Ok, a few tests, for what they're worth. Pay no attention if you don't care about graphs and numbers. I bought both the 70D and a used 702 and I decided I would keep the one the one that had better "quiet room" measurement of noise using a high quality pair of mics (I do some nature ambiences), having just got the 702 delivered today, I had to do the test, of course. I was asking for the data anyway, so...
Three recorders:
Sony PCM-M10 with internal mics, Tascam DR-70D fed by an M/S pair (Sennheiser MKH40/30), and a Sound Devices 702, fed by the same M/S pair.
Arrangement:
In the quietest room that I have at my disposal (which isn't anechoic by any means
), a 440 Hz sine wave tone was played at a fixed volume through a loudspeaker and this tone was used to set all gain settings for all recorders to a steady -20 dB readout on the level meters. Then, the tone was turned off, all lights/HVAC/etc was turned off and I hit record and walked away while 44.1/96 files were recorded for about 1 minute. I dumped all of the files in SoundForge Pro 11 and selected the quietest 20 seconds of each file and displayed the statistics. I also screen captured the power spectra for the 20-20,000 region and a more detailed region from 2-10 kHz, where the most obvious difference was. Take the information for what it is, an imperfect measurement.
Data:
Here is an Excel file to grab:
https://drive.google.com/file/d/0B-o5v-jZ9NUjVDkzMy1IUzViVW8/view?usp=sharing (This is an Excel spreadsheet. Once the file opens in the viewer, if you are using Chrome, that is, you should hit 'open in Google Sheets'. Hopefully that allows you to look at the file if you don't have Excel.)
My take:
I interpret it as under these conditions, the DR-70D holds it's own quite well against the SD702 and both are noticeably (even audibly, hiss-wise, when played through loudspeakers at my normal listening volume) than the Sony recorder/internal mic combo (uh, no surprise??).