NOTE: to keep this report as short as possible I have chosen not to illustrate all meters. There is enough info on the internet to get a view of any of these meters.

The following list of exposure meters has been evaluated.

______________________________
Weston Master V, 1963 - 1972
Sekonic L-398A Studio de Luxe III, 1978 - current
Minolta Flashmeter VI, 2003, later known as Kenko KFM-2100
Sekonic L-478D with 5° attachment, 2012 - current
Sekonic L-758D, 2001 - 2017
Sekonic L-858D, 2017(?) -current
Gossen Lunasix 3, 1966 - 1980 (Lunasix 3s)
Gossen Mastersix, 1983 - 1990
Gossen Starlite 2, 2001(Starlite) - current
Gossen Digisky, 2011 - current
____________________________

This list spans a period from 1950 to 2020. Handheld external exposure meters were required when mechanical cameras were made without a meter. At first exposure meters were considered a toy, not worthy of the true photographic craftsman who could estimate the intensity of the scene illumination by experience or used one of the many tables to ‘calculate’ the exposure values. Using an exposure meter slowed down the speed of taking pictures. It has been said of Cartier-Bresson that he could guess the exposure values (aperture and speed combinations) quite accurately. Never mind that his darkroom assistants had to cope with severe under and over exposed negatives. The other method is also true: C-B knew what his film could handle and by experience took only pictures when the ambient light fell within the latitude of his pre-set values. Two problems however had to be confronted: cope with the contrast range within a scene and the fact that the film reacted differently to intensity of the brightness areas. The classical examples are the black cat in a dark barn and the white bear in a snowy landscape. The thick two-layer emulsions of those days could handle a wide contrast range and also over- and underexposure. Experience taught the photographer to underexpose when confronted with a dark scene and to overexpose when the scene was brighter than the average one.
The problem of accurate exposure became more pressing when the miniature format gained popularity. The quality of the print depended heavily on the exposure: over- and underexposure will reduce the resolution and the tiny areas of tonal differences are more difficult to reproduce in print.
With the incorporation of through-the-lens exposure metering in so-called professional cameras, like the Konica Autoreflex-T, introduced in 1968. This particular model was designed for the professional market with its TTL metering and fully automatic exposure control, setting the norm for all professional cameras since. The lenses for the Konica model were also particularly good, optically, and suitable for professional tasks, but were overshadowed by the more popular Nikon and Canon ranges and the mythical Zeiss and Leitz ranges.
The last film-cartridge loading camera models used extremely sophisticated exposure algorithms, based on the analysis of thousands of film rolls and typical and a-typical scenes. When Nikon introduced its matrix exposure system, and Canon used its multi-segmented metering areas system, the technical integration of versatile metering, the classical hand held exposure meter faced extinction! The fatal blow for this instrument came when the solid-state sensor replaced the film emulsion. Now every pixel could become an exposure meter. The idea of combining focus points with exposure points was the next step.
Whatever the sophistication, the evaluation of the luminance values of every pixel resulted in one specific EV (exposure value). The algorithm how the software weighs and averages the individual values is unknown. The user is hardly interested, because the selected EV is almost always accurate and at least sufficient as a base for subsequent post processing. The cursory look at the histogram is not a real alternative for the function of the program, because the histogram only shows the distribution of the brightness values over the range of 0 to 255. It does not tell you which parts of the scene have which values. And it is still left to the photographer to decide what parts of the scene are important and what is the best exposure for these parts.
One of the best implementations of TTL exposure measurement can be found in the famous Olympus OM-3 and OM-4. There is a spot meter option, coupled to a high and dark selection. Use the spot to aim at a bright spot and the high option will increase EV by two stops, ensuring that this area is at the top of the characteristic curve. For the dark areas the opposite rule applies. It is a crude, but effective approximation of the Zone System. The only caveat is the fact that the spot diameter changes with the used focal length. This is a problem with all so-called narrow angle meters in a camera body. One of the advantages of the handheld spot meter is the consistency of the one degree spot. Whatever you aim at to measure, it is always a one degree measurement.
Here we have one of the advantages why it makes sense to use an external exposure meter: the spot meter option. Measuring several different brightness areas in the scene and averaging these gives a good idea of the final EV. The second advantage is the option of incident metering. Much has been written about this technique. The idea is simple: the illuminance of a scene depends on the brightness of the main source. Measuring this source is the only thing that matters. The various reflectance values of the parts of the scene are not relevant. With a sun high in the sky it makes no difference if we take a picture of a white or a dark wall. The EV is identical. Incident measurements were the favorite of many Directors of Photography on film sets. It is easy to measure the main and the secondary light sources and balance both. The Sekonic Studio de Luxe is the latest version of such a meter. It has limited sensitivity (only 4 EV at ISO100). It is a battery-less meter, a great companion for the battery-less mechanical cameras (as example: the Leica M3, Leica M-A, the Nikon F and the Canon Rangefinders).
The main advantage of this meter is the simple operation, combined with a maximum of information. The large dial shows at one glance all combinations of aperture and speed for still photography and cine, ISO settings and EV values. Under- and overexposure values by one and two stops can be easily selected. In fact this dial tells you everything you really need to know about correctly exposing a scene. The method of measuring reflected light is supported, but can not be recommended.

The other meter of this type is the Weston Master, here in its final version (V). The Weston Master V is a more modern version of the original Weston Master I, made in England from 1947. The Westons are famous for their accuracy and the special shape of the Invercone dome, converting the meter into an incident-type instrument. The shape of the dome gives accurate EVs even when the main subject is backlighted. The dial presents the user with a number of controls: in addition to the ones on the Sekonic meter, there is a brightness range scale and a method for reading extremely low light and extremely high light areas. Reflected readings are also quite easy and deliver accurate results.

The evolution of the exposure meter shows an increasing tendency to flash metering and radio control of external flash units. The parallel trend is to more sensitivity, flexibility and versatility.
More sensitivity and the simplicity of the traditional scales is the strong point of the Gossen Lunasix. Its range is now extended to -4 EV, a level almost matched by the new Sekonic L-858.
The Lunasix was a purely photographic meter. Gossen produced the next generation of meters, taking advantage of then new electric circuits and micro-processor. The Mastersix was a very versatile meter with additional accessories for specialized applications. It could master all photographic tasks and also the basic photometric tasks. It has the same sensitivity as the Lunasix (-4EV) and can be used as a scientific instrument.
Its size is a limitation and the latest addition to the Gossen range, the Digisky, is more compact. (size of the Mastersix: 70 x 130 x 34 mm versus Digisky: 60 x 139 x 16). The sensitivity has been reduced to -2.5 EV. This level is still enough for most scenes. The meter compensates this with an extended range of features, like flash readings, mixed readings for flash plus ambient light and most importantly a range of options for the cinematographer. A color display supports the analysis of the main values. The meter lacks a spot option. If this is important the Gossen range has the Starlite 2.
This meter incorporates all measurements of the Mastersix and adds the unique option of the Zone System readings. Dip switches are needed to change between function groups. The meter is very sensitive and it takes some time to get stabilized readings. The versatility for cine and photometric measurements is a bit too much for most users, but when needed everything is there, including a range of flash readings.
The layout is ergonomic, but the dip switches ask for some pre-configuration to get the optimum one-hand operation.
The change from incident light to reflected light via spot meter is done by turning a ring on the measurement unit. This is a bit inconvenient and the Minolta Flashmeter VI has a better solution. Changing between the two modes is easy. The meter is sensitive to -2EV (spot is only 2EV) and has separate buttons for Average, Highlight and Shadow readings. The range can be adjusted in the Custom Settings. Standard it is -3 stops for Shadow reading and + 2, 5 stops for Highlight readings. The latitude function shows the latitude of the preferred film emulsion and indicates when the measured subject contrast range is inside or outside the specified latitude. The Custom Functions are quite flexible and must be the maximum that an analogue meter can handle.
The Sekonic L-758D was announced in 2001. It is a unit optimized fro still photography with digital cameras. There is a wired connection with a computer to set the specific latitude for the sensor response for the minimum and maximum light levels that the sensor can handle. The cine and photometric values had to be omitted and a specific version L-758C was introduced for movie requirements. The basic L-758D has a flexible way of combining readings from ambient(spot and incident) and flash light. These modes require study and insight into what is happening at the scene. This requirement is indeed one of the main reasons to use a separate handheld meter. Using a meter gives insight into what is involved in the simple final exposure value. The camera may respond with a specific f-number and shutter speed, but there is no information why the camera has selected these parameters.
The L-758 can measure low light levels to -2EV (spot 1EV), can calculate and change the ratio of flash and ambient light readings and can do averaging readings and light contrast analysis.
During the first decades of the twenty-first century color LCD and touch-sceens became the rule for smart phones and cameras. Sekonic and Gossen went back to the drawing board and came back with the L-478D and the Digisky.
The L-478D has a sensitivity of -2 EV with the Lumisphere (reflected light with separate attachment 3EV) and incorporates all functions of the L-758D and C (but without the spot meter facility. The separate attachment (for five degree measurement) is cumbersome and not really a fine addition. For ambient (incident) light readings, flash metering and analysis and personalized display settings for still photography and cinema, the instrument is quite unique. There is one problem with all these flash meters and that is the measurement of high speed flash (stroboscopic and non-studio flash). Studio flash is rather slow and most meters can catch the peak of the flash energy. On-camera and in-camera flash has a very short duration and this cannot be measured by standard flash meters.
The recent L858 does away with this critique. It combines the touch screen of the L-478 with the options found on the L-758D/C and extends even on the software flexibility. There are no longer buttons on the instrument and everything is changed with the options on the touch screen. This is both flexible and slow: instead of just pressing a button, you now have to wait through a menu of options to find the one you need. The sensitivity has been enhanced to -5EV for incident readings and -1EV for the spot meter. The ability to measure HSS flash peak values and flash duration is a bonus for those photographers who use flash for freezing movement, but will be of limited use for the available light photographer who needs to measure dimly-lit scenes.

The measurements for comparison
All measurements were done in a diffusely lit room with all meters (in incident mode) pointing in the same direction. Spot metering was done in the same room with the target a grey card, made by Fotowand. ISO setting is 100.
Below are are the results in EV:
Weston Master V (Invercone)……………………………………..9.7
Sekonic L-398A Studio de Luxe III (Lumisphere)……………..10.3
Minolta Flashmeter VI
incident…………………………………………………………… 10.3
spot 1 degree………………………………………………………10.5
Sekonic L-478D
incident…………………………………………………………… 10.4
spot 5 degree………………………………………………………10.5
Sekonic L-758D
incident……………………………………………………………10.7
spot 1 degree……………………………………………………10.8
Sekonic L-858D
incident……………………………………………………………10.5
spot 1 degree……………………………………………………10.6
Gossen Lunasix 3
incident …………………………………………………………10.0
reflected…………………………………………………………..10.5
Gossen Mastersix
incident ……………………………………………………………10.2
reflected……………………………………………………………10.5
Gossen Starlite 2
incident……………………………………………………………10.4
spot 1 degree……………………………………………………10.8
Gossen Digisky
incident ……………………………………………………………10.0
reflected……………………………………………………………10.5
Note: The EV values for the incident and spot meter modes are not comparable, because they are generated in different conditions. The target for the spot readings is the grey card. This method is good for comparing the calibration of the meter.

Between the reading of the Weston master and the Sekonic L-758D there is a difference of one stop. This is partly due to the different method of calibration. Every meter needs to be calibrated in relation to a known light source. In addition there is also the design philosophy of where the average reading should be placed on the characteristic curve. The Weston Master clearly assumes that a slight over exposure (black and white film) is good for the shadow detail in the scene. The Sekonic on the other hand assumes that slight under exposure is the best solution (it is for slide film and general sensor capture). The Sekonic can be individually calibrated, so the choice is of secondary importance. All other meters are within a third or half stop difference. Hardly important for the average black and white or colour negative film. This difference is only significant when working at the extreme ends of the film or sensor latitude.

Why use a handheld exposure meter?
This question needs to be answered from several perspectives: functional, educative and fun.
Functionally the exposure meter is required when using a meterless camera (which is obvious), when a true one degree spot meter reading is needed and when there is a substantial use of flash light in ambient light scenes. It makes sense to use an external exposure meter even when the camera has a built-in TTL metering system and has a sensor. The exposure latitude of a sensor is less than commonly assumed: a spot meter is a great tool to make sure the exposure of the important areas of the scene is within the useable contrast range. An incident reading is also advisable when the scene has unusual brightness distributions. In all other cases there is no functional argument for using an external meter: the modern in-built TTL metering systems were and are remarkably efficient.
The educational role of the external meter is the most important asset of this method of exposure control: a spot meter reading of important areas in a scene and averaging these values and comparing them to the values that the camera proposes is often interesting and an inspiration for thinking. Using the Zone System or the Highlight and Shadow Buttons lets the photographer control effectively the tonal values of the scene, irrespective of this is recorded by film or by digital means.
The fun factor is the last but certainly not the least argument. Not being dependent on what the exposure algorithms inside the camera propose and making a decision oneself, based on a thorough analysis of the scene and a knowledgable prediction of the behavior of the recording medium, is fun. The technological dependency in photography is generally quite high and the exposure is too important to leave it to some program, however efficient.

What meter do I use?
I have two meters in my bag:
The Sekonic Studio de Luxe for battery independent incident readings
The Gossen Starlite 2 for spot metering, Zone System analysis and photometric analysis (this is fun and not directly related to the photographic tasks.

If only one meter is required (because of space requirements in my small Billingham Hadley bag I select the Gossen Digisky or the Minolta Flashmeter VI or the Sekonic L-858D (for its high sensitivity). The choice depends on the envisaged application (Gossen: compact; Minolta: ergonomics; Sekonic: flexibilty).
When using a classic camera (Leica M3, Canon VIL or Canon F-1), I match the meter to the camera (easy to handle Weston Master or the scientific Mastersix).
Luckily I have the choices!