So you've just imaged your original Quake disc to CUE/BIN or CUE/ISO/OGG. After firing up Quake in DOSBox and listening to Trent Reznor's ambient soundtrack, you notice that it sounds harsh and has an overabundance of high frequencies -- not quite the dark and moody BGM you remember. What's up with that?
The audio tracks of some (perhaps not all) Quake discs were mastered with pre-emphasis, a noise reduction technique used to increase the signal-to-noise ratio of audio recordings. A de-emphasis filter should be applied to such audio tracks for them to sound as intended. CD-ROM drives usually apply the de-emphasis, but only on their analog output -- digital output, digital audio extraction, and ripping of audio tracks does not filter the PCM stream.
DOSBox does not filter CD audio in images, but you can do the de-emphasis processing with a tool like SoX. For CUE/ISO/OGG you process the WAV files before encoding to OGG; and for CUE/BIN you first make a CUE/ISO/WAV, process the WAV files, then assemble into a CUE/BIN (e.g. mount the CUE/ISO/WAV in a virtual drive like Daemon Tools and make an image from it).
Please leave a comment in this thread with the titles of DOS games other than Quake that are known to have audio tracks with pre-emphasis. I'm not aware of any, but it would not surprise me if there are others. However, be aware that there are two ways that pre-emphasis is noted on audio tracks: a flag in the table of contents that is unmistakable, and/or a flag in the subcode that is easy to overlook. In order to be certain, EAC can do subcode analysis to determine if pre-emphasis is present, or you can play the disc in an audio CD player and see if the emphasis indicator appears.
It looks like pre-emphasis can be signalled in the CUE file using the FLAGS command, maybe dosbox could look for this and automatically apply de-emphasis during playback?
Perhaps, if enough games could benefit, but otherwise I don't think it would be worth the effort. Analog domain audio processing seems kind of heavy to run in parallel with emulation for the single-threaded DOSBox, but supporting DSP plugins on a separate CPU thread could have lots of interesting possibilities.
Perhaps, if enough games could benefit, but otherwise I don't think it would be worth the effort. Analog domain audio processing seems kind of heavy to run in parallel with emulation for the single-threaded DOSBox, but supporting DSP plugins on a separate CPU thread could have lots of interesting possibilities.
Vorbis/MP3/FLAC decoding (or whatever the cool kids are storing their audio tracks as these days) already requires that sort of processing power and seems to work just fine.
Heh, and this whole time I thought the Quake soundtrack sounded awesome without any sort of de-emphasis processing. I regularly listen to harsh industrial music though, so what do I know?
Not all Quake CDs have pre-emphasis, I think only the first release.
I noticed my Assault Rigs PC has pre-emphasis in the cue, while Assault Rigs PSX does not.
And yes I own both, as well as the Ati Rage version, (but I haven't tested the pre-emphasis on it).
Both do not have pre-emphasis in the subcode. So, I don't think Assault Rigs PC/PSX uses pre-emphasis.
Both PC/PSX Assault Rigs have identical CRCs for the extracted audio.
You can download the IBM PC cuesheets from r****p.org and search the cue files for the pre-emphasis flag, but to be certain you should also check the subcode.
You can download the IBM PC cuesheets from r****p.org and search the cue files for the pre-emphasis flag, but to be certain you should also check the subcode.
Brilliant idea, if ripsaw8080 wants to experiment... might give a rough idea of how many did this...
"I see a little silhouette-o of a man, Scaramouche, Scaramouche, will you
do the Fandango!" - Queen
Dang, am I going to be mad now that I had purged the *.sub files from my images collection? Had presumed they'd be without purpose since I wasn't hoping burn exact CD duplicates with them.
This is the result of "data mining" the PC cue sheets from r****p.org that indicate pre-emphasis on audio tracks. I can only confirm the Quake disc that I have. The KQ6 disc in my collection appears to be an earlier release than the one in the list, and I find it odd that a later release would have pre-emphasis. Because the pre-emphasis flag in the subcode is more reliable than the flag in the TOC, I used the "Detect TOC Manually" option in EAC to test the Quake and KQ6 discs.
DOS
Alone in the Dark 2 (Australia)
Assault Rigs (Europe)
CyClones (Europe) (Rerelease)
Decent II (Europe)
Decent II (USA) (v1.0)
Destruction Derby (Europe)
id Anthology (USA) (Disc 3) (Quake)
King's Quest VI - Heir Today, Gone Tomorrow (USA) (19940112)
Might and Magic - Clouds of Xeen (USA)
Out of This World (Europe)
Quake (World)
Quake (World) (Demo)
Quake (USA) (v1.01) - confirmed
Quarantine (Europe)
Ring Cycle (Europe)
Ring Cycle (Germany)
Road Warrior (Europe) - [may be Quarantine II, not sure]
Sam & Max Hit the Road (Europe)
Sam & Max Hit the Road (USA)
WipEout (Europe)
Windows / Windows 3.x
CyberSpeed (USA)
EarthSiege 2 (USA)
Silent Thunder - A-10 Tank Killer II (Germany)
Silent Thunder - A-10 Tank Killer II (USA) (v1.00)
I finished testing my own collection of PC games with CD audio (thanks Stretch for the cdda2wav info and link) and can confirm that Descent 2 version 1.0 does have the pre-emphasis flag. However, the info I got from cdda2wav looks confusing to me:
1D:\CDRTOOLS>cdda2wav dev=0,0,0 -J 2 3Type: ROM, Vendor 'ATAPI ' Model 'iHDS118 7 ' Revision 'VL08' MMC+CDDA 4983040 bytes buffer memory requested, transfer size 64512 bytes, 4 buffers, 27 sectors 5 6cdda2wav: Read TOC CD Text failed (probably not supported). 7 8#Cdda2wav version 3.00_cygwin32_nt_1.7.5(0.225-5-3)_i686_i686, real time sched., 9 soundcard, libparanoia support 10 11172297 sectors of data before track #2, ignoring. 12 13 DATAtrack recorded copy-permitted tracktype 14 1- 1 uninterrupted no data 15 16AUDIOtrack pre-emphasis copy-permitted tracktype channels 17 2-13 yes no audio 2 18 19Table of Contents: total tracks:13, (total time 68:23.59) 20 1.[38:17.22], 2.( 0:40.51), 3.( 2:36.44), 4.( 1:55.18), 5.( 2:29.42), 21 6.( 3:02.53), 7.( 2:28.33), 8.( 3:23.28), 9.( 2:50.14), 10.( 2:45.36), 22 11.( 2:42.44), 12.( 2:18.02), 13.( 2:53.47) 23 24Table of Contents: starting sectors 25 1.( 0), 2.( 172297), 3.( 175348), 4.( 187092), 5.( 195735), 26 6.( 206952), 7.( 220655), 8.( 231788), 9.( 247041), 10.( 259805), 27 11.( 272216), 12.( 284410), 13.( 294762), lead-out( 307784) 28 29CDINDEX discid: T6LOfbGwKTzPOfd98RPd4mqMnwc- 30 31CDDB discid: 0xd410070d 32 33CD-Text: not detected 34 35CD-Extra: not detected 36 37No media catalog number present. 38 39scanning for ISRCs: 13 ... 40 41Skipping data track 1... 42 43index scan: 2...difference: TOC:with, subchannel:without preemphasis 44correcting TOC... 45 46difference: TOC:without, subchannel:with preemphasis 47correcting TOC... 48 49difference: TOC:with, subchannel:without preemphasis 50correcting TOC... 51 52index scan: 3...difference: TOC:with, subchannel:without preemphasis 53correcting TOC... 54 55difference: TOC:without, subchannel:with preemphasis 56correcting TOC... 57 58difference: TOC:with, subchannel:without preemphasis 59correcting TOC... 60
I can confirm that Sam & Max Hit the Road (v1.11, U.S.) has the pre-emphasis flag set on all audio tracks, even after using the manual TOC detection feature in EAC (see also the cdda2wav output below). I can also confirm that playing the tracks unchanged makes them sound almost exactly like the Roland SC-55, from which they were recorded; if anything, the SC-55 is even brighter. De-emphasizing the tracks (sox deemph) makes them sound overly bassy and muffled. With the Roland SC-55 output as a reference, I conclude that the pre-emphasis flag was set by mistake during mastering.
1#Cdda2wav version 3.02a06_cygwin32_wow_2.4.1(0.293-5-3)_i686_i686, real time sched., soundcard, libparanoia support 2103081 sectors of data before track #2, audible data at sector 0. 3Hidden audio track with 103081 sectors found. 4 DATAtrack recorded copy-permitted tracktype 5 1- 1 uninterrupted no data 6AUDIOtrack pre-emphasis copy-permitted tracktype channels 7 2- 5 yes no audio 2 8Table of Contents: total tracks:5, (total time 35:17.62) 9 1.[22:54.31], 2.( 2:02.42), 3.( 3:36.39), 4.( 1:39.19), 5.( 5:05.06), 10 11Table of Contents: starting sectors 12 1.( 0), 2.( 103081), 3.( 112273), 4.( 128512), 5.( 135956), 13 lead-out( 158837) 14CDINDEX discid: 19tatED.8sMtWw6WacAxHueIZTg- 15CDDB discid: 0x45084505 16CD-Text: not detected 17CD-Extra: not detected 18scanning for MCN... 19No media catalog number present. 20 21scanning for ISRCs: 2 ... 22scanning for ISRCs: 3 ... 23scanning for ISRCs: 4 ... 24scanning for ISRCs: 5 ... 25Skipping data track 1... 26 27index scan: 2... 28index scan: 3... 29index scan: 4... 30index scan: 5...
