There are so few reviews and screenshots of MUSE/Hi-Vision laserdiscs. In every case there is a mention of either picture being too dark and/or having excessive compression artifacts.

MUSE compression is different than anything we know(MPEG2/4,AVC,VC etc.) and there is really no detailed explanation anywhere what these artifacts look like. From little explanation from various sources, I gather they are more like mosquito noise and/or block noise. Also these artifacts are more visible during motion scenes and low light.

I am curious if any of todays technologies can help these artifacts. Many HT receivers and even HDTVs have MNR(mosquito noise reduction) and block artifact reducer. I happen to own Algolith Mosquito HDMI(2 of them actually) analog/digital compression artifact noise reducer. It is probably the most advance device for what it does. I can tweak below adjustments;

3D Temporal Filter (for motion)
2D Spatial Filter (for static parts of the image)
MNR/DNR switch(MNR for digital/DNR for analog sources)
BAR(block artifact reducer)
Detail

These below are the explanation of each from the owner's manual

MPEG noise (or mosquito noise) is random noise visible along edges of the image and is part of any compressed signal. Today, the most widely available compression scheme is MPEG2. The Mosquito HDMI will correct all DCT based compression artifacts, such as those found in MPEG2.

When the image source is not compressed, DNR mode should be selected. «Dynamic» or «Gaussian» noise appears as dynamic or moving grain in an image. This type of noise is sometimes observed on low intensity camera captures from analog sources such as VHS, or from bad reception of analog signals from cable or an antenna. It is not the result of digital data compression, but can be diminished when your Mosquito HDMI is set to DNR mode.

Blocking artifacts or «macro-blocking» occur whenever images are digitally compressed. These artifacts are most visible in fast moving images like sports scenes and action scenes. To reduce blocking artifacts, set the BAR mode to on. Block artifact reduction blends and diminishes the inherent 8x8 pixel block structures of MPEG2 images.

I wonder if any of these could help this older compression technology. I will soon have Hi-Vision setup. I am planning to make screen shots and reviews of the components I have.

Really there is no block noise, because blocks are not used. And a noise-reducer designed for DCT (frequency-domain compression) artifacts will not function normally on a pure time-domain signal. The DNR for non-compressed images might be somewhat better, but I would be concerned about blurring.

The other thing to consider is that the entire MUSE encoding-decoding chain is essentially an elaborate process of 2D (spatial) & 3D (spatial-temporal) filtering. So, unexpected results might occur when applying other forms of filtering.

I encourage you to experiment with the functions of this device, but real improvements in image quality may prove elusive. It's worth noting that at least some MUSE decoders include controllable noise-reduction circuits.

the decoder is Sony MSC-4000.
Do you know if it has built in noise reduction?
Are these LDs encoded in YPbPr or GBR colorspace?
Do you know if MSC-4000 processes video in GBR or YPbPr?

They are YPbPr. A laserdisc movie encoded in RGB would need to be the size of King Arthur's Round Table.

MUSE discs are encoded in a line-sequential YPbPr colorspace. In other words, each line contains a Y signal, & either a Pb or a Pr signal, which are time-compressed. Here you can see what this looks like : the stairstep on the right hand is the luminance signal, & the business on the left is the two color-difference signals, overlapping because this is a time exposure of an analog oscilloscope screen, rather than a capture from a digital scope.

Because of the use of "transmission gamma" & "pseudo-constant luminance encoding", part of the processing is done in the YPbPr domain, then there is a transform to the GBR domain, & then the final YPbPr signal (if selected) is created by re-transforming the GBR.

The MSC-4000 undoubtedly includes some form of noise-reduction circuitry, but that does not appear (from the manual) to be user-controllable. The user adjustments for the picture are just tint & color saturation. The beast that is the Panasonic TU-AHD100 does have certain user-controllable features such as "MUSE-AI", which seems to be some combination of noise reduction & other enhancement features.

Do you think panasonic d100 is a better muse decoder than the rest? I always thought msc4000 was newer therefor the best

There are those who say that the Panasonic TU-MDC100 is the best MUSE decoder around. Certainly it seems to be a 1997 model, according to my information, whereas the MSC-4000 & the Victor HV-MD2 are 1996 models. But they're scarce, & I don't have one.
My MUSE decoders are :
Panasonic TU-AHD100N
Victor HV-VMD1
Sony MST-2000
Sony MSC-3000