playback deck I wish to use. I have two Hi8 decks. The first is a Sony EVO-9850 which is a high end deck with lots of adjustments available. This is the deck I would prefer to use because of the availability of these adjustments. The problem I am encountering is that there are occasional dropouts as exhibited by white streaks shown in the attached screen capture. When I play back the same tape on a Sony EV-S5000 deck, the noise isn't there. I have cleaned the heads on the EVO9850 several times, but it hasn't helped. Does anyone have any suggestions?

[edDV]

Those are dropouts which are typical on Hi8 tape. These are most common at the start and end of the tape where the bend is greatest.

I can't explain a tape playing differently on two players unless the heads are dirty. A tape dropout is caused by magnetic particles separating from the tape binding. Metal evaportative tapes were worse for dropouts vs metal particle tapes.

[Jim Norman]

I also assumed that the white lines were a function of tape dropout. I suspect that the magetic signal may be a little weak where this is occuring. I captured the same tape area on the EV-S5000 and also on a TR700 camcorder and did a same frame compare and neither of those show dropouts. The EVO-9850 has only 242 drum hours and has been cleaned several times, so it is a mystery as to why it is having these problems. I recaptured the same tape area from the EVO-9850 again and in almost all cases, the white lines occur in the same frames as before so it is tape related. I also played back a much newer back-coated tape and saw an occasional white line when playing on the EVO-9850, but not with the EV-S5000. It appears that the EVO-9850 head is not as sensitive as it should be.

Since I really wanted to use the additional video controls on the EVO-9850, I may box it up and send it to Sony for their inspection/repair.

[Jim Norman]

I just spoke with Sony and they no longer service the EVO-9850.

[jpw]

Jim-- I sympathize with you. I have an EV-S7000 prosumer Hi8 deck and it exhibits the identical behavior as your 9850. I also have a Digital8 camcorder that--even in Hi8 playback mode--virtually never exhibits a white dropout. QUESTION: Does the 9850 show dropouts when you twist the shuttle knob to the slowest few slo-mo speeds? I ask because the 7000 never seems to show dropouts this way . . . yet it will in still frames or faster slo-mo. I wonder what's going on, and if maybe it's a grounding problem with the head, and those are static-induced traces or something. . . . What's terrible is that the camcorder will cut off a portion of the very first second or two of audio on clips that follow a blank portion of tape. That is making it awfully tough to digitize my archival Hi8 tapes. It's totally disappointing. -John in Buffalo

[budwzr]

Um...I think the white streaks are the least of your troubles, no one will care or notice. It's the low light noise and darkness that's the biggest concern.

The lines seem to follow the interlace comb. Maybe if you de-interlace it later the problem will be solved.

Love them weiner dogs

[jpw]

Sgt., I'd agree with your take on no one caring if, and only if, my footage were of a snowstorm, sugar factory or maybe Johnny or Edgar Winter. . . . I think you would take my side if you had valuable archival footage.

[budwzr]

Sgt., I'd agree with your take on no one caring if, and only if, my footage were of a snowstorm, sugar factory or maybe Johnny or Edgar Winter. . . . I think you would take my side if you had valuable archival footage.

OK, buddy No harm intended.

[jpw]

Same here, thanks--I'm just trying to keep it light here while looking for that needle-in-a-haystack solution. It's a challenge. True, some folks may never even notice the static!

[elandy]

Anyone know where to fix one of these?

[Jim Norman]

Well, it's been 3 years since I posted the original question. I got busy with other projects and put the Hi8 video capture on hold. I am finally getting back to this project. I also acquired a second EVO-9850 which did not exhibit the dropouts on the same tape area. I found a comment from Adam Wilt who reported that dropouts can often be eliminated on some decks by tweeking the Dropout Compensation (DOC) control. Using a copy of the Service manual for the EVO-9850, I found that the DOC was the RV11 potentiometer on the DM-87 board which is the third board from the left in the 6 card cage on the right side of the EVO-9850 when viewed from the front of the deck. Normally this adjustment would be made using a test tape and extender board, neither of which I had or were available. I ended up pulling the board, making a minor adjustment, reinstalling and retesting. This took several trys as the adjustment was very critical. The first several times resulted in the picture getting much worst. I was almost ready to give up when I made a very minor tweek and hit the sweet spot and all dropouts in this portion of the tape went away. I now have two EVO-9850s with extremely good playback quality.

My next task is to try to make a color calibration tape (I have a calibrated source) and use it to tweek the Luminance, Chroma, Setup, and Hue adjustments before finally starting the capture of the 100 hours of Hi8 video. Any tips on making color calibration tapes or setting these adjustments would be appreciated. If someone has a color calibration tape to sell, loan, or rent out, I would be interested.

Jim Norman

[Brad]

Congrats on the success, and thanks for the detailed post.

The obvious thing to try to create a color calibration tape is recording DVD test patterns to a tape, but I suspect the results you get won't be very meaningful. Even if you had recorded bars to the original tapes using the original camcorder before the footage (and the video input on it didn't skew the results at all), the camera probably didn't record "accurate" colors anyway.

[Cornucopia]

Don't know if this is a possibility for you (depending on the durability of the tapes), but - and I know this would greatly add to the complexity of your work and the delay of its finishing - you COULD use multiple machines to do multiple passes (2x, 3x) of the same tape.
With multiple captures of analog material, the SIGNAL (program you are wanting to keep) is mixed in with noise that is a function of the recording medium/device/methodology (and is now BURNED IN as part of the signal), as well as with noise that is a function of the playback medium/device/methodology. The latter noise is NOT burned in, and will be random, changing EVERY time you play back the tape.

You can remove that 2nd noise (or rather, greatly diminish it), by combining multiple takes of the same program with a MEDIAN filter (median to the dataset of the multiple takes, not the median of a certain area as is normally defined). Similar to Time-based frame averaging, it removes/minimizes the random playback noise and retains the SIGNAL (along with the burned-in recording noise). Nothing's perfect, sorry. But less noise is always better, isn't it?.

Dropouts/sparkles/blacktails/comets are usually a playback-originated noise. The very thing that can be diminished, using this method.

<edit>BTW, in AVISynth, the filters to use for this are either MergeRGB/Merge/MergeLuma/MergeChroma (depending on your sources, etc).

Scott