Couple comments, first a disclaimer- I haven't heard a whole bunch of different Grados but haven't cared for the mid-range emphasis of those I have heard. That could be related to the 1.9kHz peak you mention in your particular model. I haven't heard the newer models. Generally I prefer the less aggressively "up-front" Sennheiser "house-sound" which is smoother to my ear and easier to manipulate to subjectively flat, at least for me.
I made two foam removal mods with the Senn HD650 which are my critical reference phones, to open up the top so that they sound naturally-neutral and properly balanced to me both for normal listening and better critical-listening mix assessments. One removed a small piece of damping foam in the driver assembly (not reversible). The other replaced the foam screens in front of the driver with thin, black nylon stocking material (easily reversible).
You may be able to get close to what you want with careful EQ. If using software EQ, you might try a few different plugins to determine which allows you to dial in the mostly subjectively well balanced "air" up top, probably using a shelf-filter.
If you want to really get fancy with EQ, you can custom EQ them with an HRTF technique to balance them for a perceptually flat response for a signal oriented directly ahead. The easier "slap-dash" way of doing this is to place a speaker with a flat or preferred timbre directly in front of you and run a pink noise source through it, an also route the pink noise through your EQ > headphone chain. Go back and forth listening to the speaker and listening to the headphones, and EQ the headphones such that they sound as close to the response from the speaker as possible. You'll need to mute the speaker output when switching to the headphones, and you should sit in the near-field range of the speaker to reduce the influence of the room.
The fancy, less subjective and more accurate way of doing this is to put binaural microphones in your ears and record both the response as heard from the from the speaker, and also the response heard with the speaker off and the headphones on, worn over the binarual microphones. Take the resulting curve of the loudspeaker response, subtract the inverted response of the headphones from it, and apply that as the corrective EQ to the headphones. That corrects for both your own HTRF and the response of the headphones, to the accuracy of the speaker. If done well, the result can fully achieve a very natural "out of head" headphone listening experience.
I'm curious about your comment about replacing the Mytek Stereo 96 DAC. I use the same DAC and am very happy with it. I felt no need at all to replace it.