SCD-777ES or SCD-1

Both units are identical inside, except that SCD-1 has transformer shields and an additional PCB at XLR to derive balanced output from single-ended analog stage. The additional processing for XLR may even degrade signal relative to RCA output. In RCA mode they are indistinguishable. In stock fitment they are average at best, but well-modified examples can hold their own with current hi-end players.

Put another way, at $800-900 for a used 777ES vs. $1800 or more for SCD-1, 777ES is a good deal for anyone interested in performance/value or mod potential. Moreover, a 777ES can easily be converted to true balanced operation through one of several pro mod shops.

Tvad, admittedly I don't get out much. The only top players I've listened to closely are Zanden, DCS stack, and top Reimyo(RBCD comparisons only), as well as comparisons to vinyl in my own system and to stock 777ES and Theta Gen 5a that I keep for reference. IMHO, none of the above approach RBCD on my battery-powered SCD-1, which is based largely on Reference Audio Mods ideas & parts, twists from Allen Wright and Yamada-san, and also various experiments of my own. I've put some things into it that from a T&M and risk perspective are not particularly viable from a commercial modder's perspective, but having moved progressively through many iterations of upgrades, my sense is that you can get fairly close to this level of performance for $2-4K spent with RAM. Unfortunately I have not heard the VSEI L7 mod, which in balanced version prices out for around $3K and offers the benefit of a productized, drop-in, warrantied purchase.

One thing I know is it's all about improvements to clock, power, and analog section. The "heart" of the unit-- the digital section and DAC-- is very strong. Too much emphasis is placed on the latest developments in the digital domain(e.g. upsampling, parallel DACs, etc.), while ignoring built-to-cost compromises in other critical areas. If one accepts this viewpoint, then these two Sonys remain good platforms for continuing mods.

As you imply, mods are not for everyone, as most of the value is lost on the resale side. But if you are a keeper, the Sony has almost unlimited potential.

I'm at a loss for having not heard the APL/Denon, but if it gets close to the big boys that I mentioned, then that underscores the value proposition of modding.

Similarly, in the RAM full-out mod, the entire power section is replaced with aftermarket discrete DC regulation circuits, the one-chip clock oscillator is replaced with a Superclock PCB, and the entire analog output section is replaced with buffered Audio Consulting silver transformers that as piece parts alone can cost up to $2K.

What's interesting is that essentially the same approach is repeatable across a variety of CDP models. The differentiation between machines that remains after these mods are done can be regarded as the inherent sonic signature of the OEM DAC and transport/RF section themselves, rather than the product of compromised stock power and analog sections.

When stripped back to laser, RF, and DAC, are all machines essentially the same? They are not. Allen Wright has written that the STACT/VC24-based Sonys and also those like the SCD-1/777ES with the sliding platter assemblies, are especially good mod platforms. When selecting a mule for modding, the cost and effort of the add-on work may exceed the value of original unit several-fold, so it is best to select a model with the greatest potential.

The Denon may be good as well, and as you point out, Alex has gone further than others by gutting the DAC from the machine and substituting AKM replacements. But even assuming that the AKM arrays are the best of all possible DACs, the question remains as to whether his analog section, clock, and power assemblies are equal to alternative assemblies and piece parts. The player's performance is the synergistic whole of these considerations.

I would add from my own experience with AC and DC mods, that properly implemented large SLA batteries take things to an altogether higher level than the best AC regulated power-- perhaps an equalizer in the tech wars. There are no firm rules as to how the mods all net out, of course.

In this context, I'll add a thought illustrating how engineering choices to improve performance are sometimes available to the modder, but not to the manufacturer.

The many PCB traces that carry power to a DAC chip are bypassed by capacitors at the chip pins, typically by a half dozen or so low-value surface-mounted ceramic caps costing pennies apiece, even in megabuck CDPs. It's an important function, but space and cost considerations preclude using better pieces. However, with a few hours work it's possible to sqeeze highest-quality Teflon caps near enough to these critical locations. Luckily, the russian cold war surplus variant of what would be a $50 piece part if manufactured today, is available direct from eastern Europe for $2. So in go six around one DAC chip for a major improvement in sound. It doesn't take a genius to engineer this modification (in fact, even a genius could not cram 72 such parts around an array of 12 AKM DACs; it is not a viable parts choice for the APL design). Nor is this choice available to a commercial manufacturer, who for this one tweak alone might need to charge something like 4x markup x $50/unit cost x qty 6 pieces per DAC = $1000+.

Just an example of the unusual options that can be cost-effective only for the modder-- and in this particular example, only for a machine using a single DAC.