Classic Phish Show From 1991 Surfaces On YouTube
By Scott Bernstein Aug 12, 2013 • 10:00 am PDT
JamBase: Clinton, can you give us a little background about yourself?
Clinton Vadnais: I’ve been listening to Phish on and off for 20 years. Back in the ’90s I was a fairly avid VHS trader. Back when we used to trade hand written lists through snail mail and dub tapes with two VCR’s. I have also been recording music/audio for over 15 years. I’ve gained a lot of experience in post production: mixing, mastering, remastering, restoration, etc. For about half of that time I have also been focusing on working with video production and post production. I’ve also been involved in some fringe projects with members of Phish over the years including quite recently. Hopefully I can talk about that more soon.
JB: How did you get involved with this project?
CV: In mid June I happened to see a Facebook post on a private Phish group by user “Jackie Treehorn,” aka Dexter Deddo. After some back and forth it turned out that his friends Mike Burns and Zach Nabor happened on these rare VHS tapes in Zach’s fathers basement. We think his friend shot it and dubbed it for him back in the ’90s, though we are not really sure. Hopefully the videographer surfaces.
JB: Can you take us through the process of adding a better audio source to the footage?
CV: While I was waiting for the tapes to arrive I decided to track down the known circulating audio recording. Funnily enough in 2013 it can be hard to find lossless audio for a particular show. MP3’s of their entire career can be had without effort. Luckily a fellow audio enthusiast “T Watts” aka Terry had it and made it available for me. The recording varied slightly in quality. The files were from an analog to digital transfer from over ten years ago by “Mattress.” Unfortunately I’m not sure who originally recorded it. It was a soundboard tape but it varied in overall quality.
Balancing the time I could spare, the tools I have available and my overall skill set I did my best to enhance what was there. This is of course a multi-step process. First, decompress the lossless files to WAV files. Then, seamlessly combining the 23 individual tracks into their original chunks in a DAW (digital audio workstation). In this case I assume this is four cassette sides. A tape flip after Reba in the first set and Stash in the second set, and after Good Times Bad Times before the encore. From there, the files are unconverted from CD standard 16bit, 44.1khz to 24bit, 48khz audio -not to increase the quality but to match the video specs and give more resolution for the upcoming processing.
From there, I did some restoration to the tracks. In 2013 there are some tools available that are almost magic compared to what I remember using in the ’90s for restoration and noise reduction. Over the course of a week or so I would carefully listen to the recording while looking closely at a spectrogram display, addressing flaws along the way. The types of things I worked on during this process included removing minor ticks or clicks, ground buzz, actual feedback from the PA system, slight dropouts and even over saturation in some passages. I went so far as to replace about a second of missing audio from the end of “The Lizards” by using the (luckily) repeated phrase from the next pass of the phrase. Here I also used some amazing tools to occasionally reduce just Trey’s overwhelming (in the mix) notes by a few decibels without effecting anything sitting around it sonically. From here, there is some more standard general sweetening. After accessing the needs which did vary from chunk to chunk I essentially balanced volumes as needed, applied equalization. I used some tools to try to better balance and enhance the stereo image and help give a strong center and widen the field some. Also overall compression to tame the dynamics, bring up and even out the overall volume.
As the source was a club soundboard recording the levels vary through the show. Sometimes Trey’s guitar is way on top of the mix, sometimes the organ is dreaming and hitting the console and/or tape pretty hard. I also assume Mike’s bass was quite loud on stage as it is low in the mix. To a great degree there is only so much that can be done to balance these things. Since these processes push the volume of the noise floor up along with the music the last thing I did was some gentle noise reduction. In fact, I only processed frequencies above 8000hz and below 250hz or so leaving a lot of mid-range untouched -a careful balance between removing some tape noise and leaving the music intact. Again, noise reduction in 2013 is a lot better than the earlier digital days but you don’t want to hear any artifacts.
JB: How did you process the video?
CV: Somewhere in this process the VHS tapes arrived safe and sound. I transferred the two VHS tapes to my non linear video editing software in a standard definition 720×480 format. At this point it was quite clear that these unfortunately were not master tapes. They showed signs of being unknown generation copies. The image was ghosting, soft, blown out, noisy and lacking detail. There was also some burned in tracking displays and the overall image had VHS distortion and overscan noise on the bottom. I had some back and forth trying to figure out if we would be able to somehow get closer to the master tape. Unfortunately this was probably the only version we would be able to put out hands on.
Being that this was indeed the uncirculated 4/5/91 show and it was complete, I decided to move forward. Since there was distortion on every edge of the image I scaled the image up slightly to keep the pixel size and hide the edge distortions without masking it with a fake border or something. The next step was to get the new audio to match the video. The original tape actually had a soundboard feed as well. Of course the quality was much worse than that of the cassette transfer. It did come in quite handy for syncing. One song at a time I used a time stretch function to slightly adjust the audio track to keep sync with the source video. Scrubbing through the whole video and adjusting as needed throughout. I then omitted full frames of distortion where they appeared. The worst being over a minutes worth during Reba. Leaving that black keeping the audio underneath. In once instance I fixed a few missing frames by manipulating the timing of adjacent frames. For the most part it was intact.
I stopped short of trying to address the overall shakiness of the footage. While it was shot with a tripod from the front of house (soundboard area) there is a lot of general shakiness. It would have taken a lot of rendering with mixed results to remove this. In some places it would have worked great. In others less so. So I left it as is. After all of the audio was in sync I rendered some new video files. These files were brought into a different video editor where I used some of the only viable commercial intelligent noise reduction software. Using this I analyzed the noise and come up with a filter that would take away quite a big of the noise without over effecting the information we did want to keep. This again is something that was in my world not possible even a few years ago. In addition to this some mild color work to punch things up a little and bring the black down a bit the image was starting to look pretty darn good for what it is. Then it was just a matter of rendering for a couple of nights leaving me with two new master video files. I then added some quick faux VHS titles, credits, chapter markers and metadata for the show creating a final 640×480 video file.
JB: How have you shared your work?
CV: While some of the final odds and ends were being completed I mocked up a bit of a sample video to show the before and after.
Though it’s not the actual final video it did help promote and get folks to subscribe to my channel in advance. When it was completed I decided to snip the songs into their own files. This would allow for individual songs to be viewed and shared more easily. I took a night to upload the nine gigs of video to my YouTube. The following day I added all of the YouTube metadata, acknowledged the third party content and created a playlist that would allow for the whole show to be played through in its entirety. Then I shared it on the various outlets online. I did get a lot of requests for the new audio so I seeded a bit torrent on eTree. I also uploaded some FLAC and MP3 files to Mega for direct download. Then a lot of folks where asking about a download for the video as well. I created and uploaded some mp4 files to also share on Mega. I really hope that nobody in the Phish organization minds it being shared. I also hope that the fans enjoy being able to see and hear this recording. Ultimately that is what it’s all about.
Without further ado, check out a playlist featuring Clinton’s work on the April 5, 1991 Phish show:
